scmrw.cpp@ 94913

Last change on this file since 94913 was 94913, checked in by vboxsync, 3 years ago
scm: New heuristic to trigger only on 'HRESULT rc' and 'int rc' declarations as the other one was too broad and triggered on things which we don't want it to trigger on, bugref:10223
Property svn:eol-style set to `native` Property svn:keywords set to `Author Date Id Revision`
File size: 137.6 KB

Line
1	/* $Id: scmrw.cpp 94913 2022-05-08 19:07:29Z vboxsync $ */
2	/** @file
3	* IPRT Testcase / Tool - Source Code Massager.
4	*/
5
6	/*
7	* Copyright (C) 2010-2022 Oracle Corporation
8	*
9	* This file is part of VirtualBox Open Source Edition (OSE), as
10	* available from http://www.virtualbox.org. This file is free software;
11	* you can redistribute it and/or modify it under the terms of the GNU
12	* General Public License (GPL) as published by the Free Software
13	* Foundation, in version 2 as it comes in the "COPYING" file of the
14	* VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15	* hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16	*/
17
18
19	/*********************************************************************************************************************************
20	* Header Files *
21	*********************************************************************************************************************************/
22	#include <iprt/assert.h>
23	#include <iprt/ctype.h>
24	#include <iprt/dir.h>
25	#include <iprt/env.h>
26	#include <iprt/file.h>
27	#include <iprt/err.h>
28	#include <iprt/getopt.h>
29	#include <iprt/initterm.h>
30	#include <iprt/mem.h>
31	#include <iprt/message.h>
32	#include <iprt/param.h>
33	#include <iprt/path.h>
34	#include <iprt/process.h>
35	#include <iprt/stream.h>
36	#include <iprt/string.h>
37
38	#include "scm.h"
39
40
41	/*********************************************************************************************************************************
42	* Structures and Typedefs *
43	*********************************************************************************************************************************/
44	/** License types. */
45	typedef enum SCMLICENSETYPE
46	{
47	kScmLicenseType_Invalid = 0,
48	kScmLicenseType_OseGpl,
49	kScmLicenseType_OseDualGplCddl,
50	kScmLicenseType_OseCddl,
51	kScmLicenseType_VBoxLgpl,
52	kScmLicenseType_Mit,
53	kScmLicenseType_Confidential
54	} SCMLICENSETYPE;
55
56	/** A license. */
57	typedef struct SCMLICENSETEXT
58	{
59	/** The license type. */
60	SCMLICENSETYPE enmType;
61	/** The license option. */
62	SCMLICENSE enmOpt;
63	/** The license text. */
64	const char *psz;
65	/** The license text length. */
66	size_t cch;
67	} SCMLICENSETEXT;
68	/** Pointer to a license. */
69	typedef SCMLICENSETEXT const *PCSCMLICENSETEXT;
70
71	/**
72	* Copyright + license rewriter state.
73	*/
74	typedef struct SCMCOPYRIGHTINFO
75	{
76	/** State. */
77	PSCMRWSTATE pState; /*< input /
78	/** The comment style (neede for C/C++). */
79	SCMCOMMENTSTYLE enmCommentStyle; /*< input /
80
81	/** Number of comments we've parsed. */
82	uint32_t cComments;
83
84	/** Copy of the contributed-by line if present. */
85	char *pszContributedBy;
86
87	/** @name Common info
88	* @{ */
89	uint32_t iLineComment;
90	uint32_t cLinesComment; /*< This excludes any external license lines. /
91	/** @} */
92
93	/** @name Copyright info
94	* @{ */
95	uint32_t iLineCopyright;
96	uint32_t uFirstYear;
97	uint32_t uLastYear;
98	bool fWellFormedCopyright;
99	bool fUpToDateCopyright;
100	/** @} */
101
102	/** @name License info
103	* @{ */
104	bool fOpenSource; /*< input /
105	PCSCMLICENSETEXT pExpectedLicense; /*< input /
106	PCSCMLICENSETEXT paLicenses; /*< input /
107	SCMLICENSE enmLicenceOpt; /*< input /
108	uint32_t iLineLicense;
109	uint32_t cLinesLicense;
110	PCSCMLICENSETEXT pCurrentLicense;
111	bool fIsCorrectLicense;
112	bool fWellFormedLicense;
113	bool fExternalLicense;
114	/** @} */
115
116	/** @name LGPL licence notice and disclaimer info
117	* @{ */
118	/** Wheter to check for LGPL license notices and disclaimers. */
119	bool fCheckforLgpl;
120	/** The approximate line we found the (first) LGPL licence notice on. */
121	uint32_t iLineLgplNotice;
122	/** The line number after the LGPL notice comment. */
123	uint32_t iLineAfterLgplComment;
124	/** The LGPL disclaimer line. */
125	uint32_t iLineLgplDisclaimer;
126	/** @} */
127
128	} SCMCOPYRIGHTINFO;
129	typedef SCMCOPYRIGHTINFO *PSCMCOPYRIGHTINFO;
130
131
132	/*********************************************************************************************************************************
133	* Global Variables *
134	*********************************************************************************************************************************/
135	/** --license-ose-gpl */
136	static const char g_szVBoxOseGpl[] =
137	"This file is part of VirtualBox Open Source Edition (OSE), as\n"
138	"available from http://www.virtualbox.org. This file is free software;\n"
139	"you can redistribute it and/or modify it under the terms of the GNU\n"
140	"General Public License (GPL) as published by the Free Software\n"
141	"Foundation, in version 2 as it comes in the \"COPYING\" file of the\n"
142	"VirtualBox OSE distribution. VirtualBox OSE is distributed in the\n"
143	"hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.\n";
144
145	/** --license-ose-dual */
146	static const char g_szVBoxOseDualGplCddl[] =
147	"This file is part of VirtualBox Open Source Edition (OSE), as\n"
148	"available from http://www.virtualbox.org. This file is free software;\n"
149	"you can redistribute it and/or modify it under the terms of the GNU\n"
150	"General Public License (GPL) as published by the Free Software\n"
151	"Foundation, in version 2 as it comes in the \"COPYING\" file of the\n"
152	"VirtualBox OSE distribution. VirtualBox OSE is distributed in the\n"
153	"hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.\n"
154	"\n"
155	"The contents of this file may alternatively be used under the terms\n"
156	"of the Common Development and Distribution License Version 1.0\n"
157	"(CDDL) only, as it comes in the \"COPYING.CDDL\" file of the\n"
158	"VirtualBox OSE distribution, in which case the provisions of the\n"
159	"CDDL are applicable instead of those of the GPL.\n"
160	"\n"
161	"You may elect to license modified versions of this file under the\n"
162	"terms and conditions of either the GPL or the CDDL or both.\n";
163
164	/** --license-ose-cddl */
165	static const char g_szVBoxOseCddl[] =
166	"This file is part of VirtualBox Open Source Edition (OSE), as\n"
167	"available from http://www.virtualbox.org. This file is free software;\n"
168	"you can redistribute it and/or modify it under the terms of the Common\n"
169	"Development and Distribution License Version 1.0 (CDDL) only, as it\n"
170	"comes in the \"COPYING.CDDL\" file of the VirtualBox OSE distribution.\n"
171	"VirtualBox OSE is distributed in the hope that it will be useful, but\n"
172	"WITHOUT ANY WARRANTY of any kind.\n";
173
174	/** --license-lgpl */
175	static const char g_szVBoxLgpl[] =
176	"This file is part of a free software library; you can redistribute\n"
177	"it and/or modify it under the terms of the GNU Lesser General\n"
178	"Public License version 2.1 as published by the Free Software\n"
179	"Foundation and shipped in the \"COPYING\" file with this library.\n"
180	"The library is distributed in the hope that it will be useful,\n"
181	"but WITHOUT ANY WARRANTY of any kind.\n"
182	"\n"
183	"Oracle LGPL Disclaimer: For the avoidance of doubt, except that if\n"
184	"any license choice other than GPL or LGPL is available it will\n"
185	"apply instead, Oracle elects to use only the Lesser General Public\n"
186	"License version 2.1 (LGPLv2) at this time for any software where\n"
187	"a choice of LGPL license versions is made available with the\n"
188	"language indicating that LGPLv2 or any later version may be used,\n"
189	"or where a choice of which version of the LGPL is applied is\n"
190	"otherwise unspecified.\n";
191
192	/** --license-mit
193	* @note This isn't detectable as VirtualBox or Oracle specific.
194	*/
195	static const char g_szMit[] =
196	"Permission is hereby granted, free of charge, to any person\n"
197	"obtaining a copy of this software and associated documentation\n"
198	"files (the \"Software\"), to deal in the Software without\n"
199	"restriction, including without limitation the rights to use,\n"
200	"copy, modify, merge, publish, distribute, sublicense, and/or sell\n"
201	"copies of the Software, and to permit persons to whom the\n"
202	"Software is furnished to do so, subject to the following\n"
203	"conditions:\n"
204	"\n"
205	"The above copyright notice and this permission notice shall be\n"
206	"included in all copies or substantial portions of the Software.\n"
207	"\n"
208	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND,\n"
209	"EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES\n"
210	"OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND\n"
211	"NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT\n"
212	"HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,\n"
213	"WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n"
214	"FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR\n"
215	"OTHER DEALINGS IN THE SOFTWARE.\n";
216
217	/** --license-mit, alternative wording \#1.
218	* @note This differes from g_szMit in "AUTHORS OR COPYRIGHT HOLDERS" is written
219	* "COPYRIGHT HOLDER(S) OR AUTHOR(S)". Its layout is wider, so it is a
220	* couple of lines shorter. */
221	static const char g_szMitAlt1[] =
222	"Permission is hereby granted, free of charge, to any person obtaining a\n"
223	"copy of this software and associated documentation files (the \"Software\"),\n"
224	"to deal in the Software without restriction, including without limitation\n"
225	"the rights to use, copy, modify, merge, publish, distribute, sublicense,\n"
226	"and/or sell copies of the Software, and to permit persons to whom the\n"
227	"Software is furnished to do so, subject to the following conditions:\n"
228	"\n"
229	"The above copyright notice and this permission notice shall be included in\n"
230	"all copies or substantial portions of the Software.\n"
231	"\n"
232	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
233	"IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
234	"FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL\n"
235	"THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR\n"
236	"OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,\n"
237	"ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR\n"
238	"OTHER DEALINGS IN THE SOFTWARE.\n";
239
240	/** --license-mit, alternative wording \#2.
241	* @note This differes from g_szMit in that "AUTHORS OR COPYRIGHT HOLDERS" is
242	* replaced with "THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS".
243	* Its layout is wider, so it is a couple of lines shorter. */
244	static const char g_szMitAlt2[] =
245	"Permission is hereby granted, free of charge, to any person obtaining a\n"
246	"copy of this software and associated documentation files (the \"Software\"),\n"
247	"to deal in the Software without restriction, including without limitation\n"
248	"the rights to use, copy, modify, merge, publish, distribute, sublicense,\n"
249	"and/or sell copies of the Software, and to permit persons to whom the\n"
250	"Software is furnished to do so, subject to the following conditions:\n"
251	"\n"
252	"The above copyright notice and this permission notice shall be included in\n"
253	"all copies or substantial portions of the Software.\n"
254	"\n"
255	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
256	"IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
257	"FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL\n"
258	"THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,\n"
259	"DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR\n"
260	"OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE\n"
261	"USE OR OTHER DEALINGS IN THE SOFTWARE.\n";
262
263	/** --license-mit, alternative wording \#3.
264	* @note This differes from g_szMitAlt2 in that the second and third sections
265	* have been switch. */
266	static const char g_szMitAlt3[] =
267	"Permission is hereby granted, free of charge, to any person obtaining a\n"
268	"copy of this software and associated documentation files (the \"Software\"),\n"
269	"to deal in the Software without restriction, including without limitation\n"
270	"the rights to use, copy, modify, merge, publish, distribute, sublicense,\n"
271	"and/or sell copies of the Software, and to permit persons to whom the\n"
272	"Software is furnished to do so, subject to the following conditions:\n"
273	"\n"
274	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
275	"IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
276	"FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL\n"
277	"THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,\n"
278	"DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR\n"
279	"OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE\n"
280	"USE OR OTHER DEALINGS IN THE SOFTWARE.\n"
281	"\n"
282	"The above copyright notice and this permission notice shall be included in\n"
283	"all copies or substantial portions of the Software.\n";
284
285	/** --license-(based-on)mit, alternative wording \#4.
286	* @note This differs from g_szMitAlt2 in injecting "(including the next
287	* paragraph)". */
288	static const char g_szMitAlt4[] =
289	"Permission is hereby granted, free of charge, to any person obtaining a\n"
290	"copy of this software and associated documentation files (the \"Software\"),\n"
291	"to deal in the Software without restriction, including without limitation\n"
292	"the rights to use, copy, modify, merge, publish, distribute, sublicense,\n"
293	"and/or sell copies of the Software, and to permit persons to whom the\n"
294	"Software is furnished to do so, subject to the following conditions:\n"
295	"\n"
296	"The above copyright notice and this permission notice (including the next\n"
297	"paragraph) shall be included in all copies or substantial portions of the\n"
298	"Software.\n"
299	"\n"
300	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
301	"IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
302	"FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL\n"
303	"THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n"
304	"LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n"
305	"FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER\n"
306	"DEALINGS IN THE SOFTWARE.\n";
307
308	/** --license-(based-on)mit, alternative wording \#5.
309	* @note This differs from g_szMitAlt3 in using "sub license" instead of
310	* "sublicense" and adding an illogical "(including the next
311	* paragraph)" remark to the final paragraph. (vbox_ttm.c) */
312	static const char g_szMitAlt5[] =
313	"Permission is hereby granted, free of charge, to any person obtaining a\n"
314	"copy of this software and associated documentation files (the\n"
315	"\"Software\"), to deal in the Software without restriction, including\n"
316	"without limitation the rights to use, copy, modify, merge, publish,\n"
317	"distribute, sub license, and/or sell copies of the Software, and to\n"
318	"permit persons to whom the Software is furnished to do so, subject to\n"
319	"the following conditions:\n"
320	"\n"
321	"THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n"
322	"IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n"
323	"FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL\n"
324	"THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,\n"
325	"DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR\n"
326	"OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE\n"
327	"USE OR OTHER DEALINGS IN THE SOFTWARE.\n"
328	"\n"
329	"The above copyright notice and this permission notice (including the\n"
330	"next paragraph) shall be included in all copies or substantial portions\n"
331	"of the Software.\n";
332
333	/** Oracle confidential. */
334	static const char g_szOracleConfidential[] =
335	"Oracle Corporation confidential\n"
336	"All rights reserved\n";
337
338	/** Licenses to detect when --license-mit isn't used. */
339	static const SCMLICENSETEXT g_aLicenses[] =
340	{
341	{ kScmLicenseType_OseGpl, kScmLicense_OseGpl, RT_STR_TUPLE(g_szVBoxOseGpl)},
342	{ kScmLicenseType_OseDualGplCddl, kScmLicense_OseDualGplCddl, RT_STR_TUPLE(g_szVBoxOseDualGplCddl) },
343	{ kScmLicenseType_OseCddl, kScmLicense_OseCddl, RT_STR_TUPLE(g_szVBoxOseCddl) },
344	{ kScmLicenseType_VBoxLgpl, kScmLicense_Lgpl, RT_STR_TUPLE(g_szVBoxLgpl)},
345	{ kScmLicenseType_Confidential, kScmLicense_End, RT_STR_TUPLE(g_szOracleConfidential) },
346	{ kScmLicenseType_Invalid, kScmLicense_End, NULL, 0 },
347	};
348
349	/** Licenses to detect when --license-mit or --license-based-on-mit are used. */
350	static const SCMLICENSETEXT g_aLicensesWithMit[] =
351	{
352	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMit) },
353	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMitAlt1) },
354	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMitAlt2) },
355	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMitAlt3) },
356	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMitAlt4) },
357	{ kScmLicenseType_Mit, kScmLicense_Mit, RT_STR_TUPLE(g_szMitAlt5) },
358	{ kScmLicenseType_OseGpl, kScmLicense_OseGpl, RT_STR_TUPLE(g_szVBoxOseGpl)},
359	{ kScmLicenseType_OseDualGplCddl, kScmLicense_OseDualGplCddl, RT_STR_TUPLE(g_szVBoxOseDualGplCddl) },
360	{ kScmLicenseType_VBoxLgpl, kScmLicense_Lgpl, RT_STR_TUPLE(g_szVBoxLgpl)},
361	{ kScmLicenseType_Confidential, kScmLicense_End, RT_STR_TUPLE(g_szOracleConfidential) },
362	{ kScmLicenseType_Invalid, kScmLicense_End, NULL, 0 },
363	};
364
365	/** Copyright holder. */
366	static const char g_szCopyrightHolder[] = "Oracle Corporation";
367
368	/** LGPL disclaimer. */
369	static const char g_szLgplDisclaimer[] =
370	"Oracle LGPL Disclaimer: For the avoidance of doubt, except that if any license choice\n"
371	"other than GPL or LGPL is available it will apply instead, Oracle elects to use only\n"
372	"the Lesser General Public License version 2.1 (LGPLv2) at this time for any software where\n"
373	"a choice of LGPL license versions is made available with the language indicating\n"
374	"that LGPLv2 or any later version may be used, or where a choice of which version\n"
375	"of the LGPL is applied is otherwise unspecified.\n";
376
377	/** Copyright+license comment start for each SCMCOMMENTSTYLE. */
378	static RTSTRTUPLE const g_aCopyrightCommentStart[] =
379	{
380	{ RT_STR_TUPLE("<invalid> ") },
381	{ RT_STR_TUPLE("/*") },
382	{ RT_STR_TUPLE("#") },
383	{ RT_STR_TUPLE("\"\"\"") },
384	{ RT_STR_TUPLE(";") },
385	{ RT_STR_TUPLE("REM") },
386	{ RT_STR_TUPLE("rem") },
387	{ RT_STR_TUPLE("Rem") },
388	{ RT_STR_TUPLE("--") },
389	{ RT_STR_TUPLE("'") },
390	{ RT_STR_TUPLE("<end>") },
391	};
392
393	/** Copyright+license line prefix for each SCMCOMMENTSTYLE. */
394	static RTSTRTUPLE const g_aCopyrightCommentPrefix[] =
395	{
396	{ RT_STR_TUPLE("<invalid> ") },
397	{ RT_STR_TUPLE(" * ") },
398	{ RT_STR_TUPLE("# ") },
399	{ RT_STR_TUPLE("") },
400	{ RT_STR_TUPLE("; ") },
401	{ RT_STR_TUPLE("REM ") },
402	{ RT_STR_TUPLE("rem ") },
403	{ RT_STR_TUPLE("Rem ") },
404	{ RT_STR_TUPLE("-- ") },
405	{ RT_STR_TUPLE("' ") },
406	{ RT_STR_TUPLE("<end>") },
407	};
408
409	/** Copyright+license empty line for each SCMCOMMENTSTYLE. */
410	static RTSTRTUPLE const g_aCopyrightCommentEmpty[] =
411	{
412	{ RT_STR_TUPLE("<invalid>") },
413	{ RT_STR_TUPLE(" *") },
414	{ RT_STR_TUPLE("#") },
415	{ RT_STR_TUPLE("") },
416	{ RT_STR_TUPLE(";") },
417	{ RT_STR_TUPLE("REM") },
418	{ RT_STR_TUPLE("rem") },
419	{ RT_STR_TUPLE("Rem") },
420	{ RT_STR_TUPLE("--") },
421	{ RT_STR_TUPLE("'") },
422	{ RT_STR_TUPLE("<end>") },
423	};
424
425	/** Copyright+license end of comment for each SCMCOMMENTSTYLE. */
426	static RTSTRTUPLE const g_aCopyrightCommentEnd[] =
427	{
428	{ RT_STR_TUPLE("<invalid> ") },
429	{ RT_STR_TUPLE(" */") },
430	{ RT_STR_TUPLE("#") },
431	{ RT_STR_TUPLE("\"\"\"") },
432	{ RT_STR_TUPLE(";") },
433	{ RT_STR_TUPLE("REM") },
434	{ RT_STR_TUPLE("rem") },
435	{ RT_STR_TUPLE("Rem") },
436	{ RT_STR_TUPLE("--") },
437	{ RT_STR_TUPLE("'") },
438	{ RT_STR_TUPLE("<end>") },
439	};
440
441
442	/**
443	* Figures out the predominant casing of the "REM" keyword in a batch file.
444	*
445	* @returns Predominant comment style.
446	* @param pIn The file to scan. Will be rewound.
447	*/
448	static SCMCOMMENTSTYLE determinBatchFileCommentStyle(PSCMSTREAM pIn)
449	{
450	/*
451	* Figure out whether it's using upper or lower case REM comments before
452	* doing the work.
453	*/
454	uint32_t cUpper = 0;
455	uint32_t cLower = 0;
456	uint32_t cCamel = 0;
457	SCMEOL enmEol;
458	size_t cchLine;
459	const char *pchLine;
460	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
461	{
462	while ( cchLine > 2
463	&& RT_C_IS_SPACE(*pchLine))
464	{
465	pchLine++;
466	cchLine--;
467	}
468	if ( ( cchLine > 3
469	&& RT_C_IS_SPACE(pchLine[2]))
470	\|\| cchLine == 3)
471	{
472	if ( pchLine[0] == 'R'
473	&& pchLine[1] == 'E'
474	&& pchLine[2] == 'M')
475	cUpper++;
476	else if ( pchLine[0] == 'r'
477	&& pchLine[1] == 'e'
478	&& pchLine[2] == 'm')
479	cLower++;
480	else if ( pchLine[0] == 'R'
481	&& pchLine[1] == 'e'
482	&& pchLine[2] == 'm')
483	cCamel++;
484	}
485	}
486
487	ScmStreamRewindForReading(pIn);
488
489	if (cLower >= cUpper && cLower >= cCamel)
490	return kScmCommentStyle_Rem_Lower;
491	if (cCamel >= cLower && cCamel >= cUpper)
492	return kScmCommentStyle_Rem_Camel;
493	return kScmCommentStyle_Rem_Upper;
494	}
495
496
497	/**
498	* Worker for isBlankLine.
499	*
500	* @returns true if blank, false if not.
501	* @param pchLine Pointer to the start of the line.
502	* @param cchLine The (encoded) length of the line, excluding EOL char.
503	*/
504	static bool isBlankLineSlow(const char *pchLine, size_t cchLine)
505	{
506	/*
507	* From the end, more likely to hit a non-blank char there.
508	*/
509	while (cchLine-- > 0)
510	if (!RT_C_IS_BLANK(pchLine[cchLine]))
511	return false;
512	return true;
513	}
514
515	/**
516	* Helper for checking whether a line is blank.
517	*
518	* @returns true if blank, false if not.
519	* @param pchLine Pointer to the start of the line.
520	* @param cchLine The (encoded) length of the line, excluding EOL char.
521	*/
522	DECLINLINE(bool) isBlankLine(const char *pchLine, size_t cchLine)
523	{
524	if (cchLine == 0)
525	return true;
526	/*
527	* We're more likely to fine a non-space char at the end of the line than
528	* at the start, due to source code indentation.
529	*/
530	if (pchLine[cchLine - 1])
531	return false;
532
533	/*
534	* Don't bother inlining loop code.
535	*/
536	return isBlankLineSlow(pchLine, cchLine);
537	}
538
539
540	/**
541	* Checks if there are @a cch blanks at @a pch.
542	*
543	* @returns true if span of @a cch blanks, false if not.
544	* @param pch The start of the span to check.
545	* @param cch The length of the span.
546	*/
547	DECLINLINE(bool) isSpanOfBlanks(const char *pch, size_t cch)
548	{
549	while (cch-- > 0)
550	{
551	char const ch = *pch++;
552	if (!RT_C_IS_BLANK(ch))
553	return false;
554	}
555	return true;
556	}
557
558
559	/**
560	* Strip trailing blanks (space & tab).
561	*
562	* @returns True if modified, false if not.
563	* @param pIn The input stream.
564	* @param pOut The output stream.
565	* @param pSettings The settings.
566	*/
567	bool rewrite_StripTrailingBlanks(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
568	{
569	if (!pSettings->fStripTrailingBlanks)
570	return false;
571
572	bool fModified = false;
573	SCMEOL enmEol;
574	size_t cchLine;
575	const char *pchLine;
576	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
577	{
578	int rc;
579	if ( cchLine == 0
580	\|\| !RT_C_IS_BLANK(pchLine[cchLine - 1]) )
581	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
582	else
583	{
584	cchLine--;
585	while (cchLine > 0 && RT_C_IS_BLANK(pchLine[cchLine - 1]))
586	cchLine--;
587	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
588	fModified = true;
589	}
590	if (RT_FAILURE(rc))
591	return false;
592	}
593	if (fModified)
594	ScmVerbose(pState, 2, " * Stripped trailing blanks\n");
595	return fModified;
596	}
597
598	/**
599	* Expand tabs.
600	*
601	* @returns True if modified, false if not.
602	* @param pIn The input stream.
603	* @param pOut The output stream.
604	* @param pSettings The settings.
605	*/
606	bool rewrite_ExpandTabs(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
607	{
608	if (!pSettings->fConvertTabs)
609	return false;
610
611	size_t const cchTab = pSettings->cchTab;
612	bool fModified = false;
613	SCMEOL enmEol;
614	size_t cchLine;
615	const char *pchLine;
616	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
617	{
618	int rc;
619	const char pchTab = (const char )memchr(pchLine, '\t', cchLine);
620	if (!pchTab)
621	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
622	else
623	{
624	size_t offTab = 0;
625	const char *pchChunk = pchLine;
626	for (;;)
627	{
628	size_t cchChunk = pchTab - pchChunk;
629	offTab += cchChunk;
630	ScmStreamWrite(pOut, pchChunk, cchChunk);
631
632	size_t cchToTab = cchTab - offTab % cchTab;
633	ScmStreamWrite(pOut, g_szTabSpaces, cchToTab);
634	offTab += cchToTab;
635
636	pchChunk = pchTab + 1;
637	size_t cchLeft = cchLine - (pchChunk - pchLine);
638	pchTab = (const char *)memchr(pchChunk, '\t', cchLeft);
639	if (!pchTab)
640	{
641	rc = ScmStreamPutLine(pOut, pchChunk, cchLeft, enmEol);
642	break;
643	}
644	}
645
646	fModified = true;
647	}
648	if (RT_FAILURE(rc))
649	return false;
650	}
651	if (fModified)
652	ScmVerbose(pState, 2, " * Expanded tabs\n");
653	return fModified;
654	}
655
656	/**
657	* Worker for rewrite_ForceNativeEol, rewrite_ForceLF and rewrite_ForceCRLF.
658	*
659	* @returns true if modifications were made, false if not.
660	* @param pIn The input stream.
661	* @param pOut The output stream.
662	* @param pSettings The settings.
663	* @param enmDesiredEol The desired end of line indicator type.
664	* @param pszDesiredSvnEol The desired svn:eol-style.
665	*/
666	static bool rewrite_ForceEol(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings,
667	SCMEOL enmDesiredEol, const char *pszDesiredSvnEol)
668	{
669	if (!pSettings->fConvertEol)
670	return false;
671
672	bool fModified = false;
673	SCMEOL enmEol;
674	size_t cchLine;
675	const char *pchLine;
676	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
677	{
678	if ( enmEol != enmDesiredEol
679	&& enmEol != SCMEOL_NONE)
680	{
681	fModified = true;
682	enmEol = enmDesiredEol;
683	}
684	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
685	if (RT_FAILURE(rc))
686	return false;
687	}
688	if (fModified)
689	ScmVerbose(pState, 2, " * Converted EOL markers\n");
690
691	/* Check svn:eol-style if appropriate */
692	if ( pSettings->fSetSvnEol
693	&& ScmSvnIsInWorkingCopy(pState))
694	{
695	char *pszEol;
696	int rc = ScmSvnQueryProperty(pState, "svn:eol-style", &pszEol);
697	if ( (RT_SUCCESS(rc) && strcmp(pszEol, pszDesiredSvnEol))
698	\|\| rc == VERR_NOT_FOUND)
699	{
700	if (rc == VERR_NOT_FOUND)
701	ScmVerbose(pState, 2, " * Setting svn:eol-style to %s (missing)\n", pszDesiredSvnEol);
702	else
703	ScmVerbose(pState, 2, " * Setting svn:eol-style to %s (was: %s)\n", pszDesiredSvnEol, pszEol);
704	int rc2 = ScmSvnSetProperty(pState, "svn:eol-style", pszDesiredSvnEol);
705	if (RT_FAILURE(rc2))
706	ScmError(pState, rc2, "ScmSvnSetProperty: %Rrc\n", rc2);
707	}
708	if (RT_SUCCESS(rc))
709	RTStrFree(pszEol);
710	}
711
712	/** @todo also check the subversion svn:eol-style state! */
713	return fModified;
714	}
715
716	/**
717	* Force native end of line indicator.
718	*
719	* @returns true if modifications were made, false if not.
720	* @param pIn The input stream.
721	* @param pOut The output stream.
722	* @param pSettings The settings.
723	*/
724	bool rewrite_ForceNativeEol(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
725	{
726	#if defined(RT_OS_WINDOWS) \|\| defined(RT_OS_OS2)
727	return rewrite_ForceEol(pState, pIn, pOut, pSettings, SCMEOL_CRLF, "native");
728	#else
729	return rewrite_ForceEol(pState, pIn, pOut, pSettings, SCMEOL_LF, "native");
730	#endif
731	}
732
733	/**
734	* Force the stream to use LF as the end of line indicator.
735	*
736	* @returns true if modifications were made, false if not.
737	* @param pIn The input stream.
738	* @param pOut The output stream.
739	* @param pSettings The settings.
740	*/
741	bool rewrite_ForceLF(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
742	{
743	return rewrite_ForceEol(pState, pIn, pOut, pSettings, SCMEOL_LF, "LF");
744	}
745
746	/**
747	* Force the stream to use CRLF as the end of line indicator.
748	*
749	* @returns true if modifications were made, false if not.
750	* @param pIn The input stream.
751	* @param pOut The output stream.
752	* @param pSettings The settings.
753	*/
754	bool rewrite_ForceCRLF(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
755	{
756	return rewrite_ForceEol(pState, pIn, pOut, pSettings, SCMEOL_CRLF, "CRLF");
757	}
758
759	/**
760	* Strip trailing blank lines and/or make sure there is exactly one blank line
761	* at the end of the file.
762	*
763	* @returns true if modifications were made, false if not.
764	* @param pIn The input stream.
765	* @param pOut The output stream.
766	* @param pSettings The settings.
767	*
768	* @remarks ASSUMES trailing white space has been removed already.
769	*/
770	bool rewrite_AdjustTrailingLines(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
771	{
772	if ( !pSettings->fStripTrailingLines
773	&& !pSettings->fForceTrailingLine
774	&& !pSettings->fForceFinalEol)
775	return false;
776
777	size_t const cLines = ScmStreamCountLines(pIn);
778
779	/* Empty files remains empty. */
780	if (cLines <= 1)
781	return false;
782
783	/* Figure out if we need to adjust the number of lines or not. */
784	size_t cLinesNew = cLines;
785
786	if ( pSettings->fStripTrailingLines
787	&& ScmStreamIsWhiteLine(pIn, cLinesNew - 1))
788	{
789	while ( cLinesNew > 1
790	&& ScmStreamIsWhiteLine(pIn, cLinesNew - 2))
791	cLinesNew--;
792	}
793
794	if ( pSettings->fForceTrailingLine
795	&& !ScmStreamIsWhiteLine(pIn, cLinesNew - 1))
796	cLinesNew++;
797
798	bool fFixMissingEol = pSettings->fForceFinalEol
799	&& ScmStreamGetEolByLine(pIn, cLinesNew - 1) == SCMEOL_NONE;
800
801	if ( !fFixMissingEol
802	&& cLines == cLinesNew)
803	return false;
804
805	/* Copy the number of lines we've arrived at. */
806	ScmStreamRewindForReading(pIn);
807
808	size_t cCopied = RT_MIN(cLinesNew, cLines);
809	ScmStreamCopyLines(pOut, pIn, cCopied);
810
811	if (cCopied != cLinesNew)
812	{
813	while (cCopied++ < cLinesNew)
814	ScmStreamPutLine(pOut, "", 0, ScmStreamGetEol(pIn));
815	}
816	/* Fix missing EOL if required. */
817	else if (fFixMissingEol)
818	{
819	if (ScmStreamGetEol(pIn) == SCMEOL_LF)
820	ScmStreamWrite(pOut, "\n", 1);
821	else
822	ScmStreamWrite(pOut, "\r\n", 2);
823	}
824
825	ScmVerbose(pState, 2, " * Adjusted trailing blank lines\n");
826	return true;
827	}
828
829	/**
830	* Make sure there is no svn:executable property on the current file.
831	*
832	* @returns false - the state carries these kinds of changes.
833	* @param pState The rewriter state.
834	* @param pIn The input stream.
835	* @param pOut The output stream.
836	* @param pSettings The settings.
837	*/
838	bool rewrite_SvnNoExecutable(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
839	{
840	RT_NOREF2(pIn, pOut);
841	if ( !pSettings->fSetSvnExecutable
842	\|\| !ScmSvnIsInWorkingCopy(pState))
843	return false;
844
845	int rc = ScmSvnQueryProperty(pState, "svn:executable", NULL);
846	if (RT_SUCCESS(rc))
847	{
848	ScmVerbose(pState, 2, " * removing svn:executable\n");
849	rc = ScmSvnDelProperty(pState, "svn:executable");
850	if (RT_FAILURE(rc))
851	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
852	}
853	return false;
854	}
855
856	/**
857	* Make sure there is no svn:keywords property on the current file.
858	*
859	* @returns false - the state carries these kinds of changes.
860	* @param pState The rewriter state.
861	* @param pIn The input stream.
862	* @param pOut The output stream.
863	* @param pSettings The settings.
864	*/
865	bool rewrite_SvnNoKeywords(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
866	{
867	RT_NOREF2(pIn, pOut);
868	if ( !pSettings->fSetSvnExecutable
869	\|\| !ScmSvnIsInWorkingCopy(pState))
870	return false;
871
872	int rc = ScmSvnQueryProperty(pState, "svn:keywords", NULL);
873	if (RT_SUCCESS(rc))
874	{
875	ScmVerbose(pState, 2, " * removing svn:keywords\n");
876	rc = ScmSvnDelProperty(pState, "svn:keywords");
877	if (RT_FAILURE(rc))
878	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
879	}
880	return false;
881	}
882
883	/**
884	* Make sure there is no svn:eol-style property on the current file.
885	*
886	* @returns false - the state carries these kinds of changes.
887	* @param pState The rewriter state.
888	* @param pIn The input stream.
889	* @param pOut The output stream.
890	* @param pSettings The settings.
891	*/
892	bool rewrite_SvnNoEolStyle(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
893	{
894	RT_NOREF2(pIn, pOut);
895	if ( !pSettings->fSetSvnExecutable
896	\|\| !ScmSvnIsInWorkingCopy(pState))
897	return false;
898
899	int rc = ScmSvnQueryProperty(pState, "svn:eol-style", NULL);
900	if (RT_SUCCESS(rc))
901	{
902	ScmVerbose(pState, 2, " * removing svn:eol-style\n");
903	rc = ScmSvnDelProperty(pState, "svn:eol-style");
904	if (RT_FAILURE(rc))
905	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
906	}
907	return false;
908	}
909
910	/**
911	* Makes sure the svn properties are appropriate for a binary.
912	*
913	* @returns false - the state carries these kinds of changes.
914	* @param pState The rewriter state.
915	* @param pIn The input stream.
916	* @param pOut The output stream.
917	* @param pSettings The settings.
918	*/
919	bool rewrite_SvnBinary(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
920	{
921	RT_NOREF2(pIn, pOut);
922	if ( !pSettings->fSetSvnExecutable
923	\|\| !ScmSvnIsInWorkingCopy(pState))
924	return false;
925
926	/* remove svn:eol-style and svn:keywords */
927	static const char * const s_apszRemove[] = { "svn:eol-style", "svn:keywords" };
928	for (uint32_t i = 0; i < RT_ELEMENTS(s_apszRemove); i++)
929	{
930	char *pszValue;
931	int rc = ScmSvnQueryProperty(pState, s_apszRemove[i], &pszValue);
932	if (RT_SUCCESS(rc))
933	{
934	ScmVerbose(pState, 2, " * removing %s=%s\n", s_apszRemove[i], pszValue);
935	RTStrFree(pszValue);
936	rc = ScmSvnDelProperty(pState, s_apszRemove[i]);
937	if (RT_FAILURE(rc))
938	ScmError(pState, rc, "ScmSvnSetProperty(,%s): %Rrc\n", s_apszRemove[i], rc);
939	}
940	else if (rc != VERR_NOT_FOUND)
941	ScmError(pState, rc, "ScmSvnQueryProperty: %Rrc\n", rc);
942	}
943
944	/* Make sure there is a svn:mime-type set. */
945	int rc = ScmSvnQueryProperty(pState, "svn:mime-type", NULL);
946	if (rc == VERR_NOT_FOUND)
947	{
948	ScmVerbose(pState, 2, " * settings svn:mime-type\n");
949	rc = ScmSvnSetProperty(pState, "svn:mime-type", "application/octet-stream");
950	if (RT_FAILURE(rc))
951	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
952	}
953	else if (RT_FAILURE(rc))
954	ScmError(pState, rc, "ScmSvnQueryProperty: %Rrc\n", rc);
955
956	return false;
957	}
958
959	/**
960	* Make sure the Id and Revision keywords are expanded.
961	*
962	* @returns false - the state carries these kinds of changes.
963	* @param pState The rewriter state.
964	* @param pIn The input stream.
965	* @param pOut The output stream.
966	* @param pSettings The settings.
967	*/
968	bool rewrite_SvnKeywords(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
969	{
970	RT_NOREF2(pIn, pOut);
971	if ( !pSettings->fSetSvnKeywords
972	\|\| !ScmSvnIsInWorkingCopy(pState))
973	return false;
974
975	char *pszKeywords;
976	int rc = ScmSvnQueryProperty(pState, "svn:keywords", &pszKeywords);
977	if ( RT_SUCCESS(rc)
978	&& ( !strstr(pszKeywords, "Id") /** @todo need some function for finding a word in a string. */
979	\|\| !strstr(pszKeywords, "Revision")) )
980	{
981	if (!strstr(pszKeywords, "Id") && !strstr(pszKeywords, "Revision"))
982	rc = RTStrAAppend(&pszKeywords, " Id Revision");
983	else if (!strstr(pszKeywords, "Id"))
984	rc = RTStrAAppend(&pszKeywords, " Id");
985	else
986	rc = RTStrAAppend(&pszKeywords, " Revision");
987	if (RT_SUCCESS(rc))
988	{
989	ScmVerbose(pState, 2, " * changing svn:keywords to '%s'\n", pszKeywords);
990	rc = ScmSvnSetProperty(pState, "svn:keywords", pszKeywords);
991	if (RT_FAILURE(rc))
992	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
993	}
994	else
995	ScmError(pState, rc, "RTStrAppend: %Rrc\n", rc);
996	RTStrFree(pszKeywords);
997	}
998	else if (rc == VERR_NOT_FOUND)
999	{
1000	ScmVerbose(pState, 2, " * setting svn:keywords to 'Id Revision'\n");
1001	rc = ScmSvnSetProperty(pState, "svn:keywords", "Id Revision");
1002	if (RT_FAILURE(rc))
1003	ScmError(pState, rc, "ScmSvnSetProperty: %Rrc\n", rc);
1004	}
1005	else if (RT_SUCCESS(rc))
1006	RTStrFree(pszKeywords);
1007
1008	return false;
1009	}
1010
1011	/**
1012	* Checks the svn:sync-process value and that parent is exported too.
1013	*
1014	* @returns false - the state carries these kinds of changes.
1015	* @param pState The rewriter state.
1016	* @param pIn The input stream.
1017	* @param pOut The output stream.
1018	* @param pSettings The settings.
1019	*/
1020	bool rewrite_SvnSyncProcess(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
1021	{
1022	RT_NOREF2(pIn, pOut);
1023	if ( pSettings->fSkipSvnSyncProcess
1024	\|\| !ScmSvnIsInWorkingCopy(pState))
1025	return false;
1026
1027	char *pszSyncProcess;
1028	int rc = ScmSvnQueryProperty(pState, "svn:sync-process", &pszSyncProcess);
1029	if (RT_SUCCESS(rc))
1030	{
1031	if (strcmp(pszSyncProcess, "export") == 0)
1032	{
1033	char *pszParentSyncProcess;
1034	rc = ScmSvnQueryParentProperty(pState, "svn:sync-process", &pszParentSyncProcess);
1035	if (RT_SUCCESS(rc))
1036	{
1037	if (strcmp(pszSyncProcess, "export") != 0)
1038	ScmError(pState, VERR_INVALID_STATE,
1039	"svn:sync-process=export, but parent directory differs: %s\n"
1040	"WARNING! Make sure to unexport everything inside the directory first!\n"
1041	" Then you may export the directory and stuff inside it if you want.\n"
1042	" (Just exporting the directory will not make anything inside it externally visible.)\n"
1043	, pszParentSyncProcess);
1044	RTStrFree(pszParentSyncProcess);
1045	}
1046	else if (rc == VERR_NOT_FOUND)
1047	ScmError(pState, VERR_NOT_FOUND,
1048	"svn:sync-process=export, but parent directory is not exported!\n"
1049	"WARNING! Make sure to unexport everything inside the directory first!\n"
1050	" Then you may export the directory and stuff inside it if you want.\n"
1051	" (Just exporting the directory will not make anything inside it externally visible.)\n");
1052	else
1053	ScmError(pState, rc, "ScmSvnQueryParentProperty: %Rrc\n", rc);
1054	}
1055	else if (strcmp(pszSyncProcess, "ignore") != 0)
1056	ScmError(pState, VERR_INVALID_NAME, "Bad sync-process value: %s\n", pszSyncProcess);
1057	RTStrFree(pszSyncProcess);
1058	}
1059	else if (rc != VERR_NOT_FOUND)
1060	ScmError(pState, rc, "ScmSvnQueryProperty: %Rrc\n", rc);
1061
1062	return false;
1063	}
1064
1065	/**
1066	* Checks the that there is no bidirectional unicode fun in the file.
1067	*
1068	* @returns false - the state carries these kinds of changes.
1069	* @param pState The rewriter state.
1070	* @param pIn The input stream.
1071	* @param pOut The output stream.
1072	* @param pSettings The settings.
1073	*/
1074	bool rewrite_UnicodeChecks(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
1075	{
1076	RT_NOREF2(pIn, pOut);
1077	if (pSettings->fSkipUnicodeChecks)
1078	return false;
1079
1080	/*
1081	* Just scan the input for weird stuff and fail if we find anything we don't like.
1082	*/
1083	uint32_t iLine = 0;
1084	SCMEOL enmEol;
1085	size_t cchLine;
1086	const char *pchLine;
1087	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
1088	{
1089	iLine++;
1090	const char *pchCur = pchLine;
1091	size_t cchLeft = cchLine;
1092	while (cchLeft > 0)
1093	{
1094	RTUNICP uc = 0;
1095	int rc = RTStrGetCpNEx(&pchCur, &cchLeft, &uc);
1096	if (RT_SUCCESS(rc))
1097	{
1098	const char *pszWhat;
1099	switch (uc)
1100	{
1101	default:
1102	continue;
1103
1104	/* Potentially evil bi-directional control codes (Table I, trojan-source.pdf): */
1105	case 0x202a: pszWhat = "LRE - left-to-right embedding"; break;
1106	case 0x202b: pszWhat = "RLE - right-to-left embedding"; break;
1107	case 0x202d: pszWhat = "LRO - left-to-right override"; break;
1108	case 0x202e: pszWhat = "RLO - right-to-left override"; break;
1109	case 0x2066: pszWhat = "LRI - left-to-right isolate"; break;
1110	case 0x2067: pszWhat = "RLI - right-to-left isolate"; break;
1111	case 0x2068: pszWhat = "FSI - first strong isolate"; break;
1112	case 0x202c: pszWhat = "PDF - pop directional formatting (LRE, RLE, LRO, RLO)"; break;
1113	case 0x2069: pszWhat = "PDI - pop directional isolate (LRI, RLI)"; break;
1114
1115	/** @todo add checks for homoglyphs too. */
1116	}
1117	ScmFixManually(pState, "%u:%zu: Evil unicode codepoint: %s\n", iLine, pchCur - pchLine, pszWhat);
1118	}
1119	else
1120	ScmFixManually(pState, "%u:%zu: Invalid UTF-8 encoding: %Rrc\n", iLine, pchCur - pchLine, rc);
1121	}
1122	}
1123
1124	return false;
1125	}
1126
1127	/**
1128	* Compares two strings word-by-word, ignoring spaces, punctuation and case.
1129	*
1130	* Assumes ASCII strings.
1131	*
1132	* @returns true if they match, false if not.
1133	* @param psz1 The first string. This is typically the known one.
1134	* @param psz2 The second string. This is typically the unknown one,
1135	* which is why we return a next pointer for this one.
1136	* @param ppsz2Next Where to return the next part of the 2nd string. If
1137	* this is NULL, the whole string must match.
1138	*/
1139	static bool IsEqualWordByWordIgnoreCase(const char psz1, const char psz2, const char **ppsz2Next)
1140	{
1141	for (;;)
1142	{
1143	/* Try compare raw strings first. */
1144	char ch1 = *psz1;
1145	char ch2 = *psz2;
1146	if ( ch1 == ch2
1147	\|\| RT_C_TO_LOWER(ch1) == RT_C_TO_LOWER(ch2))
1148	{
1149	if (ch1)
1150	{
1151	psz1++;
1152	psz2++;
1153	}
1154	else
1155	{
1156	if (ppsz2Next)
1157	*ppsz2Next = psz2;
1158	return true;
1159	}
1160	}
1161	else
1162	{
1163	/* Try skip spaces an punctuation. */
1164	while ( RT_C_IS_SPACE(ch1)
1165	\|\| RT_C_IS_PUNCT(ch1))
1166	ch1 = *++psz1;
1167
1168	if (ch1 == '\0' && ppsz2Next)
1169	{
1170	*ppsz2Next = psz2;
1171	return true;
1172	}
1173
1174	while ( RT_C_IS_SPACE(ch2)
1175	\|\| RT_C_IS_PUNCT(ch2))
1176	ch2 = *++psz2;
1177
1178	if ( ch1 != ch2
1179	&& RT_C_TO_LOWER(ch1) != RT_C_TO_LOWER(ch2))
1180	{
1181	if (ppsz2Next)
1182	*ppsz2Next = psz2;
1183	return false;
1184	}
1185	}
1186	}
1187	}
1188
1189	/**
1190	* Looks for @a pszFragment anywhere in @a pszText, ignoring spaces, punctuation
1191	* and case.
1192	*
1193	* @returns true if found, false if not.
1194	* @param pszText The haystack to search in.
1195	* @param cchText The length @a pszText.
1196	* @param pszFragment The needle to search for.
1197	* @param ppszStart Where to return the address in @a pszText where
1198	* the fragment was found. Optional.
1199	* @param ppszNext Where to return the pointer to the first char in
1200	* @a pszText after the fragment. Optional.
1201	*
1202	* @remarks First character of @a pszFragment must be an 7-bit ASCII character!
1203	* This character must not be space or punctuation.
1204	*/
1205	static bool scmContainsWordByWordIgnoreCase(const char pszText, size_t cchText, const char pszFragment,
1206	const char ppszStart, const char ppszNext)
1207	{
1208	Assert(!((unsigned)*pszFragment & 0x80));
1209	Assert(pszText[cchText] == '\0');
1210	Assert(!RT_C_IS_BLANK(*pszFragment));
1211	Assert(!RT_C_IS_PUNCT(*pszFragment));
1212
1213	char chLower = RT_C_TO_LOWER(*pszFragment);
1214	char chUpper = RT_C_TO_UPPER(*pszFragment);
1215	for (;;)
1216	{
1217	const char pszHit = (const char )memchr(pszText, chLower, cchText);
1218	const char pszHit2 = (const char )memchr(pszText, chUpper, cchText);
1219	if (!pszHit && !pszHit2)
1220	{
1221	if (ppszStart)
1222	*ppszStart = NULL;
1223	if (ppszNext)
1224	*ppszNext = NULL;
1225	return false;
1226	}
1227
1228	if ( pszHit == NULL
1229	\|\| ( pszHit2 != NULL
1230	&& ((uintptr_t)pszHit2 < (uintptr_t)pszHit)) )
1231	pszHit = pszHit2;
1232
1233	const char *pszNext;
1234	if (IsEqualWordByWordIgnoreCase(pszFragment, pszHit, &pszNext))
1235	{
1236	if (ppszStart)
1237	*ppszStart = pszHit;
1238	if (ppszNext)
1239	*ppszNext = pszNext;
1240	return true;
1241	}
1242
1243	cchText -= pszHit - pszText + 1;
1244	pszText = pszHit + 1;
1245	}
1246	}
1247
1248
1249	/**
1250	* Counts the number of lines in the given substring.
1251	*
1252	* @returns The number of lines.
1253	* @param psz The start of the substring.
1254	* @param cch The length of the substring.
1255	*/
1256	static uint32_t CountLinesInSubstring(const char *psz, size_t cch)
1257	{
1258	uint32_t cLines = 0;
1259	for (;;)
1260	{
1261	const char pszEol = (const char )memchr(psz, '\n', cch);
1262	if (pszEol)
1263	cLines++;
1264	else
1265	return cLines + (*psz != '\0');
1266	cch -= pszEol + 1 - psz;
1267	if (!cch)
1268	return cLines;
1269	psz = pszEol + 1;
1270	}
1271	}
1272
1273
1274	/**
1275	* Comment parser callback for locating copyright and license.
1276	*/
1277	static DECLCALLBACK(int)
1278	rewrite_Copyright_CommentCallback(PCSCMCOMMENTINFO pInfo, const char pszBody, size_t cchBody, void pvUser)
1279	{
1280	PSCMCOPYRIGHTINFO pState = (PSCMCOPYRIGHTINFO)pvUser;
1281	Assert(strlen(pszBody) == cchBody);
1282	//RTPrintf("--- comment at %u, type %u ---\n%s\n--- end ---\n", pInfo->iLineStart, pInfo->enmType, pszBody);
1283	ScmVerbose(pState->pState, 5,
1284	"--- comment at %u col %u, %u lines, type %u, %u lines before body, %u lines after body\n",
1285	pInfo->iLineStart, pInfo->offStart, pInfo->iLineEnd - pInfo->iLineStart + 1, pInfo->enmType,
1286	pInfo->cBlankLinesBefore, pInfo->cBlankLinesAfter);
1287
1288	pState->cComments++;
1289
1290	uint32_t iLine = pInfo->iLineStart + pInfo->cBlankLinesBefore;
1291
1292	/*
1293	* Look for a 'contributed by' or 'includes contributions from' line, these
1294	* comes first when present.
1295	*/
1296	const char *pchContributedBy = NULL;
1297	size_t cchContributedBy = 0;
1298	size_t cBlankLinesAfterContributedBy = 0;
1299	if ( pState->pszContributedBy == NULL
1300	&& ( pState->iLineCopyright == UINT32_MAX
1301	\|\| pState->iLineLicense == UINT32_MAX)
1302	&& ( ( cchBody > sizeof("Contributed by")
1303	&& RTStrNICmp(pszBody, RT_STR_TUPLE("contributed by")) == 0)
1304	\|\| ( cchBody > sizeof("Includes contributions from")
1305	&& RTStrNICmp(pszBody, RT_STR_TUPLE("Includes contributions from")) == 0) ) )
1306	{
1307	const char pszNextLine = (const char )memchr(pszBody, '\n', cchBody);
1308	while (pszNextLine && pszNextLine[1] != '\n')
1309	pszNextLine = (const char *)memchr(pszNextLine + 1, '\n', cchBody);
1310	if (pszNextLine)
1311	{
1312	pchContributedBy = pszBody;
1313	cchContributedBy = pszNextLine - pszBody;
1314
1315	/* Skip the copyright line and any blank lines following it. */
1316	cchBody -= cchContributedBy + 1;
1317	pszBody = pszNextLine + 1;
1318	iLine += 1;
1319	while (*pszBody == '\n')
1320	{
1321	pszBody++;
1322	cchBody--;
1323	iLine++;
1324	cBlankLinesAfterContributedBy++;
1325	}
1326	}
1327	}
1328
1329	/*
1330	* Look for the copyright line.
1331	*/
1332	bool fFoundCopyright = false;
1333	uint32_t cBlankLinesAfterCopyright = 0;
1334	if ( pState->iLineCopyright == UINT32_MAX
1335	&& cchBody > sizeof("Copyright") + sizeof(g_szCopyrightHolder)
1336	&& RTStrNICmp(pszBody, RT_STR_TUPLE("copyright")) == 0)
1337	{
1338	const char pszNextLine = (const char )memchr(pszBody, '\n', cchBody);
1339
1340	/* Oracle copyright? */
1341	const char *pszEnd = pszNextLine ? pszNextLine : &pszBody[cchBody];
1342	while (RT_C_IS_SPACE(pszEnd[-1]))
1343	pszEnd--;
1344	if ( (uintptr_t)(pszEnd - pszBody) > sizeof(g_szCopyrightHolder)
1345	&& ((unsigned char )(pszEnd - sizeof(g_szCopyrightHolder) + 1) & 0x80) == 0 /* to avoid annoying assertion */
1346	&& RTStrNICmp(pszEnd - sizeof(g_szCopyrightHolder) + 1, RT_STR_TUPLE(g_szCopyrightHolder)) == 0)
1347	{
1348	/* Parse out the year(s). */
1349	const char *psz = pszBody + sizeof("copyright");
1350	while ((uintptr_t)psz < (uintptr_t)pszEnd && !RT_C_IS_DIGIT(*psz))
1351	psz++;
1352	if (RT_C_IS_DIGIT(*psz))
1353	{
1354	char *pszNext;
1355	int rc = RTStrToUInt32Ex(psz, &pszNext, 10, &pState->uFirstYear);
1356	if ( RT_SUCCESS(rc)
1357	&& rc != VWRN_NUMBER_TOO_BIG
1358	&& rc != VWRN_NEGATIVE_UNSIGNED)
1359	{
1360	if ( pState->uFirstYear < 1975
1361	\|\| pState->uFirstYear > 3000)
1362	{
1363	ScmError(pState->pState, VERR_OUT_OF_RANGE, "Copyright year is out of range: %u ('%.*s')\n",
1364	pState->uFirstYear, pszEnd - pszBody, pszBody);
1365	pState->uFirstYear = UINT32_MAX;
1366	}
1367
1368	while (RT_C_IS_SPACE(*pszNext))
1369	pszNext++;
1370	if (*pszNext == '-')
1371	{
1372	do
1373	pszNext++;
1374	while (RT_C_IS_SPACE(*pszNext));
1375	rc = RTStrToUInt32Ex(pszNext, &pszNext, 10, &pState->uLastYear);
1376	if ( RT_SUCCESS(rc)
1377	&& rc != VWRN_NUMBER_TOO_BIG
1378	&& rc != VWRN_NEGATIVE_UNSIGNED)
1379	{
1380	if ( pState->uLastYear < 1975
1381	\|\| pState->uLastYear > 3000)
1382	{
1383	ScmError(pState->pState, VERR_OUT_OF_RANGE, "Second copyright year is out of range: %u ('%.*s')\n",
1384	pState->uLastYear, pszEnd - pszBody, pszBody);
1385	pState->uLastYear = UINT32_MAX;
1386	}
1387	else if (pState->uFirstYear > pState->uLastYear)
1388	{
1389	RTMsgWarning("Copyright years switched(?): '%.*s'\n", pszEnd - pszBody, pszBody);
1390	uint32_t iTmp = pState->uLastYear;
1391	pState->uLastYear = pState->uFirstYear;
1392	pState->uFirstYear = iTmp;
1393	}
1394	}
1395	else
1396	{
1397	pState->uLastYear = UINT32_MAX;
1398	ScmError(pState->pState, RT_SUCCESS(rc) ? -rc : rc,
1399	"Failed to parse second copyright year: '%.*s'\n", pszEnd - pszBody, pszBody);
1400	}
1401	}
1402	else if (*pszNext != g_szCopyrightHolder[0])
1403	ScmError(pState->pState, VERR_PARSE_ERROR,
1404	"Failed to parse copyright: '%.*s'\n", pszEnd - pszBody, pszBody);
1405	else
1406	pState->uLastYear = pState->uFirstYear;
1407	}
1408	else
1409	{
1410	pState->uFirstYear = UINT32_MAX;
1411	ScmError(pState->pState, RT_SUCCESS(rc) ? -rc : rc,
1412	"Failed to parse copyright year: '%.*s'\n", pszEnd - pszBody, pszBody);
1413	}
1414	}
1415
1416	/* The copyright comment must come before the license. */
1417	if (pState->iLineLicense != UINT32_MAX)
1418	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright (line %u) must come before the license (line %u)!\n",
1419	iLine, pState->iLineLicense);
1420
1421	/* In C/C++ code, this must be a multiline comment. While in python it
1422	must be a */
1423	if (pState->enmCommentStyle == kScmCommentStyle_C && pInfo->enmType != kScmCommentType_MultiLine)
1424	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright must appear in a multiline comment (no doxygen stuff)\n");
1425	else if (pState->enmCommentStyle == kScmCommentStyle_Python && pInfo->enmType != kScmCommentType_DocString)
1426	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright must appear in a doc-string\n");
1427
1428	/* The copyright must be followed by the license. */
1429	if (!pszNextLine)
1430	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright should be followed by the license text!\n");
1431
1432	/* Quit if we've flagged a failure. */
1433	if (RT_FAILURE(pState->pState->rc))
1434	return VERR_CALLBACK_RETURN;
1435
1436	/* Check if it's well formed and up to date. */
1437	char szWellFormed[256];
1438	size_t cchWellFormed;
1439	if (pState->uFirstYear == pState->uLastYear)
1440	cchWellFormed = RTStrPrintf(szWellFormed, sizeof(szWellFormed), "Copyright (C) %u %s",
1441	pState->uFirstYear, g_szCopyrightHolder);
1442	else
1443	cchWellFormed = RTStrPrintf(szWellFormed, sizeof(szWellFormed), "Copyright (C) %u-%u %s",
1444	pState->uFirstYear, pState->uLastYear, g_szCopyrightHolder);
1445	pState->fUpToDateCopyright = pState->uLastYear == g_uYear;
1446	pState->iLineCopyright = iLine;
1447	pState->fWellFormedCopyright = cchWellFormed == (uintptr_t)(pszEnd - pszBody)
1448	&& memcmp(pszBody, szWellFormed, cchWellFormed) == 0;
1449	if (!pState->fWellFormedCopyright)
1450	ScmVerbose(pState->pState, 1, "* copyright isn't well formed\n");
1451
1452	/* If there wasn't exactly one blank line before the comment, trigger a rewrite. */
1453	if (pInfo->cBlankLinesBefore != 1)
1454	{
1455	ScmVerbose(pState->pState, 1, "* copyright comment is preceeded by %u blank lines instead of 1\n",
1456	pInfo->cBlankLinesBefore);
1457	pState->fWellFormedCopyright = false;
1458	}
1459
1460	/* If the comment doesn't start in column 1, trigger rewrite. */
1461	if (pInfo->offStart != 0)
1462	{
1463	ScmVerbose(pState->pState, 1, "* copyright comment starts in column %u instead of 1\n", pInfo->offStart + 1);
1464	pState->fWellFormedCopyright = false;
1465	/** @todo check that there isn't any code preceeding the comment. */
1466	}
1467
1468	if (pchContributedBy)
1469	{
1470	pState->pszContributedBy = RTStrDupN(pchContributedBy, cchContributedBy);
1471	if (cBlankLinesAfterContributedBy != 1)
1472	{
1473	ScmVerbose(pState->pState, 1, "* %u blank lines between contributed by and copyright, should be 1\n",
1474	cBlankLinesAfterContributedBy);
1475	pState->fWellFormedCopyright = false;
1476	}
1477	}
1478
1479	fFoundCopyright = true;
1480	ScmVerbose(pState->pState, 3, "oracle copyright %u-%u: up-to-date=%RTbool well-formed=%RTbool\n",
1481	pState->uFirstYear, pState->uLastYear, pState->fUpToDateCopyright, pState->fWellFormedCopyright);
1482	}
1483	else
1484	ScmVerbose(pState->pState, 3, "not oracle copyright: '%.*s'\n", pszEnd - pszBody, pszBody);
1485
1486	if (!pszNextLine)
1487	return VINF_SUCCESS;
1488
1489	/* Skip the copyright line and any blank lines following it. */
1490	cchBody -= pszNextLine - pszBody + 1;
1491	pszBody = pszNextLine + 1;
1492	iLine += 1;
1493	while (*pszBody == '\n')
1494	{
1495	pszBody++;
1496	cchBody--;
1497	iLine++;
1498	cBlankLinesAfterCopyright++;
1499	}
1500
1501	/*
1502	* If we have a based-on-mit scenario, check for the lead in now and
1503	* complain if not found.
1504	*/
1505	if ( fFoundCopyright
1506	&& pState->enmLicenceOpt == kScmLicense_BasedOnMit
1507	&& pState->iLineLicense == UINT32_MAX)
1508	{
1509	if (RTStrNICmp(pszBody, RT_STR_TUPLE("This file is based on ")) == 0)
1510	{
1511	/* Take down a comment area which goes up to 'this file is based on'.
1512	The license line and length isn't used but gets set to cover the current line. */
1513	pState->iLineComment = pInfo->iLineStart;
1514	pState->cLinesComment = iLine - pInfo->iLineStart;
1515	pState->iLineLicense = iLine;
1516	pState->cLinesLicense = 1;
1517	pState->fExternalLicense = true;
1518	pState->fIsCorrectLicense = true;
1519	pState->fWellFormedLicense = true;
1520
1521	/* Check if we've got a MIT a license here or not. */
1522	pState->pCurrentLicense = NULL;
1523	do
1524	{
1525	const char pszEol = (const char )memchr(pszBody, '\n', cchBody);
1526	if (!pszEol \|\| pszEol[1] == '\0')
1527	{
1528	pszBody += cchBody;
1529	cchBody = 0;
1530	break;
1531	}
1532	cchBody -= pszEol - pszBody + 1;
1533	pszBody = pszEol + 1;
1534	iLine++;
1535
1536	for (PCSCMLICENSETEXT pCur = pState->paLicenses; pCur->cch > 0; pCur++)
1537	{
1538	const char *pszNext;
1539	if ( pCur->cch <= cchBody + 32 /* (+ 32 since we don't compare spaces and punctuation) */
1540	&& IsEqualWordByWordIgnoreCase(pCur->psz, pszBody, &pszNext))
1541	{
1542	pState->pCurrentLicense = pCur;
1543	break;
1544	}
1545	}
1546	} while (!pState->pCurrentLicense);
1547	if (!pState->pCurrentLicense)
1548	ScmError(pState->pState, VERR_NOT_FOUND, "Could not find the based-on license!\n");
1549	else if (pState->pCurrentLicense->enmType != kScmLicenseType_Mit)
1550	ScmError(pState->pState, VERR_NOT_FOUND, "The based-on license is not MIT (%.32s...)\n",
1551	pState->pCurrentLicense->psz);
1552	}
1553	else
1554	ScmError(pState->pState, VERR_WRONG_ORDER, "Expected 'This file is based on ...' after our copyright!\n");
1555	return VINF_SUCCESS;
1556	}
1557	}
1558
1559	/*
1560	* Look for LGPL like text in the comment.
1561	*/
1562	if (pState->fCheckforLgpl && cchBody > 128)
1563	{
1564	/* We look for typical LGPL notices. */
1565	if (pState->iLineLgplNotice == UINT32_MAX)
1566	{
1567	static const char * const s_apszFragments[] =
1568	{
1569	"under the terms of the GNU Lesser General Public License",
1570	};
1571	for (unsigned i = 0; i < RT_ELEMENTS(s_apszFragments); i++)
1572	if (scmContainsWordByWordIgnoreCase(pszBody, cchBody, s_apszFragments[i], NULL, NULL))
1573	{
1574	pState->iLineLgplNotice = iLine;
1575	pState->iLineAfterLgplComment = pInfo->iLineEnd + 1;
1576	ScmVerbose(pState->pState, 3, "Found LGPL notice at %u\n", iLine);
1577	break;
1578	}
1579	}
1580
1581	if ( pState->iLineLgplDisclaimer == UINT32_MAX
1582	&& scmContainsWordByWordIgnoreCase(pszBody, cchBody, g_szLgplDisclaimer, NULL, NULL))
1583	{
1584	pState->iLineLgplDisclaimer = iLine;
1585	ScmVerbose(pState->pState, 3, "Found LGPL disclaimer at %u\n", iLine);
1586	}
1587	}
1588
1589	/*
1590	* Look for the license text.
1591	*/
1592	if (pState->iLineLicense == UINT32_MAX)
1593	{
1594	for (PCSCMLICENSETEXT pCur = pState->paLicenses; pCur->cch > 0; pCur++)
1595	{
1596	const char *pszNext;
1597	if ( pCur->cch <= cchBody + 32 /* (+ 32 since we don't compare spaces and punctuation) */
1598	&& IsEqualWordByWordIgnoreCase(pCur->psz, pszBody, &pszNext))
1599	{
1600	while ( RT_C_IS_SPACE(*pszNext)
1601	\|\| (RT_C_IS_PUNCT(pszNext) && pszNext != '-'))
1602	pszNext++;
1603
1604	uint32_t cDashes = 0;
1605	while (*pszNext == '-')
1606	cDashes++, pszNext++;
1607	bool fExternal = cDashes > 10;
1608
1609	if ( *pszNext == '\0'
1610	\|\| fExternal)
1611	{
1612	/* In C/C++ code, this must be a multiline comment. While in python it
1613	must be a */
1614	if (pState->enmCommentStyle == kScmCommentStyle_C && pInfo->enmType != kScmCommentType_MultiLine)
1615	ScmError(pState->pState, VERR_WRONG_ORDER, "License must appear in a multiline comment (no doxygen stuff)\n");
1616	else if (pState->enmCommentStyle == kScmCommentStyle_Python && pInfo->enmType != kScmCommentType_DocString)
1617	ScmError(pState->pState, VERR_WRONG_ORDER, "License must appear in a doc-string\n");
1618
1619	/* Quit if we've flagged a failure. */
1620	if (RT_FAILURE(pState->pState->rc))
1621	return VERR_CALLBACK_RETURN;
1622
1623	/* Record it. */
1624	pState->iLineLicense = iLine;
1625	pState->cLinesLicense = CountLinesInSubstring(pszBody, pszNext - pszBody) - fExternal;
1626	pState->pCurrentLicense = pCur;
1627	pState->fExternalLicense = fExternal;
1628	pState->fIsCorrectLicense = pState->fOpenSource
1629	? pCur == pState->pExpectedLicense
1630	: pCur->enmType == kScmLicenseType_Confidential;
1631	pState->fWellFormedLicense = memcmp(pszBody, pCur->psz, pCur->cch - 1) == 0;
1632	if (!pState->fWellFormedLicense)
1633	ScmVerbose(pState->pState, 1, "* license text isn't well-formed\n");
1634
1635	/* If there was more than one blank line between the copyright and the
1636	license text, extend the license text area and force a rewrite of it. */
1637	if (cBlankLinesAfterCopyright > 1)
1638	{
1639	ScmVerbose(pState->pState, 1, "* %u blank lines between copyright and license text, instead of 1\n",
1640	cBlankLinesAfterCopyright);
1641	pState->iLineLicense -= cBlankLinesAfterCopyright - 1;
1642	pState->cLinesLicense += cBlankLinesAfterCopyright - 1;
1643	pState->fWellFormedLicense = false;
1644	}
1645
1646	/* If there was more than one blank line after the license, trigger a rewrite. */
1647	if (!fExternal && pInfo->cBlankLinesAfter != 1)
1648	{
1649	ScmVerbose(pState->pState, 1, "* copyright comment is followed by %u blank lines instead of 1\n",
1650	pInfo->cBlankLinesAfter);
1651	pState->fWellFormedLicense = false;
1652	}
1653
1654	/** @todo Check that the last comment line doesn't have any code on it. */
1655	/** @todo Check that column 2 contains '' for C/C++ files. /
1656
1657	ScmVerbose(pState->pState, 3,
1658	"Found license %d/%d at %u..%u: is-correct=%RTbool well-formed=%RTbool external-part=%RTbool open-source=%RTbool\n",
1659	pCur->enmType, pCur->enmOpt, pState->iLineLicense, pState->iLineLicense + pState->cLinesLicense,
1660	pState->fIsCorrectLicense, pState->fWellFormedLicense,
1661	pState->fExternalLicense, pState->fOpenSource);
1662
1663	if (fFoundCopyright)
1664	{
1665	pState->iLineComment = pInfo->iLineStart;
1666	pState->cLinesComment = (fExternal ? pState->iLineLicense + pState->cLinesLicense : pInfo->iLineEnd + 1)
1667	- pInfo->iLineStart;
1668	}
1669	else
1670	ScmError(pState->pState, VERR_WRONG_ORDER, "License should be preceeded by the copyright!\n");
1671	break;
1672	}
1673	}
1674	}
1675	}
1676
1677	if (fFoundCopyright && pState->iLineLicense == UINT32_MAX)
1678	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright should be followed by the license text!\n");
1679
1680	/*
1681	* Stop looking for stuff after 100 comments.
1682	*/
1683	if (pState->cComments > 100)
1684	return VERR_CALLBACK_RETURN;
1685	return VINF_SUCCESS;
1686	}
1687
1688	/**
1689	* Writes comment body text.
1690	*
1691	* @returns Stream status.
1692	* @param pOut The output stream.
1693	* @param pszText The text to write.
1694	* @param cchText The length of the text.
1695	* @param enmCommentStyle The comment style.
1696	* @param enmEol The EOL style.
1697	*/
1698	static int scmWriteCommentBody(PSCMSTREAM pOut, const char *pszText, size_t cchText,
1699	SCMCOMMENTSTYLE enmCommentStyle, SCMEOL enmEol)
1700	{
1701	Assert(pszText[cchText - 1] == '\n');
1702	Assert(pszText[cchText - 2] != '\n');
1703	NOREF(cchText);
1704	do
1705	{
1706	const char *pszEol = strchr(pszText, '\n');
1707	if (pszEol != pszText)
1708	{
1709	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1710	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1711	ScmStreamWrite(pOut, pszText, pszEol - pszText);
1712	ScmStreamPutEol(pOut, enmEol);
1713	}
1714	else
1715	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1716	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1717	pszText = pszEol + 1;
1718	} while (*pszText != '\0');
1719	return ScmStreamGetStatus(pOut);
1720	}
1721
1722
1723	/**
1724	* Updates the copyright year and/or license text.
1725	*
1726	* @returns true if modifications were made, false if not.
1727	* @param pState The rewriter state.
1728	* @param pIn The input stream.
1729	* @param pOut The output stream.
1730	* @param pSettings The settings.
1731	* @param enmCommentStyle The comment style used by the file.
1732	*/
1733	static bool rewrite_Copyright_Common(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings,
1734	SCMCOMMENTSTYLE enmCommentStyle)
1735	{
1736	if ( !pSettings->fUpdateCopyrightYear
1737	&& pSettings->enmUpdateLicense == kScmLicense_LeaveAlone)
1738	return false;
1739
1740	/*
1741	* Try locate the relevant comments.
1742	*/
1743	SCMCOPYRIGHTINFO Info =
1744	{
1745	/.pState = / pState,
1746	/.enmCommentStyle = / enmCommentStyle,
1747
1748	/.cComments = / 0,
1749
1750	/.pszContributedBy = / NULL,
1751
1752	/.iLineComment = / UINT32_MAX,
1753	/.cLinesComment = / 0,
1754
1755	/.iLineCopyright = / UINT32_MAX,
1756	/.uFirstYear = / UINT32_MAX,
1757	/.uLastYear = / UINT32_MAX,
1758	/.fWellFormedCopyright = / false,
1759	/.fUpToDateCopyright = / false,
1760
1761	/.fOpenSource = / true,
1762	/.pExpectedLicense = / NULL,
1763	/.paLicenses = / pSettings->enmUpdateLicense != kScmLicense_Mit
1764	&& pSettings->enmUpdateLicense != kScmLicense_BasedOnMit
1765	? &g_aLicenses[0] : &g_aLicensesWithMit[0],
1766	/.enmLicenceOpt = / pSettings->enmUpdateLicense,
1767	/.iLineLicense = / UINT32_MAX,
1768	/.cLinesLicense = / 0,
1769	/.pCurrentLicense = / NULL,
1770	/.fIsCorrectLicense = / false,
1771	/.fWellFormedLicense = / false,
1772	/.fExternalLicense = / false,
1773
1774	/.fCheckForLgpl = / true,
1775	/.iLineLgplNotice = / UINT32_MAX,
1776	/.iLineAfterLgplComment = / UINT32_MAX,
1777	/.iLineLgplDisclaimer = / UINT32_MAX,
1778	};
1779
1780	/* Figure Info.fOpenSource and the desired license: */
1781	char *pszSyncProcess;
1782	int rc = ScmSvnQueryProperty(pState, "svn:sync-process", &pszSyncProcess);
1783	if (RT_SUCCESS(rc))
1784	{
1785	Info.fOpenSource = strcmp(RTStrStrip(pszSyncProcess), "export") == 0;
1786	RTStrFree(pszSyncProcess);
1787	}
1788	else if (rc == VERR_NOT_FOUND)
1789	Info.fOpenSource = false;
1790	else
1791	return ScmError(pState, rc, "ScmSvnQueryProperty(svn:sync-process): %Rrc\n", rc);
1792
1793	Info.pExpectedLicense = Info.paLicenses;
1794	if (Info.fOpenSource)
1795	{
1796	if ( pSettings->enmUpdateLicense != kScmLicense_Mit
1797	&& pSettings->enmUpdateLicense != kScmLicense_BasedOnMit)
1798	while (Info.pExpectedLicense->enmOpt != pSettings->enmUpdateLicense)
1799	Info.pExpectedLicense++;
1800	else
1801	Assert(Info.pExpectedLicense->enmOpt == kScmLicense_Mit);
1802	}
1803	else
1804	while (Info.pExpectedLicense->enmType != kScmLicenseType_Confidential)
1805	Info.pExpectedLicense++;
1806
1807	/* Scan the comments. */
1808	rc = ScmEnumerateComments(pIn, enmCommentStyle, rewrite_Copyright_CommentCallback, &Info);
1809	if ( (rc == VERR_CALLBACK_RETURN \|\| RT_SUCCESS(rc))
1810	&& RT_SUCCESS(pState->rc))
1811	{
1812	/*
1813	* Do conformity checks.
1814	*/
1815	bool fAddLgplDisclaimer = false;
1816	if (Info.fCheckforLgpl)
1817	{
1818	if ( Info.iLineLgplNotice != UINT32_MAX
1819	&& Info.iLineLgplDisclaimer == UINT32_MAX)
1820	{
1821	if (!pSettings->fLgplDisclaimer) /** @todo reconcile options with common sense. */
1822	ScmError(pState, VERR_NOT_FOUND, "LGPL licence notice on line %u, but no LGPL disclaimer was found!\n",
1823	Info.iLineLgplNotice + 1);
1824	else
1825	{
1826	ScmVerbose(pState, 1, "* Need to add LGPL disclaimer\n");
1827	fAddLgplDisclaimer = true;
1828	}
1829	}
1830	else if ( Info.iLineLgplNotice == UINT32_MAX
1831	&& Info.iLineLgplDisclaimer != UINT32_MAX)
1832	ScmError(pState, VERR_NOT_FOUND, "LGPL disclaimer on line %u, but no LGPL copyright notice!\n",
1833	Info.iLineLgplDisclaimer + 1);
1834	}
1835
1836	if (!pSettings->fExternalCopyright)
1837	{
1838	if (Info.iLineCopyright == UINT32_MAX)
1839	ScmError(pState, VERR_NOT_FOUND, "Missing copyright!\n");
1840	if (Info.iLineLicense == UINT32_MAX)
1841	ScmError(pState, VERR_NOT_FOUND, "Missing license!\n");
1842	}
1843	else if (Info.iLineCopyright != UINT32_MAX)
1844	ScmError(pState, VERR_NOT_FOUND,
1845	"Marked as external copyright only, but found non-external copyright statement at line %u!\n",
1846	Info.iLineCopyright + 1);
1847
1848
1849	if (RT_SUCCESS(pState->rc))
1850	{
1851	/*
1852	* Do we need to make any changes?
1853	*/
1854	bool fUpdateCopyright = !pSettings->fExternalCopyright
1855	&& ( !Info.fWellFormedCopyright
1856	\|\| (!Info.fUpToDateCopyright && pSettings->fUpdateCopyrightYear));
1857	bool fUpdateLicense = !pSettings->fExternalCopyright
1858	&& Info.enmLicenceOpt != kScmLicense_LeaveAlone
1859	&& ( !Info.fWellFormedLicense
1860	\|\| !Info.fIsCorrectLicense);
1861	if ( fUpdateCopyright
1862	\|\| fUpdateLicense
1863	\|\| fAddLgplDisclaimer)
1864	{
1865	Assert(Info.iLineComment != UINT32_MAX);
1866	Assert(Info.cLinesComment > 0);
1867
1868	/*
1869	* Okay, do the work.
1870	*/
1871	ScmStreamRewindForReading(pIn);
1872
1873	if (pSettings->fUpdateCopyrightYear)
1874	Info.uLastYear = g_uYear;
1875
1876	uint32_t iLine = 0;
1877	SCMEOL enmEol;
1878	size_t cchLine;
1879	const char *pchLine;
1880	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
1881	{
1882	if ( iLine == Info.iLineComment
1883	&& (fUpdateCopyright \|\| fUpdateLicense) )
1884	{
1885	/* Leading blank line. */
1886	ScmStreamPutLine(pOut, g_aCopyrightCommentStart[enmCommentStyle].psz,
1887	g_aCopyrightCommentStart[enmCommentStyle].cch, enmEol);
1888
1889	/* Contributed by someone? */
1890	if (Info.pszContributedBy)
1891	{
1892	const char *psz = Info.pszContributedBy;
1893	for (;;)
1894	{
1895	const char *pszEol = strchr(psz, '\n');
1896	size_t cchContribLine = pszEol ? pszEol - psz : strlen(psz);
1897	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1898	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1899	ScmStreamWrite(pOut, psz, cchContribLine);
1900	ScmStreamPutEol(pOut, enmEol);
1901	if (!pszEol)
1902	break;
1903	psz = pszEol + 1;
1904	}
1905
1906	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1907	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1908	}
1909
1910	/* Write the copyright comment line. */
1911	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1912	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1913
1914	char szCopyright[256];
1915	size_t cchCopyright;
1916	if (Info.uFirstYear == Info.uLastYear)
1917	cchCopyright = RTStrPrintf(szCopyright, sizeof(szCopyright), "Copyright (C) %u %s",
1918	Info.uFirstYear, g_szCopyrightHolder);
1919	else
1920	cchCopyright = RTStrPrintf(szCopyright, sizeof(szCopyright), "Copyright (C) %u-%u %s",
1921	Info.uFirstYear, Info.uLastYear, g_szCopyrightHolder);
1922
1923	ScmStreamWrite(pOut, szCopyright, cchCopyright);
1924	ScmStreamPutEol(pOut, enmEol);
1925
1926	if (pSettings->enmUpdateLicense != kScmLicense_BasedOnMit)
1927	{
1928	/* Blank line separating the two. */
1929	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1930	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1931
1932	/* Write the license text. */
1933	scmWriteCommentBody(pOut, Info.pExpectedLicense->psz, Info.pExpectedLicense->cch,
1934	enmCommentStyle, enmEol);
1935
1936	/* Final comment line. */
1937	if (!Info.fExternalLicense)
1938	ScmStreamPutLine(pOut, g_aCopyrightCommentEnd[enmCommentStyle].psz,
1939	g_aCopyrightCommentEnd[enmCommentStyle].cch, enmEol);
1940	}
1941	else
1942	Assert(Info.fExternalLicense);
1943
1944	/* Skip the copyright and license text in the input file. */
1945	rc = ScmStreamGetStatus(pOut);
1946	if (RT_SUCCESS(rc))
1947	{
1948	iLine = Info.iLineComment + Info.cLinesComment;
1949	rc = ScmStreamSeekByLine(pIn, iLine);
1950	}
1951	}
1952	/*
1953	* Add LGPL disclaimer?
1954	*/
1955	else if ( iLine == Info.iLineAfterLgplComment
1956	&& fAddLgplDisclaimer)
1957	{
1958	ScmStreamPutEol(pOut, enmEol);
1959	ScmStreamPutLine(pOut, g_aCopyrightCommentStart[enmCommentStyle].psz,
1960	g_aCopyrightCommentStart[enmCommentStyle].cch, enmEol);
1961	scmWriteCommentBody(pOut, g_szLgplDisclaimer, sizeof(g_szLgplDisclaimer) - 1,
1962	enmCommentStyle, enmEol);
1963	ScmStreamPutLine(pOut, g_aCopyrightCommentEnd[enmCommentStyle].psz,
1964	g_aCopyrightCommentEnd[enmCommentStyle].cch, enmEol);
1965
1966	/* put the actual line */
1967	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
1968	iLine++;
1969	}
1970	else
1971	{
1972	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
1973	iLine++;
1974	}
1975	if (RT_FAILURE(rc))
1976	{
1977	RTStrFree(Info.pszContributedBy);
1978	return false;
1979	}
1980	} /* for each source line */
1981
1982	RTStrFree(Info.pszContributedBy);
1983	return true;
1984	}
1985	}
1986	}
1987	else
1988	ScmError(pState, rc, "ScmEnumerateComments: %Rrc\n", rc);
1989	NOREF(pState); NOREF(pOut);
1990	RTStrFree(Info.pszContributedBy);
1991	return false;
1992	}
1993
1994
1995	/** Copyright updater for C-style comments. */
1996	bool rewrite_Copyright_CstyleComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
1997	{
1998	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_C);
1999	}
2000
2001	/** Copyright updater for hash-prefixed comments. */
2002	bool rewrite_Copyright_HashComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2003	{
2004	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Hash);
2005	}
2006
2007	/** Copyright updater for REM-prefixed comments. */
2008	bool rewrite_Copyright_RemComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2009	{
2010	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, determinBatchFileCommentStyle(pIn));
2011	}
2012
2013	/** Copyright updater for python comments. */
2014	bool rewrite_Copyright_PythonComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2015	{
2016	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Python);
2017	}
2018
2019	/** Copyright updater for semicolon-prefixed comments. */
2020	bool rewrite_Copyright_SemicolonComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2021	{
2022	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Semicolon);
2023	}
2024
2025	/** Copyright updater for sql comments. */
2026	bool rewrite_Copyright_SqlComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2027	{
2028	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Sql);
2029	}
2030
2031	/** Copyright updater for tick-prefixed comments. */
2032	bool rewrite_Copyright_TickComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2033	{
2034	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Tick);
2035	}
2036
2037
2038	/**
2039	* Makefile.kup are empty files, enforce this.
2040	*
2041	* @returns true if modifications were made, false if not.
2042	* @param pIn The input stream.
2043	* @param pOut The output stream.
2044	* @param pSettings The settings.
2045	*/
2046	bool rewrite_Makefile_kup(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2047	{
2048	RT_NOREF2(pOut, pSettings);
2049
2050	/* These files should be zero bytes. */
2051	if (pIn->cb == 0)
2052	return false;
2053	ScmVerbose(pState, 2, " * Truncated file to zero bytes\n");
2054	return true;
2055	}
2056
2057	/**
2058	* Rewrite a kBuild makefile.
2059	*
2060	* @returns true if modifications were made, false if not.
2061	* @param pIn The input stream.
2062	* @param pOut The output stream.
2063	* @param pSettings The settings.
2064	*
2065	* @todo
2066	*
2067	* Ideas for Makefile.kmk and Config.kmk:
2068	* - sort if1of/ifn1of sets.
2069	* - line continuation slashes should only be preceded by one space.
2070	*/
2071	bool rewrite_Makefile_kmk(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2072	{
2073	RT_NOREF4(pState, pIn, pOut, pSettings);
2074	return false;
2075	}
2076
2077
2078	static bool isFlowerBoxSectionMarker(PSCMSTREAM pIn, const char *pchLine, size_t cchLine, uint32_t cchWidth,
2079	const char *ppchText, size_t pcchText, bool *pfNeedFixing)
2080	{
2081	*ppchText = NULL;
2082	*pcchText = 0;
2083	*pfNeedFixing = false;
2084
2085	/*
2086	* The first line.
2087	*/
2088	if (pchLine[0] != '/')
2089	return false;
2090	size_t offLine = 1;
2091	while (offLine < cchLine && pchLine[offLine] == '*')
2092	offLine++;
2093	if (offLine < 20) /* (Code below depend on a reasonable minimum here.) */
2094	return false;
2095	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2096	offLine++;
2097	if (offLine != cchLine)
2098	return false;
2099
2100	size_t const cchBox = cchLine;
2101	*pfNeedFixing = cchBox != cchWidth;
2102
2103	/*
2104	* The next line, extracting the text.
2105	*/
2106	SCMEOL enmEol;
2107	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2108	if (cchLine < cchBox - 3)
2109	return false;
2110
2111	offLine = 0;
2112	if (RT_C_IS_BLANK(pchLine[0]))
2113	{
2114	*pfNeedFixing = true;
2115	offLine = RT_C_IS_BLANK(pchLine[1]) ? 2 : 1;
2116	}
2117
2118	if (pchLine[offLine] != '*')
2119	return false;
2120	offLine++;
2121
2122	if (!RT_C_IS_BLANK(pchLine[offLine + 1]))
2123	return false;
2124	offLine++;
2125
2126	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2127	offLine++;
2128	if (offLine >= cchLine)
2129	return false;
2130	if (!RT_C_IS_UPPER(pchLine[offLine]))
2131	return false;
2132
2133	if (offLine != 4 \|\| cchLine != cchBox)
2134	*pfNeedFixing = true;
2135
2136	*ppchText = &pchLine[offLine];
2137	size_t const offText = offLine;
2138
2139	/* From the end now. */
2140	offLine = cchLine - 1;
2141	while (RT_C_IS_BLANK(pchLine[offLine]))
2142	offLine--;
2143
2144	if (pchLine[offLine] != '*')
2145	return false;
2146	offLine--;
2147	if (!RT_C_IS_BLANK(pchLine[offLine]))
2148	return false;
2149	offLine--;
2150	while (RT_C_IS_BLANK(pchLine[offLine]))
2151	offLine--;
2152	*pcchText = offLine - offText + 1;
2153
2154	/*
2155	* Third line closes the box.
2156	*/
2157	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2158	if (cchLine < cchBox - 3)
2159	return false;
2160
2161	offLine = 0;
2162	if (RT_C_IS_BLANK(pchLine[0]))
2163	{
2164	*pfNeedFixing = true;
2165	offLine = RT_C_IS_BLANK(pchLine[1]) ? 2 : 1;
2166	}
2167	while (offLine < cchLine && pchLine[offLine] == '*')
2168	offLine++;
2169	if (offLine < cchBox - 4)
2170	return false;
2171
2172	if (pchLine[offLine] != '/')
2173	return false;
2174	offLine++;
2175
2176	if (offLine != cchBox)
2177	*pfNeedFixing = true;
2178
2179	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2180	offLine++;
2181	if (offLine != cchLine)
2182	return false;
2183
2184	return true;
2185	}
2186
2187
2188	/**
2189	* Flower box marker comments in C and C++ code.
2190	*
2191	* @returns true if modifications were made, false if not.
2192	* @param pIn The input stream.
2193	* @param pOut The output stream.
2194	* @param pSettings The settings.
2195	*/
2196	bool rewrite_FixFlowerBoxMarkers(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2197	{
2198	if (!pSettings->fFixFlowerBoxMarkers)
2199	return false;
2200
2201	/*
2202	* Work thru the file line by line looking for flower box markers.
2203	*/
2204	size_t cChanges = 0;
2205	size_t cBlankLines = 0;
2206	SCMEOL enmEol;
2207	size_t cchLine;
2208	const char *pchLine;
2209	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2210	{
2211	/*
2212	* Get a likely match for a first line.
2213	*/
2214	if ( pchLine[0] == '/'
2215	&& cchLine > 20
2216	&& pchLine[1] == '*'
2217	&& pchLine[2] == '*'
2218	&& pchLine[3] == '*')
2219	{
2220	size_t const offSaved = ScmStreamTell(pIn);
2221	char const *pchText;
2222	size_t cchText;
2223	bool fNeedFixing;
2224	bool fIsFlowerBoxSection = isFlowerBoxSectionMarker(pIn, pchLine, cchLine, pSettings->cchWidth,
2225	&pchText, &cchText, &fNeedFixing);
2226	if ( fIsFlowerBoxSection
2227	&& ( fNeedFixing
2228	\|\| cBlankLines < pSettings->cMinBlankLinesBeforeFlowerBoxMakers) )
2229	{
2230	while (cBlankLines < pSettings->cMinBlankLinesBeforeFlowerBoxMakers)
2231	{
2232	ScmStreamPutEol(pOut, enmEol);
2233	cBlankLines++;
2234	}
2235
2236	ScmStreamPutCh(pOut, '/');
2237	ScmStreamWrite(pOut, g_szAsterisks, pSettings->cchWidth - 1);
2238	ScmStreamPutEol(pOut, enmEol);
2239
2240	static const char s_szLead[] = "* ";
2241	ScmStreamWrite(pOut, s_szLead, sizeof(s_szLead) - 1);
2242	ScmStreamWrite(pOut, pchText, cchText);
2243	size_t offCurPlus1 = sizeof(s_szLead) - 1 + cchText + 1;
2244	ScmStreamWrite(pOut, g_szSpaces, offCurPlus1 < pSettings->cchWidth ? pSettings->cchWidth - offCurPlus1 : 1);
2245	ScmStreamPutCh(pOut, '*');
2246	ScmStreamPutEol(pOut, enmEol);
2247
2248	ScmStreamWrite(pOut, g_szAsterisks, pSettings->cchWidth - 1);
2249	ScmStreamPutCh(pOut, '/');
2250	ScmStreamPutEol(pOut, enmEol);
2251
2252	cChanges++;
2253	cBlankLines = 0;
2254	continue;
2255	}
2256
2257	int rc = ScmStreamSeekAbsolute(pIn, offSaved);
2258	if (RT_FAILURE(rc))
2259	return false;
2260	}
2261
2262	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2263	if (RT_FAILURE(rc))
2264	return false;
2265
2266	/* Do blank line accounting so we can ensure at least two blank lines
2267	before each section marker. */
2268	if (!isBlankLine(pchLine, cchLine))
2269	cBlankLines = 0;
2270	else
2271	cBlankLines++;
2272	}
2273	if (cChanges > 0)
2274	ScmVerbose(pState, 2, " * Converted %zu flower boxer markers\n", cChanges);
2275	return cChanges != 0;
2276	}
2277
2278
2279	/**
2280	* Looks for the start of a todo comment.
2281	*
2282	* @returns Offset into the line of the comment start sequence.
2283	* @param pchLine The line to search.
2284	* @param cchLineBeforeTodo The length of the line before the todo.
2285	* @param pfSameLine Indicates whether it's refering to a statemtn on
2286	* the same line comment (true), or the next
2287	* statement (false).
2288	*/
2289	static size_t findTodoCommentStart(char const pchLine, size_t cchLineBeforeTodo, bool pfSameLine)
2290	{
2291	*pfSameLine = false;
2292
2293	/* Skip one '@' or '\\'. */
2294	char ch;
2295	if ( cchLineBeforeTodo > 2
2296	&& ( ((ch = pchLine[cchLineBeforeTodo - 1]) == '@')
2297	\|\| ch == '\\' ) )
2298	cchLineBeforeTodo--;
2299
2300	/* Skip blanks. */
2301	while ( cchLineBeforeTodo > 2
2302	&& RT_C_IS_BLANK(pchLine[cchLineBeforeTodo - 1]))
2303	cchLineBeforeTodo--;
2304
2305	/* Look for same line indicator. */
2306	if ( cchLineBeforeTodo > 0
2307	&& pchLine[cchLineBeforeTodo - 1] == '<')
2308	{
2309	*pfSameLine = true;
2310	cchLineBeforeTodo--;
2311	}
2312
2313	/* Skip s /
2314	while ( cchLineBeforeTodo > 1
2315	&& pchLine[cchLineBeforeTodo - 1] == '*')
2316	cchLineBeforeTodo--;
2317
2318	/* Do we have a comment opening sequence. */
2319	if ( cchLineBeforeTodo > 0
2320	&& pchLine[cchLineBeforeTodo - 1] == '/'
2321	&& ( ( cchLineBeforeTodo >= 2
2322	&& pchLine[cchLineBeforeTodo - 2] == '/')
2323	\|\| pchLine[cchLineBeforeTodo] == '*'))
2324	{
2325	/* Skip slashes at the start. */
2326	while ( cchLineBeforeTodo > 0
2327	&& pchLine[cchLineBeforeTodo - 1] == '/')
2328	cchLineBeforeTodo--;
2329
2330	return cchLineBeforeTodo;
2331	}
2332
2333	return ~(size_t)0;
2334	}
2335
2336
2337	/**
2338	* Looks for a TODO or todo in the given line.
2339	*
2340	* @returns Offset into the line of found, ~(size_t)0 if not.
2341	* @param pchLine The line to search.
2342	* @param cchLine The length of the line.
2343	*/
2344	static size_t findTodo(char const *pchLine, size_t cchLine)
2345	{
2346	if (cchLine >= 4 + 2)
2347	{
2348	/* We don't search the first to chars because we need the start of a comment.
2349	Also, skip the last three chars since we need at least four for a match. */
2350	size_t const cchLineT = cchLine - 3;
2351	if ( memchr(pchLine + 2, 't', cchLineT - 2) != NULL
2352	\|\| memchr(pchLine + 2, 'T', cchLineT - 2) != NULL)
2353	{
2354	for (size_t off = 2; off < cchLineT; off++)
2355	{
2356	char ch = pchLine[off];
2357	if ( ( ch != 't'
2358	&& ch != 'T')
2359	\|\| ( (ch = pchLine[off + 1]) != 'o'
2360	&& ch != 'O')
2361	\|\| ( (ch = pchLine[off + 2]) != 'd'
2362	&& ch != 'D')
2363	\|\| ( (ch = pchLine[off + 3]) != 'o'
2364	&& ch != 'O')
2365	\|\| ( off + 4 != cchLine
2366	&& (ch = pchLine[off + 4]) != ' '
2367	&& ch != '\t'
2368	&& ch != ':' /** @todo */
2369	&& (ch != '' \|\| off + 5 > cchLine \|\| pchLine[off + 5] != '/') /* @todo */
2370	) )
2371	{ /* not a hit - likely */ }
2372	else
2373	return off;
2374	}
2375	}
2376	}
2377	return ~(size_t)0;
2378	}
2379
2380
2381	/**
2382	* Doxygen todos in C and C++ code.
2383	*
2384	* @returns true if modifications were made, false if not.
2385	* @param pState The rewriter state.
2386	* @param pIn The input stream.
2387	* @param pOut The output stream.
2388	* @param pSettings The settings.
2389	*/
2390	bool rewrite_Fix_C_and_CPP_Todos(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2391	{
2392	if (!pSettings->fFixTodos)
2393	return false;
2394
2395	/*
2396	* Work thru the file line by line looking for the start of todo comments.
2397	*/
2398	size_t cChanges = 0;
2399	SCMEOL enmEol;
2400	size_t cchLine;
2401	const char *pchLine;
2402	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2403	{
2404	/*
2405	* Look for the word 'todo' in the line. We're currently only trying
2406	* to catch comments starting with the word todo and adjust the start of
2407	* the doxygen statement.
2408	*/
2409	size_t offTodo = findTodo(pchLine, cchLine);
2410	if ( offTodo != ~(size_t)0
2411	&& offTodo >= 2)
2412	{
2413	/* Work backwards to find the start of the comment. */
2414	bool fSameLine = false;
2415	size_t offCommentStart = findTodoCommentStart(pchLine, offTodo, &fSameLine);
2416	if (offCommentStart != ~(size_t)0)
2417	{
2418	char szNew[64];
2419	size_t cchNew = 0;
2420	szNew[cchNew++] = '/';
2421	szNew[cchNew++] = pchLine[offCommentStart + 1];
2422	szNew[cchNew++] = pchLine[offCommentStart + 1];
2423	if (fSameLine)
2424	szNew[cchNew++] = '<';
2425	szNew[cchNew++] = ' ';
2426	szNew[cchNew++] = '@';
2427	szNew[cchNew++] = 't';
2428	szNew[cchNew++] = 'o';
2429	szNew[cchNew++] = 'd';
2430	szNew[cchNew++] = 'o';
2431
2432	/* Figure out wheter to continue after the @todo statement opening, we'll strip ':'
2433	but need to take into account that we might be at the end of the line before
2434	adding the space. */
2435	size_t offTodoAfter = offTodo + 4;
2436	if ( offTodoAfter < cchLine
2437	&& pchLine[offTodoAfter] == ':')
2438	offTodoAfter++;
2439	if ( offTodoAfter < cchLine
2440	&& RT_C_IS_BLANK(pchLine[offTodoAfter]))
2441	offTodoAfter++;
2442	if (offTodoAfter < cchLine)
2443	szNew[cchNew++] = ' ';
2444
2445	/* Write it out. */
2446	ScmStreamWrite(pOut, pchLine, offCommentStart);
2447	ScmStreamWrite(pOut, szNew, cchNew);
2448	if (offTodoAfter < cchLine)
2449	ScmStreamWrite(pOut, &pchLine[offTodoAfter], cchLine - offTodoAfter);
2450	ScmStreamPutEol(pOut, enmEol);
2451
2452	/* Check whether we actually made any changes. */
2453	if ( cchNew != offTodoAfter - offCommentStart
2454	\|\| memcmp(szNew, &pchLine[offCommentStart], cchNew))
2455	cChanges++;
2456	continue;
2457	}
2458	}
2459
2460	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2461	if (RT_FAILURE(rc))
2462	return false;
2463	}
2464	if (cChanges > 0)
2465	ScmVerbose(pState, 2, " * Converted %zu todo statements.\n", cChanges);
2466	return cChanges != 0;
2467	}
2468
2469
2470	/**
2471	* Tries to parse a C/C++ preprocessor include directive.
2472	*
2473	* This is resonably forgiving and expects sane input.
2474	*
2475	* @retval kScmIncludeDir_Invalid if not a valid include directive.
2476	* @retval kScmIncludeDir_Quoted
2477	* @retval kScmIncludeDir_Bracketed
2478	* @retval kScmIncludeDir_Macro
2479	*
2480	* @param pState The rewriter state (for repording malformed
2481	* directives).
2482	* @param pchLine The line to try parse as an include statement.
2483	* @param cchLine The line length.
2484	* @param ppchFilename Where to return the pointer to the filename part.
2485	* @param pcchFilename Where to return the length of the filename.
2486	*/
2487	SCMINCLUDEDIR ScmMaybeParseCIncludeLine(PSCMRWSTATE pState, const char *pchLine, size_t cchLine,
2488	const char *ppchFilename, size_t pcchFilename)
2489	{
2490	/* Skip leading spaces: */
2491	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2492	cchLine--, pchLine++;
2493
2494	/* Check for '#': */
2495	if (cchLine > 0 && *pchLine == '#')
2496	{
2497	cchLine--;
2498	pchLine++;
2499
2500	/* Skip spaces after '#' (optional): */
2501	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2502	cchLine--, pchLine++;
2503
2504	/* Check for 'include': */
2505	static char const s_szInclude[] = "include";
2506	if ( cchLine >= sizeof(s_szInclude)
2507	&& memcmp(pchLine, RT_STR_TUPLE(s_szInclude)) == 0)
2508	{
2509	cchLine -= sizeof(s_szInclude) - 1;
2510	pchLine += sizeof(s_szInclude) - 1;
2511
2512	/* Skip spaces after 'include' word (optional): */
2513	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2514	cchLine--, pchLine++;
2515	if (cchLine > 0)
2516	{
2517	/* Quoted or bracketed? */
2518	char const chFirst = *pchLine;
2519	if (chFirst == '"' \|\| chFirst == '<')
2520	{
2521	cchLine--;
2522	pchLine++;
2523	const char pchEnd = (const char )memchr(pchLine, chFirst == '"' ? '"' : '>', cchLine);
2524	if (pchEnd)
2525	{
2526	if (ppchFilename)
2527	*ppchFilename = pchLine;
2528	if (pcchFilename)
2529	*pcchFilename = pchEnd - pchLine;
2530	return chFirst == '"' ? kScmIncludeDir_Quoted : kScmIncludeDir_Bracketed;
2531	}
2532	ScmError(pState, VERR_PARSE_ERROR, "Unbalanced #include filename %s: %.*s\n",
2533	chFirst == '"' ? "quotes" : "brackets" , cchLine, pchLine);
2534	}
2535	/* C prepreprocessor macro? */
2536	else if (ScmIsCIdentifierLeadChar(chFirst))
2537	{
2538	size_t cchFilename = 1;
2539	while ( cchFilename < cchLine
2540	&& ScmIsCIdentifierChar(pchLine[cchFilename]))
2541	cchFilename++;
2542	if (ppchFilename)
2543	*ppchFilename = pchLine;
2544	if (pcchFilename)
2545	*pcchFilename = cchFilename;
2546	return kScmIncludeDir_Macro;
2547	}
2548	else
2549	ScmError(pState, VERR_PARSE_ERROR, "Malformed #include filename part: %.*s\n", cchLine, pchLine);
2550	}
2551	else
2552	ScmError(pState, VERR_PARSE_ERROR, "Missing #include filename!\n");
2553	}
2554	}
2555
2556	if (ppchFilename)
2557	*ppchFilename = NULL;
2558	if (pcchFilename)
2559	*pcchFilename = 0;
2560	return kScmIncludeDir_Invalid;
2561	}
2562
2563
2564	/**
2565	* Fix err.h/errcore.h usage.
2566	*
2567	* @returns true if modifications were made, false if not.
2568	* @param pIn The input stream.
2569	* @param pOut The output stream.
2570	* @param pSettings The settings.
2571	*/
2572	bool rewrite_Fix_Err_H(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2573	{
2574	if (!pSettings->fFixErrH)
2575	return false;
2576
2577	static struct
2578	{
2579	const char *pszHeader;
2580	unsigned cchHeader;
2581	int iLevel;
2582	} const s_aHeaders[] =
2583	{
2584	{ RT_STR_TUPLE("iprt/errcore.h"), 1 },
2585	{ RT_STR_TUPLE("iprt/err.h"), 2 },
2586	{ RT_STR_TUPLE("VBox/err.h"), 3 },
2587	};
2588	static RTSTRTUPLE const g_aLevel1Statuses[] = /* Note! Keep in sync with errcore.h content! */
2589	{
2590	{ RT_STR_TUPLE("VINF_SUCCESS") },
2591	{ RT_STR_TUPLE("VERR_GENERAL_FAILURE") },
2592	{ RT_STR_TUPLE("VERR_INVALID_PARAMETER") },
2593	{ RT_STR_TUPLE("VWRN_INVALID_PARAMETER") },
2594	{ RT_STR_TUPLE("VERR_INVALID_MAGIC") },
2595	{ RT_STR_TUPLE("VWRN_INVALID_MAGIC") },
2596	{ RT_STR_TUPLE("VERR_INVALID_HANDLE") },
2597	{ RT_STR_TUPLE("VWRN_INVALID_HANDLE") },
2598	{ RT_STR_TUPLE("VERR_INVALID_POINTER") },
2599	{ RT_STR_TUPLE("VERR_NO_MEMORY") },
2600	{ RT_STR_TUPLE("VERR_PERMISSION_DENIED") },
2601	{ RT_STR_TUPLE("VINF_PERMISSION_DENIED") },
2602	{ RT_STR_TUPLE("VERR_VERSION_MISMATCH") },
2603	{ RT_STR_TUPLE("VERR_NOT_IMPLEMENTED") },
2604	{ RT_STR_TUPLE("VERR_INVALID_FLAGS") },
2605	{ RT_STR_TUPLE("VERR_WRONG_ORDER") },
2606	{ RT_STR_TUPLE("VERR_INVALID_FUNCTION") },
2607	{ RT_STR_TUPLE("VERR_NOT_SUPPORTED") },
2608	{ RT_STR_TUPLE("VINF_NOT_SUPPORTED") },
2609	{ RT_STR_TUPLE("VERR_ACCESS_DENIED") },
2610	{ RT_STR_TUPLE("VERR_INTERRUPTED") },
2611	{ RT_STR_TUPLE("VINF_INTERRUPTED") },
2612	{ RT_STR_TUPLE("VERR_TIMEOUT") },
2613	{ RT_STR_TUPLE("VINF_TIMEOUT") },
2614	{ RT_STR_TUPLE("VERR_BUFFER_OVERFLOW") },
2615	{ RT_STR_TUPLE("VINF_BUFFER_OVERFLOW") },
2616	{ RT_STR_TUPLE("VERR_TOO_MUCH_DATA") },
2617	{ RT_STR_TUPLE("VERR_TRY_AGAIN") },
2618	{ RT_STR_TUPLE("VINF_TRY_AGAIN") },
2619	{ RT_STR_TUPLE("VERR_PARSE_ERROR") },
2620	{ RT_STR_TUPLE("VERR_OUT_OF_RANGE") },
2621	{ RT_STR_TUPLE("VERR_NUMBER_TOO_BIG") },
2622	{ RT_STR_TUPLE("VWRN_NUMBER_TOO_BIG") },
2623	{ RT_STR_TUPLE("VERR_CANCELLED") },
2624	{ RT_STR_TUPLE("VERR_TRAILING_CHARS") },
2625	{ RT_STR_TUPLE("VWRN_TRAILING_CHARS") },
2626	{ RT_STR_TUPLE("VERR_TRAILING_SPACES") },
2627	{ RT_STR_TUPLE("VWRN_TRAILING_SPACES") },
2628	{ RT_STR_TUPLE("VERR_NOT_FOUND") },
2629	{ RT_STR_TUPLE("VWRN_NOT_FOUND") },
2630	{ RT_STR_TUPLE("VERR_INVALID_STATE") },
2631	{ RT_STR_TUPLE("VWRN_INVALID_STATE") },
2632	{ RT_STR_TUPLE("VERR_OUT_OF_RESOURCES") },
2633	{ RT_STR_TUPLE("VWRN_OUT_OF_RESOURCES") },
2634	{ RT_STR_TUPLE("VERR_END_OF_STRING") },
2635	{ RT_STR_TUPLE("VERR_CALLBACK_RETURN") },
2636	{ RT_STR_TUPLE("VINF_CALLBACK_RETURN") },
2637	{ RT_STR_TUPLE("VERR_DUPLICATE") },
2638	{ RT_STR_TUPLE("VERR_MISSING") },
2639	{ RT_STR_TUPLE("VERR_BUFFER_UNDERFLOW") },
2640	{ RT_STR_TUPLE("VINF_BUFFER_UNDERFLOW") },
2641	{ RT_STR_TUPLE("VERR_NOT_AVAILABLE") },
2642	{ RT_STR_TUPLE("VERR_MISMATCH") },
2643	{ RT_STR_TUPLE("VERR_WRONG_TYPE") },
2644	{ RT_STR_TUPLE("VWRN_WRONG_TYPE") },
2645	{ RT_STR_TUPLE("VERR_WRONG_PARAMETER_COUNT") },
2646	{ RT_STR_TUPLE("VERR_WRONG_PARAMETER_TYPE") },
2647	{ RT_STR_TUPLE("VERR_INVALID_CLIENT_ID") },
2648	{ RT_STR_TUPLE("VERR_INVALID_SESSION_ID") },
2649	{ RT_STR_TUPLE("VERR_INCOMPATIBLE_CONFIG") },
2650	{ RT_STR_TUPLE("VERR_INTERNAL_ERROR") },
2651	{ RT_STR_TUPLE("VINF_GETOPT_NOT_OPTION") },
2652	{ RT_STR_TUPLE("VERR_GETOPT_UNKNOWN_OPTION") },
2653	};
2654
2655	/*
2656	* First pass: Scout #include err.h/errcore.h locations and usage.
2657	*
2658	* Note! This isn't entirely optimal since it's also parsing comments and
2659	* strings, not just code. However it does a decent job for now.
2660	*/
2661	int iIncludeLevel = 0;
2662	int iUsageLevel = 0;
2663	uint32_t iLine = 0;
2664	SCMEOL enmEol;
2665	size_t cchLine;
2666	const char *pchLine;
2667	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2668	{
2669	iLine++;
2670	if (cchLine < 6)
2671	continue;
2672
2673	/*
2674	* Look for #includes.
2675	*/
2676	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2677	if ( pchHash
2678	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
2679	{
2680	const char *pchFilename;
2681	size_t cchFilename;
2682	SCMINCLUDEDIR enmIncDir = ScmMaybeParseCIncludeLine(pState, pchLine, cchLine, &pchFilename, &cchFilename);
2683	if ( enmIncDir == kScmIncludeDir_Bracketed
2684	\|\| enmIncDir == kScmIncludeDir_Quoted)
2685	{
2686	unsigned i = RT_ELEMENTS(s_aHeaders);
2687	while (i-- > 0)
2688	if ( s_aHeaders[i].cchHeader == cchFilename
2689	&& RTStrNICmpAscii(pchFilename, s_aHeaders[i].pszHeader, cchFilename) == 0)
2690	{
2691	if (iIncludeLevel < s_aHeaders[i].iLevel)
2692	iIncludeLevel = s_aHeaders[i].iLevel;
2693	break;
2694	}
2695
2696	/* Special hack for error info. */
2697	if (cchFilename == sizeof("errmsgdata.h") - 1 && memcmp(pchFilename, RT_STR_TUPLE("errmsgdata.h")) == 0)
2698	iUsageLevel = 4;
2699
2700	/* Special hack for code templates. */
2701	if ( cchFilename >= sizeof(".cpp.h")
2702	&& memcmp(&pchFilename[cchFilename - sizeof(".cpp.h") + 1], RT_STR_TUPLE(".cpp.h")) == 0)
2703	iUsageLevel = 4;
2704	continue;
2705	}
2706	}
2707	/*
2708	* Look for VERR_, VWRN_, VINF_ prefixed identifiers in the current line.
2709	*/
2710	const char pchHit = (const char )memchr(pchLine, 'V', cchLine);
2711	if (pchHit)
2712	{
2713	const char *pchLeft = pchLine;
2714	size_t cchLeft = cchLine;
2715	do
2716	{
2717	size_t cchLeftHit = &pchLeft[cchLeft] - pchHit;
2718	if (cchLeftHit < 6)
2719	break;
2720	if ( pchHit[4] == '_'
2721	&& ( pchHit == pchLine
2722	\|\| !ScmIsCIdentifierChar(pchHit[-1]))
2723	&& ( (pchHit[1] == 'E' && pchHit[2] == 'R' && pchHit[3] == 'R')
2724	\|\| (pchHit[1] == 'W' && pchHit[2] == 'R' && pchHit[3] == 'N')
2725	\|\| (pchHit[1] == 'I' && pchHit[2] == 'N' && pchHit[3] == 'F') ) )
2726	{
2727	size_t cchIdentifier = 5;
2728	while (cchIdentifier < cchLeftHit && ScmIsCIdentifierChar(pchHit[cchIdentifier]))
2729	cchIdentifier++;
2730	ScmVerbose(pState, 4, "--- status code at %u col %zu: %.*s\n",
2731	iLine, pchHit - pchLine, cchIdentifier, pchHit);
2732
2733	if (iUsageLevel <= 1)
2734	{
2735	iUsageLevel = 3; /* Cannot distingish between iprt/err.h and VBox/err.h, so pick the latter for now. */
2736	for (unsigned i = 0; i < RT_ELEMENTS(g_aLevel1Statuses); i++)
2737	if ( cchIdentifier == g_aLevel1Statuses[i].cch
2738	&& memcmp(pchHit, g_aLevel1Statuses[i].psz, cchIdentifier) == 0)
2739	{
2740	iUsageLevel = 1;
2741	break;
2742	}
2743	}
2744
2745	pchLeft = pchHit + cchIdentifier;
2746	cchLeft = cchLeftHit - cchIdentifier;
2747	}
2748	else
2749	{
2750	pchLeft = pchHit + 1;
2751	cchLeft = cchLeftHit - 1;
2752	}
2753	pchHit = (const char *)memchr(pchLeft, 'V', cchLeft);
2754	} while (pchHit != NULL);
2755	}
2756	}
2757	ScmVerbose(pState, 3, "--- iIncludeLevel=%d iUsageLevel=%d\n", iIncludeLevel, iUsageLevel);
2758
2759	/*
2760	* Second pass: Change err.h to errcore.h if we detected a need for change.
2761	*/
2762	if ( iIncludeLevel <= iUsageLevel
2763	\|\| iIncludeLevel <= 1 /* we cannot safely eliminate errcore.h includes atm. */)
2764	return false;
2765
2766	unsigned cChanges = 0;
2767	ScmStreamRewindForReading(pIn);
2768	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2769	{
2770	/*
2771	* Look for #includes to modify.
2772	*/
2773	if (cchLine >= 6)
2774	{
2775	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2776	if ( pchHash
2777	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
2778	{
2779	const char *pchFilename;
2780	size_t cchFilename;
2781	SCMINCLUDEDIR enmIncDir = ScmMaybeParseCIncludeLine(pState, pchLine, cchLine, &pchFilename, &cchFilename);
2782	if ( enmIncDir == kScmIncludeDir_Bracketed
2783	\|\| enmIncDir == kScmIncludeDir_Quoted)
2784	{
2785	unsigned i = RT_ELEMENTS(s_aHeaders);
2786	while (i-- > 0)
2787	if ( s_aHeaders[i].cchHeader == cchFilename
2788	&& RTStrNICmpAscii(pchFilename, s_aHeaders[i].pszHeader, cchFilename) == 0)
2789	{
2790	ScmStreamWrite(pOut, pchLine, pchFilename - pchLine - 1);
2791	ScmStreamWrite(pOut, RT_STR_TUPLE("<iprt/errcore.h>"));
2792	size_t cchTrailing = &pchLine[cchLine] - &pchFilename[cchFilename + 1];
2793	if (cchTrailing > 0)
2794	ScmStreamWrite(pOut, &pchFilename[cchFilename + 1], cchTrailing);
2795	ScmStreamPutEol(pOut, enmEol);
2796	cChanges++;
2797	pchLine = NULL;
2798	break;
2799	}
2800	if (!pchLine)
2801	continue;
2802	}
2803	}
2804	}
2805
2806	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2807	if (RT_FAILURE(rc))
2808	return false;
2809	}
2810	ScmVerbose(pState, 2, " * Converted %zu err.h/errcore.h include statements.\n", cChanges);
2811	return true;
2812	}
2813
2814	typedef struct
2815	{
2816	const char *pch;
2817	uint8_t cch;
2818	uint8_t cchSpaces; /*< Number of expected spaces before the word. /
2819	bool fSpacesBefore : 1; /*< Whether there may be spaces or tabs before the word. /
2820	bool fIdentifier : 1; /*< Whether we're to expect a C/C++ identifier rather than pch/cch. /
2821	} SCMMATCHWORD;
2822
2823
2824	int ScmMatchWords(const char pchLine, size_t cchLine, SCMMATCHWORD const paWords, size_t cWords,
2825	size_t *poffNext, PRTSTRTUPLE paIdentifiers, PRTERRINFO pErrInfo)
2826	{
2827	int rc = VINF_SUCCESS;
2828
2829	size_t offLine = 0;
2830	for (size_t i = 0; i < cWords; i++)
2831	{
2832	SCMMATCHWORD const *pWord = &paWords[i];
2833
2834	/*
2835	* Deal with spaces preceeding the word first:
2836	*/
2837	if (pWord->fSpacesBefore)
2838	{
2839	size_t cchSpaces = 0;
2840	size_t cchTabs = 0;
2841	while (offLine < cchLine)
2842	{
2843	const char ch = pchLine[offLine];
2844	if (ch == ' ')
2845	cchSpaces++;
2846	else if (ch == '\t')
2847	cchTabs++;
2848	else
2849	break;
2850	offLine++;
2851	}
2852
2853	if (cchSpaces == pWord->cchSpaces && cchTabs == 0)
2854	{ /* likely */ }
2855	else if (cchSpaces == 0 && cchTabs == 0)
2856	return RTErrInfoSetF(pErrInfo, VERR_PARSE_ERROR, "expected space at offset %u", offLine);
2857	else
2858	rc = VWRN_TRAILING_SPACES;
2859	}
2860	else
2861	Assert(pWord->cchSpaces == 0);
2862
2863	/*
2864	* C/C++ identifier?
2865	*/
2866	if (pWord->fIdentifier)
2867	{
2868	if (offLine >= cchLine)
2869	return RTErrInfoSetF(pErrInfo, VERR_END_OF_STRING,
2870	"expected '%.*s' (C/C++ identifier) at offset %u, not end of string",
2871	pWord->cch, pWord->pch, offLine);
2872	if (!ScmIsCIdentifierLeadChar(pchLine[offLine]))
2873	return RTErrInfoSetF(pErrInfo, VERR_MISMATCH, "expected '%.*s' (C/C++ identifier) at offset %u",
2874	pWord->cch, pWord->pch, offLine);
2875	size_t const offStart = offLine++;
2876	while (offLine < cchLine && ScmIsCIdentifierChar(pchLine[offLine]))
2877	offLine++;
2878	if (paIdentifiers)
2879	{
2880	paIdentifiers->cch = offLine - offStart;
2881	paIdentifiers->psz = &pchLine[offStart];
2882	paIdentifiers++;
2883	}
2884	}
2885	/*
2886	* Match the exact word.
2887	*/
2888	else if ( pWord->cch == 0
2889	\|\| ( pWord->cch <= cchLine - offLine
2890	&& !memcmp(pWord->pch, &pchLine[offLine], pWord->cch)))
2891	offLine += pWord->cch;
2892	else
2893	return RTErrInfoSetF(pErrInfo, VERR_MISMATCH, "expected '%.*s' at offset %u", pWord->cch, pWord->pch, offLine);
2894	}
2895
2896	/*
2897	* Check for trailing characters/whatnot.
2898	*/
2899	if (poffNext)
2900	*poffNext = offLine;
2901	else if (offLine != cchLine)
2902	rc = RTErrInfoSetF(pErrInfo, VERR_TRAILING_CHARS, "unexpected trailing characters at offset %u", offLine);
2903	return rc;
2904	}
2905
2906
2907	/**
2908	* Fix header file include guards and \#pragma once.
2909	*
2910	* @returns true if modifications were made, false if not.
2911	* @param pIn The input stream.
2912	* @param pOut The output stream.
2913	* @param pSettings The settings.
2914	*/
2915	bool rewrite_FixHeaderGuards(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2916	{
2917	if (!pSettings->fFixHeaderGuards)
2918	return false;
2919
2920	/* always skip .cpp.h files */
2921	size_t cchFilename = strlen(pState->pszFilename);
2922	if ( cchFilename > sizeof(".cpp.h")
2923	&& RTStrICmpAscii(&pState->pszFilename[cchFilename - sizeof(".cpp.h") + 1], ".cpp.h") == 0)
2924	return false;
2925
2926	RTERRINFOSTATIC ErrInfo;
2927	char szNormalized[168];
2928	size_t cchNormalized = 0;
2929	int rc;
2930	bool fRet = false;
2931
2932	/*
2933	* Calculate the expected guard for this file, if so tasked.
2934	* ASSUMES pState->pszFilename is absolute as is pSettings->pszGuardRelativeToDir.
2935	*/
2936	szNormalized[0] = '\0';
2937	if (pSettings->pszGuardRelativeToDir)
2938	{
2939	rc = RTStrCopy(szNormalized, sizeof(szNormalized), pSettings->pszGuardPrefix);
2940	if (RT_FAILURE(rc))
2941	return ScmError(pState, rc, "Guard prefix too long (or something): %s\n", pSettings->pszGuardPrefix);
2942	cchNormalized = strlen(szNormalized);
2943	if (strcmp(pSettings->pszGuardRelativeToDir, "{dir}") == 0)
2944	rc = RTStrCopy(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized,
2945	RTPathFilename(pState->pszFilename));
2946	else if (strcmp(pSettings->pszGuardRelativeToDir, "{parent}") == 0)
2947	{
2948	const char *pszSrc = RTPathFilename(pState->pszFilename);
2949	if (!pszSrc \|\| (uintptr_t)&pszSrc[-2] < (uintptr_t)pState->pszFilename \|\| !RTPATH_IS_SLASH(pszSrc[-1]))
2950	return ScmError(pState, VERR_INTERNAL_ERROR, "Error calculating {parent} header guard!\n");
2951	pszSrc -= 2;
2952	while ( (uintptr_t)pszSrc > (uintptr_t)pState->pszFilename
2953	&& !RTPATH_IS_SLASH(pszSrc[-1])
2954	&& !RTPATH_IS_VOLSEP(pszSrc[-1]))
2955	pszSrc--;
2956	rc = RTStrCopy(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized, pszSrc);
2957	}
2958	else
2959	rc = RTPathCalcRelative(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized,
2960	pSettings->pszGuardRelativeToDir, false /fFromFile/, pState->pszFilename);
2961	if (RT_FAILURE(rc))
2962	return ScmError(pState, rc, "Error calculating guard prefix (RTPathCalcRelative): %Rrc\n", rc);
2963	char ch;
2964	while ((ch = szNormalized[cchNormalized]) != '\0')
2965	{
2966	if (!ScmIsCIdentifierChar(ch))
2967	szNormalized[cchNormalized] = '_';
2968	cchNormalized++;
2969	}
2970	}
2971
2972	/*
2973	* First part looks for the #ifndef xxxx paired with #define xxxx.
2974	*
2975	* We blindly assume the first preprocessor directive in the file is the guard
2976	* and will be upset if this isn't the case.
2977	*/
2978	RTSTRTUPLE Guard = { NULL, 0 };
2979	uint32_t cBlankLines = 0;
2980	SCMEOL enmEol;
2981	size_t cchLine;
2982	const char *pchLine;
2983	for (;;)
2984	{
2985	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2986	if (pchLine == NULL)
2987	return ScmError(pState, VERR_PARSE_ERROR, "Did not find any include guards!\n");
2988	if (cchLine >= 2)
2989	{
2990	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2991	if ( pchHash
2992	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
2993	{
2994	/* #ifndef xxxx */
2995	static const SCMMATCHWORD s_aIfndefGuard[] =
2996	{
2997	{ RT_STR_TUPLE("#"), 0, true, false },
2998	{ RT_STR_TUPLE("ifndef"), 0, true, false },
2999	{ RT_STR_TUPLE("IDENTIFIER"), 1, true, true },
3000	{ RT_STR_TUPLE(""), 0, true, false },
3001	};
3002	rc = ScmMatchWords(pchLine, cchLine, s_aIfndefGuard, RT_ELEMENTS(s_aIfndefGuard),
3003	NULL /poffNext/, &Guard, RTErrInfoInitStatic(&ErrInfo));
3004	if (RT_FAILURE(rc))
3005	return ScmError(pState, rc, "%u: Expected first preprocessor directive to be '#ifndef xxxx'. %s (%.*s)\n",
3006	ScmStreamTellLine(pIn) - 1, ErrInfo.Core.pszMsg, cchLine, pchLine);
3007	fRet \|= rc != VINF_SUCCESS;
3008	ScmVerbose(pState, 3, "line %u in %s: #ifndef %.*s\n",
3009	ScmStreamTellLine(pIn) - 1, pState->pszFilename, Guard.cch, Guard.psz);
3010
3011	/* #define xxxx */
3012	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3013	if (!pchLine)
3014	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef %.*s'\n",
3015	ScmStreamTellLine(pIn) - 1, Guard.cch, Guard.psz);
3016	const SCMMATCHWORD aDefineGuard[] =
3017	{
3018	{ RT_STR_TUPLE("#"), 0, true, false },
3019	{ RT_STR_TUPLE("define"), 0, true, false },
3020	{ Guard.psz, (uint8_t)Guard.cch, 1, true, false },
3021	{ RT_STR_TUPLE(""), 0, true, false },
3022	};
3023	rc = ScmMatchWords(pchLine, cchLine, aDefineGuard, RT_ELEMENTS(aDefineGuard),
3024	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3025	if (RT_FAILURE(rc))
3026	return ScmError(pState, rc, "%u: Expected '#define %.s' to follow '#ifndef %.s'. %s (%.*s)\n",
3027	ScmStreamTellLine(pIn) - 1, Guard.cch, Guard.psz, Guard.cch, Guard.psz,
3028	ErrInfo.Core.pszMsg, cchLine, pchLine);
3029	fRet \|= rc != VINF_SUCCESS;
3030
3031	if (Guard.cch >= sizeof(szNormalized))
3032	return ScmError(pState, VERR_BUFFER_OVERFLOW, "%u: Guard macro too long! %.*s\n",
3033	ScmStreamTellLine(pIn) - 2, Guard.cch, Guard.psz);
3034
3035	if (szNormalized[0] != '\0')
3036	{
3037	if ( Guard.cch != cchNormalized
3038	\|\| memcmp(Guard.psz, szNormalized, cchNormalized) != 0)
3039	{
3040	ScmVerbose(pState, 2, "guard changed from %.*s to %s\n", Guard.cch, Guard.psz, szNormalized);
3041	ScmVerbose(pState, 2, "grep -rw %.*s ${WCROOT} \| grep -Fv %s\n",
3042	Guard.cch, Guard.psz, pState->pszFilename);
3043	fRet = true;
3044	}
3045	Guard.psz = szNormalized;
3046	Guard.cch = cchNormalized;
3047	}
3048
3049	/*
3050	* Write guard, making sure we've got a single blank line preceeding it.
3051	*/
3052	ScmStreamPutEol(pOut, enmEol);
3053	ScmStreamWrite(pOut, RT_STR_TUPLE("#ifndef "));
3054	ScmStreamWrite(pOut, Guard.psz, Guard.cch);
3055	ScmStreamPutEol(pOut, enmEol);
3056	ScmStreamWrite(pOut, RT_STR_TUPLE("#define "));
3057	ScmStreamWrite(pOut, Guard.psz, Guard.cch);
3058	rc = ScmStreamPutEol(pOut, enmEol);
3059	if (RT_FAILURE(rc))
3060	return false;
3061	break;
3062	}
3063	}
3064
3065	if (!isBlankLine(pchLine, cchLine))
3066	{
3067	while (cBlankLines-- > 0)
3068	ScmStreamPutEol(pOut, enmEol);
3069	cBlankLines = 0;
3070	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3071	if (RT_FAILURE(rc))
3072	return false;
3073	}
3074	else
3075	cBlankLines++;
3076	}
3077
3078	/*
3079	* Look for pragma once wrapped in #ifndef RT_WITHOUT_PRAGMA_ONCE.
3080	*/
3081	size_t const iPragmaOnce = ScmStreamTellLine(pIn);
3082	static const SCMMATCHWORD s_aIfndefRtWithoutPragmaOnce[] =
3083	{
3084	{ RT_STR_TUPLE("#"), 0, true, false },
3085	{ RT_STR_TUPLE("ifndef"), 0, true, false },
3086	{ RT_STR_TUPLE("RT_WITHOUT_PRAGMA_ONCE"), 1, true, false },
3087	{ RT_STR_TUPLE(""), 0, true, false },
3088	};
3089	static const SCMMATCHWORD s_aPragmaOnce[] =
3090	{
3091	{ RT_STR_TUPLE("#"), 0, true, false },
3092	{ RT_STR_TUPLE("pragma"), 1, true, false },
3093	{ RT_STR_TUPLE("once"), 1, true, false},
3094	{ RT_STR_TUPLE(""), 0, true, false },
3095	};
3096	static const SCMMATCHWORD s_aEndif[] =
3097	{
3098	{ RT_STR_TUPLE("#"), 0, true, false },
3099	{ RT_STR_TUPLE("endif"), 0, true, false },
3100	{ RT_STR_TUPLE(""), 0, true, false },
3101	};
3102
3103	/* #ifndef RT_WITHOUT_PRAGMA_ONCE */
3104	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3105	if (!pchLine)
3106	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after header guard!\n", iPragmaOnce + 1);
3107	size_t offNext;
3108	rc = ScmMatchWords(pchLine, cchLine, s_aIfndefRtWithoutPragmaOnce, RT_ELEMENTS(s_aIfndefRtWithoutPragmaOnce),
3109	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3110	if (RT_SUCCESS(rc))
3111	{
3112	fRet \|= rc != VINF_SUCCESS;
3113	if (offNext != cchLine)
3114	return ScmError(pState, VERR_PARSE_ERROR, "%u: Characters trailing '#ifndef RT_WITHOUT_PRAGMA_ONCE' (%.*s)\n",
3115	iPragmaOnce + 1, cchLine, pchLine);
3116
3117	/* # pragma once */
3118	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3119	if (!pchLine)
3120	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef RT_WITHOUT_PRAGMA_ONCE'\n",
3121	iPragmaOnce + 2);
3122	rc = ScmMatchWords(pchLine, cchLine, s_aPragmaOnce, RT_ELEMENTS(s_aPragmaOnce),
3123	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3124	if (RT_SUCCESS(rc))
3125	fRet \|= rc != VINF_SUCCESS;
3126	else
3127	return ScmError(pState, rc, "%u: Expected '# pragma once' to follow '#ifndef RT_WITHOUT_PRAGMA_ONCE'! %s (%.*s)\n",
3128	iPragmaOnce + 2, ErrInfo.Core.pszMsg, cchLine, pchLine);
3129
3130	/* #endif */
3131	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3132	if (!pchLine)
3133	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef RT_WITHOUT_PRAGMA_ONCE' and '#pragma once'\n",
3134	iPragmaOnce + 3);
3135	rc = ScmMatchWords(pchLine, cchLine, s_aEndif, RT_ELEMENTS(s_aEndif),
3136	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3137	if (RT_SUCCESS(rc))
3138	fRet \|= rc != VINF_SUCCESS;
3139	else
3140	return ScmError(pState, rc,
3141	"%u: Expected '#endif' to follow '#ifndef RT_WITHOUT_PRAGMA_ONCE' and '# pragma once'! %s (%.*s)\n",
3142	iPragmaOnce + 3, ErrInfo.Core.pszMsg, cchLine, pchLine);
3143	ScmVerbose(pState, 3, "Found pragma once\n");
3144	fRet \|= !pSettings->fPragmaOnce;
3145	}
3146	else
3147	{
3148	rc = ScmStreamSeekByLine(pIn, iPragmaOnce);
3149	if (RT_FAILURE(rc))
3150	return ScmError(pState, rc, "seek error\n");
3151	fRet \|= pSettings->fPragmaOnce;
3152	ScmVerbose(pState, pSettings->fPragmaOnce ? 2 : 3, "Missing #pragma once\n");
3153	}
3154
3155	/*
3156	* Write the pragma once stuff.
3157	*/
3158	if (pSettings->fPragmaOnce)
3159	{
3160	ScmStreamPutLine(pOut, RT_STR_TUPLE("#ifndef RT_WITHOUT_PRAGMA_ONCE"), enmEol);
3161	ScmStreamPutLine(pOut, RT_STR_TUPLE("# pragma once"), enmEol);
3162	rc = ScmStreamPutLine(pOut, RT_STR_TUPLE("#endif"), enmEol);
3163	if (RT_FAILURE(rc))
3164	return false;
3165	}
3166
3167	/*
3168	* Copy the rest of the file and remove pragma once statements, while
3169	* looking for the last #endif in the file.
3170	*/
3171	size_t iEndIfIn = 0;
3172	size_t iEndIfOut = 0;
3173	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3174	{
3175	if (cchLine > 2)
3176	{
3177	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
3178	if ( pchHash
3179	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
3180	{
3181	size_t off = pchHash - pchLine + 1;
3182	while (off < cchLine && RT_C_IS_BLANK(pchLine[off]))
3183	off++;
3184	/* #pragma once */
3185	if ( off + sizeof("pragma") - 1 <= cchLine
3186	&& !memcmp(&pchLine[off], RT_STR_TUPLE("pragma")))
3187	{
3188	rc = ScmMatchWords(pchLine, cchLine, s_aPragmaOnce, RT_ELEMENTS(s_aPragmaOnce),
3189	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3190	if (RT_SUCCESS(rc))
3191	{
3192	fRet = true;
3193	continue;
3194	}
3195	}
3196	/* #endif */
3197	else if ( off + sizeof("endif") - 1 <= cchLine
3198	&& !memcmp(&pchLine[off], RT_STR_TUPLE("endif")))
3199	{
3200	iEndIfIn = ScmStreamTellLine(pIn) - 1;
3201	iEndIfOut = ScmStreamTellLine(pOut);
3202	}
3203	}
3204	}
3205
3206	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3207	if (RT_FAILURE(rc))
3208	return false;
3209	}
3210
3211	/*
3212	* Check out the last endif, making sure it's well formed and make sure it has the
3213	* right kind of comment following it.
3214	*/
3215	if (pSettings->fFixHeaderGuardEndif)
3216	{
3217	if (iEndIfOut == 0)
3218	return ScmError(pState, VERR_PARSE_ERROR, "Expected '#endif' at the end of the file...\n");
3219	rc = ScmStreamSeekByLine(pIn, iEndIfIn);
3220	if (RT_FAILURE(rc))
3221	return false;
3222	rc = ScmStreamSeekByLine(pOut, iEndIfOut);
3223	if (RT_FAILURE(rc))
3224	return false;
3225
3226	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3227	if (!pchLine)
3228	return ScmError(pState, VERR_INTERNAL_ERROR, "ScmStreamGetLine failed re-reading #endif!\n");
3229
3230	char szTmp[64 + sizeof(szNormalized)];
3231	size_t cchTmp;
3232	if (pSettings->fEndifGuardComment)
3233	cchTmp = RTStrPrintf(szTmp, sizeof(szTmp), "#endif /* !%.s /", Guard.cch, Guard.psz);
3234	else
3235	cchTmp = RTStrPrintf(szTmp, sizeof(szTmp), "#endif"); /* lazy bird */
3236	fRet \|= cchTmp != cchLine \|\| memcmp(szTmp, pchLine, cchTmp) != 0;
3237	rc = ScmStreamPutLine(pOut, szTmp, cchTmp, enmEol);
3238	if (RT_FAILURE(rc))
3239	return false;
3240
3241	/* Copy out the remaining lines (assumes no #pragma once here). */
3242	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3243	{
3244	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3245	if (RT_FAILURE(rc))
3246	return false;
3247	}
3248	}
3249
3250	return fRet;
3251	}
3252
3253
3254	/**
3255	* Checks for PAGE_SIZE, PAGE_SHIFT and PAGE_OFFSET_MASK w/o a GUEST_ or HOST_
3256	* prefix as well as banning PAGE_BASE_HC_MASK, PAGE_BASE_GC_MASK and
3257	* PAGE_BASE_MASK.
3258	*
3259	* @returns true if modifications were made, false if not.
3260	* @param pIn The input stream.
3261	* @param pOut The output stream.
3262	* @param pSettings The settings.
3263	*/
3264	bool rewrite_PageChecks(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
3265	{
3266	RT_NOREF(pOut);
3267	if (!pSettings->fOnlyGuestHostPage && !pSettings->fNoASMMemPageUse)
3268	return false;
3269
3270	static RTSTRTUPLE const g_aWords[] =
3271	{
3272	{ RT_STR_TUPLE("PAGE_SIZE") },
3273	{ RT_STR_TUPLE("PAGE_SHIFT") },
3274	{ RT_STR_TUPLE("PAGE_OFFSET_MASK") },
3275	{ RT_STR_TUPLE("PAGE_BASE_MASK") },
3276	{ RT_STR_TUPLE("PAGE_BASE_GC_MASK") },
3277	{ RT_STR_TUPLE("PAGE_BASE_HC_MASK") },
3278	{ RT_STR_TUPLE("PAGE_ADDRESS") },
3279	{ RT_STR_TUPLE("PHYS_PAGE_ADDRESS") },
3280	{ RT_STR_TUPLE("ASMMemIsZeroPage") },
3281	{ RT_STR_TUPLE("ASMMemZeroPage") },
3282	};
3283	size_t const iFirstWord = pSettings->fOnlyGuestHostPage ? 0 : 7;
3284	size_t const iEndWords = pSettings->fNoASMMemPageUse ? 9 : 7;
3285
3286	uint32_t iLine = 0;
3287	SCMEOL enmEol;
3288	size_t cchLine;
3289	const char *pchLine;
3290	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3291	{
3292	iLine++;
3293	for (size_t i = iFirstWord; i < iEndWords; i++)
3294	{
3295	size_t const cchWord = g_aWords[i].cch;
3296	if (cchLine >= cchWord)
3297	{
3298	const char * const pszWord = g_aWords[i].psz;
3299	const char pchHit = (const char )memchr(pchLine, *pszWord, cchLine);
3300	while (pchHit)
3301	{
3302	size_t cchLeft = (uintptr_t)&pchLine[cchLine] - (uintptr_t)pchHit;
3303	if ( cchLeft >= cchWord
3304	&& memcmp(pchHit, pszWord, cchWord) == 0
3305	&& ( pchHit == pchLine
3306	\|\| !ScmIsCIdentifierChar(pchHit[-1]))
3307	&& ( cchLeft == cchWord
3308	\|\| !ScmIsCIdentifierChar(pchHit[cchWord])) )
3309	{
3310	if (i < 3)
3311	ScmFixManually(pState, "%u:%zu: %s is not allow! Use GUEST_%s or HOST_%s instead.\n",
3312	iLine, pchHit - pchLine + 1, pszWord, pszWord, pszWord);
3313	else if (i < 7)
3314	ScmFixManually(pState, "%u:%zu: %s is not allow! Rewrite using GUEST/HOST_PAGE_OFFSET_MASK.\n",
3315	iLine, pchHit - pchLine + 1, pszWord);
3316	else
3317	ScmFixManually(pState, "%u:%zu: %s is not allow! Use %s with correct page size instead.\n",
3318	iLine, pchHit - pchLine + 1, pszWord, i == 3 ? "ASMMemIsZero" : "RT_BZERO");
3319	}
3320
3321	/* next */
3322	cchLeft -= 1;
3323	if (cchLeft < cchWord)
3324	break;
3325	pchHit = (const char )memchr(pchHit + 1, pszWord, cchLeft);
3326	}
3327	}
3328	}
3329	}
3330
3331	return false;
3332	}
3333
3334
3335	/**
3336	* Checks for usage of rc in code instead of vrc for IPRT status codes (int) and hrc for COM
3337	* status codes (HRESULT).
3338	*
3339	* @returns true if modifications were made, false if not.
3340	* @param pIn The input stream.
3341	* @param pOut The output stream.
3342	* @param pSettings The settings.
3343	*
3344	* @note Used in Main to avoid ambiguity when just using rc.
3345	*/
3346	bool rewrite_ForceHrcVrcInsteadOfRc(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
3347	{
3348	RT_NOREF(pOut);
3349	if (!pSettings->fOnlyHrcVrcInsteadOfRc)
3350	return false;
3351
3352	static const SCMMATCHWORD s_aHresultVrc[] =
3353	{
3354	{ RT_STR_TUPLE("HRESULT"), 0, true, false },
3355	{ RT_STR_TUPLE("vrc"), 1, true, false }
3356	};
3357
3358	static const SCMMATCHWORD s_aIntHrc[] =
3359	{
3360	{ RT_STR_TUPLE("int"), 0, true, false },
3361	{ RT_STR_TUPLE("hrc"), 1, true, false }
3362	};
3363
3364	uint32_t iLine = 0;
3365	SCMEOL enmEol;
3366	size_t cchLine;
3367	const char *pchLine;
3368	RTERRINFOSTATIC ErrInfo;
3369	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3370	{
3371	iLine++;
3372
3373	/* Look for forbidden declarations first. */
3374	size_t offNext = 0;
3375	int rc = ScmMatchWords(pchLine, cchLine, s_aHresultVrc, RT_ELEMENTS(s_aHresultVrc),
3376	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3377	if (RT_SUCCESS(rc))
3378	{
3379	ScmFixManually(pState, "%u:%zu: 'HRESULT vrc' is not allowed! Use 'HRESULT hrc' instead.\n",
3380	iLine, offNext);
3381	continue;
3382	}
3383
3384	rc = ScmMatchWords(pchLine, cchLine, s_aIntHrc, RT_ELEMENTS(s_aIntHrc),
3385	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3386	if (RT_SUCCESS(rc))
3387	{
3388	ScmFixManually(pState, "%u:%zu: 'int hrc' is not allowed! Use 'int vrc' instead.\n",
3389	iLine, offNext);
3390	continue;
3391	}
3392
3393	#if 0 /* This is too broad and triggers on things we don't want to trigger on (like autoCaller.rc()). */
3394	const RTSTRTUPLE RcTuple = { RT_STR_TUPLE("rc") };
3395	size_t const cchWord = RcTuple.cch;
3396	if (cchLine >= cchWord)
3397	{
3398	const char pchHit = (const char )memchr(pchLine, *RcTuple.psz, cchLine);
3399	while (pchHit)
3400	{
3401	size_t cchLeft = (uintptr_t)&pchLine[cchLine] - (uintptr_t)pchHit;
3402	if ( cchLeft >= cchWord
3403	&& memcmp(pchHit, RcTuple.psz, cchWord) == 0
3404	&& ( pchHit == pchLine
3405	\|\| !ScmIsCIdentifierChar(pchHit[-1]))
3406	&& ( cchLeft == cchWord
3407	\|\| !ScmIsCIdentifierChar(pchHit[cchWord])) )
3408	ScmFixManually(pState, "%u:%zu: %s is not allowed! Use hrc or vrc instead.\n",
3409	iLine, pchHit - pchLine + 1, RcTuple.psz);
3410
3411	/* next */
3412	cchLeft -= 1;
3413	if (cchLeft < cchWord)
3414	break;
3415	pchHit = (const char )memchr(pchHit + 1, RcTuple.psz, cchLeft);
3416	}
3417	}
3418	#else
3419	/* Trigger on declarations of 'HRESULT rc' and 'int rc'. */
3420	static const SCMMATCHWORD s_aHresultRc[] =
3421	{
3422	{ RT_STR_TUPLE("HRESULT"), 0, true, false },
3423	{ RT_STR_TUPLE("rc"), 1, true, false }
3424	};
3425
3426	static const SCMMATCHWORD s_aIntRc[] =
3427	{
3428	{ RT_STR_TUPLE("int"), 0, true, false },
3429	{ RT_STR_TUPLE("rc"), 1, true, false }
3430	};
3431
3432	rc = ScmMatchWords(pchLine, cchLine, s_aHresultRc, RT_ELEMENTS(s_aHresultRc),
3433	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3434	if (RT_SUCCESS(rc))
3435	{
3436	ScmFixManually(pState, "%u:%zu: 'HRESULT rc' is not allowed! Use 'HRESULT hrc' instead.\n",
3437	iLine, offNext);
3438	continue;
3439	}
3440
3441	rc = ScmMatchWords(pchLine, cchLine, s_aIntRc, RT_ELEMENTS(s_aIntRc),
3442	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3443	if (RT_SUCCESS(rc))
3444	{
3445	ScmFixManually(pState, "%u:%zu: 'int rc' is not allowed! Use 'int vrc' instead.\n",
3446	iLine, offNext);
3447	continue;
3448	}
3449	#endif
3450	}
3451
3452	return false;
3453	}
3454
3455
3456	/**
3457	* Rewrite a C/C++ source or header file.
3458	*
3459	* @returns true if modifications were made, false if not.
3460	* @param pIn The input stream.
3461	* @param pOut The output stream.
3462	* @param pSettings The settings.
3463	*
3464	* @todo
3465	*
3466	* Ideas for C/C++:
3467	* - space after if, while, for, switch
3468	* - spaces in for (i=0;i<x;i++)
3469	* - complex conditional, bird style.
3470	* - remove unnecessary parentheses.
3471	* - sort defined RT_OS_*\|\| and RT_ARCH
3472	* - sizeof without parenthesis.
3473	* - defined without parenthesis.
3474	* - trailing spaces.
3475	* - parameter indentation.
3476	* - space after comma.
3477	* - while (x--); -> multi line + comment.
3478	* - else statement;
3479	* - space between function and left parenthesis.
3480	* - TODO, XXX, @todo cleanup.
3481	* - Space before/after '*'.
3482	* - ensure new line at end of file.
3483	* - Indentation of precompiler statements (#ifdef, #defines).
3484	* - space between functions.
3485	* - string.h -> iprt/string.h, stdarg.h -> iprt/stdarg.h, etc.
3486	*/
3487	bool rewrite_C_and_CPP(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
3488	{
3489
3490	RT_NOREF4(pState, pIn, pOut, pSettings);
3491	return false;
3492	}
3493

Note: See TracBrowser for help on using the repository browser.

source: vbox/trunk/src/bldprogs/scmrw.cpp@ 94913

Download in other formats: