scmrw.cpp@ 92709

Last change on this file since 92709 was 92185, checked in by vboxsync, 3 years ago
scm: Check UTF-8 encoding and reject files with bidirectional control codes in them.
Property svn:eol-style set to `native` Property svn:keywords set to `Author Date Id Revision`
File size: 130.0 KB

Line
1	/* $Id: scmrw.cpp 92185 2021-11-02 21:52:35Z vboxsync $ */
2	/** @file
3	* IPRT Testcase / Tool - Source Code Massager.
4	*/
5
6	/*
7	* Copyright (C) 2010-2020 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	#if 1
1118	ScmError(pState, rc, "%u:%zu: Evil unicode codepoint: %s\n", iLine, pchCur - pchLine, pszWhat);
1119	#else
1120	ScmVerbose(pState, 0, "%u:%zu: Evil unicode codepoint: %s\n", iLine, pchCur - pchLine, pszWhat);
1121	#endif
1122	}
1123	else
1124	#if 1
1125	ScmError(pState, rc, "%u:%zu: Invalid UTF-8 encoding: %Rrc\n", iLine, pchCur - pchLine, rc);
1126	#else
1127	ScmVerbose(pState, 0, "%u:%zu: Invalid UTF-8 encoding: %Rrc\n", iLine, pchCur - pchLine, rc);
1128	#endif
1129	}
1130	}
1131
1132	return false;
1133	}
1134
1135	/**
1136	* Compares two strings word-by-word, ignoring spaces, punctuation and case.
1137	*
1138	* Assumes ASCII strings.
1139	*
1140	* @returns true if they match, false if not.
1141	* @param psz1 The first string. This is typically the known one.
1142	* @param psz2 The second string. This is typically the unknown one,
1143	* which is why we return a next pointer for this one.
1144	* @param ppsz2Next Where to return the next part of the 2nd string. If
1145	* this is NULL, the whole string must match.
1146	*/
1147	static bool IsEqualWordByWordIgnoreCase(const char psz1, const char psz2, const char **ppsz2Next)
1148	{
1149	for (;;)
1150	{
1151	/* Try compare raw strings first. */
1152	char ch1 = *psz1;
1153	char ch2 = *psz2;
1154	if ( ch1 == ch2
1155	\|\| RT_C_TO_LOWER(ch1) == RT_C_TO_LOWER(ch2))
1156	{
1157	if (ch1)
1158	{
1159	psz1++;
1160	psz2++;
1161	}
1162	else
1163	{
1164	if (ppsz2Next)
1165	*ppsz2Next = psz2;
1166	return true;
1167	}
1168	}
1169	else
1170	{
1171	/* Try skip spaces an punctuation. */
1172	while ( RT_C_IS_SPACE(ch1)
1173	\|\| RT_C_IS_PUNCT(ch1))
1174	ch1 = *++psz1;
1175
1176	if (ch1 == '\0' && ppsz2Next)
1177	{
1178	*ppsz2Next = psz2;
1179	return true;
1180	}
1181
1182	while ( RT_C_IS_SPACE(ch2)
1183	\|\| RT_C_IS_PUNCT(ch2))
1184	ch2 = *++psz2;
1185
1186	if ( ch1 != ch2
1187	&& RT_C_TO_LOWER(ch1) != RT_C_TO_LOWER(ch2))
1188	{
1189	if (ppsz2Next)
1190	*ppsz2Next = psz2;
1191	return false;
1192	}
1193	}
1194	}
1195	}
1196
1197	/**
1198	* Looks for @a pszFragment anywhere in @a pszText, ignoring spaces, punctuation
1199	* and case.
1200	*
1201	* @returns true if found, false if not.
1202	* @param pszText The haystack to search in.
1203	* @param cchText The length @a pszText.
1204	* @param pszFragment The needle to search for.
1205	* @param ppszStart Where to return the address in @a pszText where
1206	* the fragment was found. Optional.
1207	* @param ppszNext Where to return the pointer to the first char in
1208	* @a pszText after the fragment. Optional.
1209	*
1210	* @remarks First character of @a pszFragment must be an 7-bit ASCII character!
1211	* This character must not be space or punctuation.
1212	*/
1213	static bool scmContainsWordByWordIgnoreCase(const char pszText, size_t cchText, const char pszFragment,
1214	const char ppszStart, const char ppszNext)
1215	{
1216	Assert(!((unsigned)*pszFragment & 0x80));
1217	Assert(pszText[cchText] == '\0');
1218	Assert(!RT_C_IS_BLANK(*pszFragment));
1219	Assert(!RT_C_IS_PUNCT(*pszFragment));
1220
1221	char chLower = RT_C_TO_LOWER(*pszFragment);
1222	char chUpper = RT_C_TO_UPPER(*pszFragment);
1223	for (;;)
1224	{
1225	const char pszHit = (const char )memchr(pszText, chLower, cchText);
1226	const char pszHit2 = (const char )memchr(pszText, chUpper, cchText);
1227	if (!pszHit && !pszHit2)
1228	{
1229	if (ppszStart)
1230	*ppszStart = NULL;
1231	if (ppszNext)
1232	*ppszNext = NULL;
1233	return false;
1234	}
1235
1236	if ( pszHit == NULL
1237	\|\| ( pszHit2 != NULL
1238	&& ((uintptr_t)pszHit2 < (uintptr_t)pszHit)) )
1239	pszHit = pszHit2;
1240
1241	const char *pszNext;
1242	if (IsEqualWordByWordIgnoreCase(pszFragment, pszHit, &pszNext))
1243	{
1244	if (ppszStart)
1245	*ppszStart = pszHit;
1246	if (ppszNext)
1247	*ppszNext = pszNext;
1248	return true;
1249	}
1250
1251	cchText -= pszHit - pszText + 1;
1252	pszText = pszHit + 1;
1253	}
1254	}
1255
1256
1257	/**
1258	* Counts the number of lines in the given substring.
1259	*
1260	* @returns The number of lines.
1261	* @param psz The start of the substring.
1262	* @param cch The length of the substring.
1263	*/
1264	static uint32_t CountLinesInSubstring(const char *psz, size_t cch)
1265	{
1266	uint32_t cLines = 0;
1267	for (;;)
1268	{
1269	const char pszEol = (const char )memchr(psz, '\n', cch);
1270	if (pszEol)
1271	cLines++;
1272	else
1273	return cLines + (*psz != '\0');
1274	cch -= pszEol + 1 - psz;
1275	if (!cch)
1276	return cLines;
1277	psz = pszEol + 1;
1278	}
1279	}
1280
1281
1282	/**
1283	* Comment parser callback for locating copyright and license.
1284	*/
1285	static DECLCALLBACK(int)
1286	rewrite_Copyright_CommentCallback(PCSCMCOMMENTINFO pInfo, const char pszBody, size_t cchBody, void pvUser)
1287	{
1288	PSCMCOPYRIGHTINFO pState = (PSCMCOPYRIGHTINFO)pvUser;
1289	Assert(strlen(pszBody) == cchBody);
1290	//RTPrintf("--- comment at %u, type %u ---\n%s\n--- end ---\n", pInfo->iLineStart, pInfo->enmType, pszBody);
1291	ScmVerbose(pState->pState, 5,
1292	"--- comment at %u col %u, %u lines, type %u, %u lines before body, %u lines after body\n",
1293	pInfo->iLineStart, pInfo->offStart, pInfo->iLineEnd - pInfo->iLineStart + 1, pInfo->enmType,
1294	pInfo->cBlankLinesBefore, pInfo->cBlankLinesAfter);
1295
1296	pState->cComments++;
1297
1298	uint32_t iLine = pInfo->iLineStart + pInfo->cBlankLinesBefore;
1299
1300	/*
1301	* Look for a 'contributed by' or 'includes contributions from' line, these
1302	* comes first when present.
1303	*/
1304	const char *pchContributedBy = NULL;
1305	size_t cchContributedBy = 0;
1306	size_t cBlankLinesAfterContributedBy = 0;
1307	if ( pState->pszContributedBy == NULL
1308	&& ( pState->iLineCopyright == UINT32_MAX
1309	\|\| pState->iLineLicense == UINT32_MAX)
1310	&& ( ( cchBody > sizeof("Contributed by")
1311	&& RTStrNICmp(pszBody, RT_STR_TUPLE("contributed by")) == 0)
1312	\|\| ( cchBody > sizeof("Includes contributions from")
1313	&& RTStrNICmp(pszBody, RT_STR_TUPLE("Includes contributions from")) == 0) ) )
1314	{
1315	const char pszNextLine = (const char )memchr(pszBody, '\n', cchBody);
1316	while (pszNextLine && pszNextLine[1] != '\n')
1317	pszNextLine = (const char *)memchr(pszNextLine + 1, '\n', cchBody);
1318	if (pszNextLine)
1319	{
1320	pchContributedBy = pszBody;
1321	cchContributedBy = pszNextLine - pszBody;
1322
1323	/* Skip the copyright line and any blank lines following it. */
1324	cchBody -= cchContributedBy + 1;
1325	pszBody = pszNextLine + 1;
1326	iLine += 1;
1327	while (*pszBody == '\n')
1328	{
1329	pszBody++;
1330	cchBody--;
1331	iLine++;
1332	cBlankLinesAfterContributedBy++;
1333	}
1334	}
1335	}
1336
1337	/*
1338	* Look for the copyright line.
1339	*/
1340	bool fFoundCopyright = false;
1341	uint32_t cBlankLinesAfterCopyright = 0;
1342	if ( pState->iLineCopyright == UINT32_MAX
1343	&& cchBody > sizeof("Copyright") + sizeof(g_szCopyrightHolder)
1344	&& RTStrNICmp(pszBody, RT_STR_TUPLE("copyright")) == 0)
1345	{
1346	const char pszNextLine = (const char )memchr(pszBody, '\n', cchBody);
1347
1348	/* Oracle copyright? */
1349	const char *pszEnd = pszNextLine ? pszNextLine : &pszBody[cchBody];
1350	while (RT_C_IS_SPACE(pszEnd[-1]))
1351	pszEnd--;
1352	if ( (uintptr_t)(pszEnd - pszBody) > sizeof(g_szCopyrightHolder)
1353	&& ((unsigned char )(pszEnd - sizeof(g_szCopyrightHolder) + 1) & 0x80) == 0 /* to avoid annoying assertion */
1354	&& RTStrNICmp(pszEnd - sizeof(g_szCopyrightHolder) + 1, RT_STR_TUPLE(g_szCopyrightHolder)) == 0)
1355	{
1356	/* Parse out the year(s). */
1357	const char *psz = pszBody + sizeof("copyright");
1358	while ((uintptr_t)psz < (uintptr_t)pszEnd && !RT_C_IS_DIGIT(*psz))
1359	psz++;
1360	if (RT_C_IS_DIGIT(*psz))
1361	{
1362	char *pszNext;
1363	int rc = RTStrToUInt32Ex(psz, &pszNext, 10, &pState->uFirstYear);
1364	if ( RT_SUCCESS(rc)
1365	&& rc != VWRN_NUMBER_TOO_BIG
1366	&& rc != VWRN_NEGATIVE_UNSIGNED)
1367	{
1368	if ( pState->uFirstYear < 1975
1369	\|\| pState->uFirstYear > 3000)
1370	{
1371	ScmError(pState->pState, VERR_OUT_OF_RANGE, "Copyright year is out of range: %u ('%.*s')\n",
1372	pState->uFirstYear, pszEnd - pszBody, pszBody);
1373	pState->uFirstYear = UINT32_MAX;
1374	}
1375
1376	while (RT_C_IS_SPACE(*pszNext))
1377	pszNext++;
1378	if (*pszNext == '-')
1379	{
1380	do
1381	pszNext++;
1382	while (RT_C_IS_SPACE(*pszNext));
1383	rc = RTStrToUInt32Ex(pszNext, &pszNext, 10, &pState->uLastYear);
1384	if ( RT_SUCCESS(rc)
1385	&& rc != VWRN_NUMBER_TOO_BIG
1386	&& rc != VWRN_NEGATIVE_UNSIGNED)
1387	{
1388	if ( pState->uLastYear < 1975
1389	\|\| pState->uLastYear > 3000)
1390	{
1391	ScmError(pState->pState, VERR_OUT_OF_RANGE, "Second copyright year is out of range: %u ('%.*s')\n",
1392	pState->uLastYear, pszEnd - pszBody, pszBody);
1393	pState->uLastYear = UINT32_MAX;
1394	}
1395	else if (pState->uFirstYear > pState->uLastYear)
1396	{
1397	RTMsgWarning("Copyright years switched(?): '%.*s'\n", pszEnd - pszBody, pszBody);
1398	uint32_t iTmp = pState->uLastYear;
1399	pState->uLastYear = pState->uFirstYear;
1400	pState->uFirstYear = iTmp;
1401	}
1402	}
1403	else
1404	{
1405	pState->uLastYear = UINT32_MAX;
1406	ScmError(pState->pState, RT_SUCCESS(rc) ? -rc : rc,
1407	"Failed to parse second copyright year: '%.*s'\n", pszEnd - pszBody, pszBody);
1408	}
1409	}
1410	else if (*pszNext != g_szCopyrightHolder[0])
1411	ScmError(pState->pState, VERR_PARSE_ERROR,
1412	"Failed to parse copyright: '%.*s'\n", pszEnd - pszBody, pszBody);
1413	else
1414	pState->uLastYear = pState->uFirstYear;
1415	}
1416	else
1417	{
1418	pState->uFirstYear = UINT32_MAX;
1419	ScmError(pState->pState, RT_SUCCESS(rc) ? -rc : rc,
1420	"Failed to parse copyright year: '%.*s'\n", pszEnd - pszBody, pszBody);
1421	}
1422	}
1423
1424	/* The copyright comment must come before the license. */
1425	if (pState->iLineLicense != UINT32_MAX)
1426	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright (line %u) must come before the license (line %u)!\n",
1427	iLine, pState->iLineLicense);
1428
1429	/* In C/C++ code, this must be a multiline comment. While in python it
1430	must be a */
1431	if (pState->enmCommentStyle == kScmCommentStyle_C && pInfo->enmType != kScmCommentType_MultiLine)
1432	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright must appear in a multiline comment (no doxygen stuff)\n");
1433	else if (pState->enmCommentStyle == kScmCommentStyle_Python && pInfo->enmType != kScmCommentType_DocString)
1434	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright must appear in a doc-string\n");
1435
1436	/* The copyright must be followed by the license. */
1437	if (!pszNextLine)
1438	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright should be followed by the license text!\n");
1439
1440	/* Quit if we've flagged a failure. */
1441	if (RT_FAILURE(pState->pState->rc))
1442	return VERR_CALLBACK_RETURN;
1443
1444	/* Check if it's well formed and up to date. */
1445	char szWellFormed[256];
1446	size_t cchWellFormed;
1447	if (pState->uFirstYear == pState->uLastYear)
1448	cchWellFormed = RTStrPrintf(szWellFormed, sizeof(szWellFormed), "Copyright (C) %u %s",
1449	pState->uFirstYear, g_szCopyrightHolder);
1450	else
1451	cchWellFormed = RTStrPrintf(szWellFormed, sizeof(szWellFormed), "Copyright (C) %u-%u %s",
1452	pState->uFirstYear, pState->uLastYear, g_szCopyrightHolder);
1453	pState->fUpToDateCopyright = pState->uLastYear == g_uYear;
1454	pState->iLineCopyright = iLine;
1455	pState->fWellFormedCopyright = cchWellFormed == (uintptr_t)(pszEnd - pszBody)
1456	&& memcmp(pszBody, szWellFormed, cchWellFormed) == 0;
1457	if (!pState->fWellFormedCopyright)
1458	ScmVerbose(pState->pState, 1, "* copyright isn't well formed\n");
1459
1460	/* If there wasn't exactly one blank line before the comment, trigger a rewrite. */
1461	if (pInfo->cBlankLinesBefore != 1)
1462	{
1463	ScmVerbose(pState->pState, 1, "* copyright comment is preceeded by %u blank lines instead of 1\n",
1464	pInfo->cBlankLinesBefore);
1465	pState->fWellFormedCopyright = false;
1466	}
1467
1468	/* If the comment doesn't start in column 1, trigger rewrite. */
1469	if (pInfo->offStart != 0)
1470	{
1471	ScmVerbose(pState->pState, 1, "* copyright comment starts in column %u instead of 1\n", pInfo->offStart + 1);
1472	pState->fWellFormedCopyright = false;
1473	/** @todo check that there isn't any code preceeding the comment. */
1474	}
1475
1476	if (pchContributedBy)
1477	{
1478	pState->pszContributedBy = RTStrDupN(pchContributedBy, cchContributedBy);
1479	if (cBlankLinesAfterContributedBy != 1)
1480	{
1481	ScmVerbose(pState->pState, 1, "* %u blank lines between contributed by and copyright, should be 1\n",
1482	cBlankLinesAfterContributedBy);
1483	pState->fWellFormedCopyright = false;
1484	}
1485	}
1486
1487	fFoundCopyright = true;
1488	ScmVerbose(pState->pState, 3, "oracle copyright %u-%u: up-to-date=%RTbool well-formed=%RTbool\n",
1489	pState->uFirstYear, pState->uLastYear, pState->fUpToDateCopyright, pState->fWellFormedCopyright);
1490	}
1491	else
1492	ScmVerbose(pState->pState, 3, "not oracle copyright: '%.*s'\n", pszEnd - pszBody, pszBody);
1493
1494	if (!pszNextLine)
1495	return VINF_SUCCESS;
1496
1497	/* Skip the copyright line and any blank lines following it. */
1498	cchBody -= pszNextLine - pszBody + 1;
1499	pszBody = pszNextLine + 1;
1500	iLine += 1;
1501	while (*pszBody == '\n')
1502	{
1503	pszBody++;
1504	cchBody--;
1505	iLine++;
1506	cBlankLinesAfterCopyright++;
1507	}
1508
1509	/*
1510	* If we have a based-on-mit scenario, check for the lead in now and
1511	* complain if not found.
1512	*/
1513	if ( fFoundCopyright
1514	&& pState->enmLicenceOpt == kScmLicense_BasedOnMit
1515	&& pState->iLineLicense == UINT32_MAX)
1516	{
1517	if (RTStrNICmp(pszBody, RT_STR_TUPLE("This file is based on ")) == 0)
1518	{
1519	/* Take down a comment area which goes up to 'this file is based on'.
1520	The license line and length isn't used but gets set to cover the current line. */
1521	pState->iLineComment = pInfo->iLineStart;
1522	pState->cLinesComment = iLine - pInfo->iLineStart;
1523	pState->iLineLicense = iLine;
1524	pState->cLinesLicense = 1;
1525	pState->fExternalLicense = true;
1526	pState->fIsCorrectLicense = true;
1527	pState->fWellFormedLicense = true;
1528
1529	/* Check if we've got a MIT a license here or not. */
1530	pState->pCurrentLicense = NULL;
1531	do
1532	{
1533	const char pszEol = (const char )memchr(pszBody, '\n', cchBody);
1534	if (!pszEol \|\| pszEol[1] == '\0')
1535	{
1536	pszBody += cchBody;
1537	cchBody = 0;
1538	break;
1539	}
1540	cchBody -= pszEol - pszBody + 1;
1541	pszBody = pszEol + 1;
1542	iLine++;
1543
1544	for (PCSCMLICENSETEXT pCur = pState->paLicenses; pCur->cch > 0; pCur++)
1545	{
1546	const char *pszNext;
1547	if ( pCur->cch <= cchBody + 32 /* (+ 32 since we don't compare spaces and punctuation) */
1548	&& IsEqualWordByWordIgnoreCase(pCur->psz, pszBody, &pszNext))
1549	{
1550	pState->pCurrentLicense = pCur;
1551	break;
1552	}
1553	}
1554	} while (!pState->pCurrentLicense);
1555	if (!pState->pCurrentLicense)
1556	ScmError(pState->pState, VERR_NOT_FOUND, "Could not find the based-on license!\n");
1557	else if (pState->pCurrentLicense->enmType != kScmLicenseType_Mit)
1558	ScmError(pState->pState, VERR_NOT_FOUND, "The based-on license is not MIT (%.32s...)\n",
1559	pState->pCurrentLicense->psz);
1560	}
1561	else
1562	ScmError(pState->pState, VERR_WRONG_ORDER, "Expected 'This file is based on ...' after our copyright!\n");
1563	return VINF_SUCCESS;
1564	}
1565	}
1566
1567	/*
1568	* Look for LGPL like text in the comment.
1569	*/
1570	if (pState->fCheckforLgpl && cchBody > 128)
1571	{
1572	/* We look for typical LGPL notices. */
1573	if (pState->iLineLgplNotice == UINT32_MAX)
1574	{
1575	static const char * const s_apszFragments[] =
1576	{
1577	"under the terms of the GNU Lesser General Public License",
1578	};
1579	for (unsigned i = 0; i < RT_ELEMENTS(s_apszFragments); i++)
1580	if (scmContainsWordByWordIgnoreCase(pszBody, cchBody, s_apszFragments[i], NULL, NULL))
1581	{
1582	pState->iLineLgplNotice = iLine;
1583	pState->iLineAfterLgplComment = pInfo->iLineEnd + 1;
1584	ScmVerbose(pState->pState, 3, "Found LGPL notice at %u\n", iLine);
1585	break;
1586	}
1587	}
1588
1589	if ( pState->iLineLgplDisclaimer == UINT32_MAX
1590	&& scmContainsWordByWordIgnoreCase(pszBody, cchBody, g_szLgplDisclaimer, NULL, NULL))
1591	{
1592	pState->iLineLgplDisclaimer = iLine;
1593	ScmVerbose(pState->pState, 3, "Found LGPL disclaimer at %u\n", iLine);
1594	}
1595	}
1596
1597	/*
1598	* Look for the license text.
1599	*/
1600	if (pState->iLineLicense == UINT32_MAX)
1601	{
1602	for (PCSCMLICENSETEXT pCur = pState->paLicenses; pCur->cch > 0; pCur++)
1603	{
1604	const char *pszNext;
1605	if ( pCur->cch <= cchBody + 32 /* (+ 32 since we don't compare spaces and punctuation) */
1606	&& IsEqualWordByWordIgnoreCase(pCur->psz, pszBody, &pszNext))
1607	{
1608	while ( RT_C_IS_SPACE(*pszNext)
1609	\|\| (RT_C_IS_PUNCT(pszNext) && pszNext != '-'))
1610	pszNext++;
1611
1612	uint32_t cDashes = 0;
1613	while (*pszNext == '-')
1614	cDashes++, pszNext++;
1615	bool fExternal = cDashes > 10;
1616
1617	if ( *pszNext == '\0'
1618	\|\| fExternal)
1619	{
1620	/* In C/C++ code, this must be a multiline comment. While in python it
1621	must be a */
1622	if (pState->enmCommentStyle == kScmCommentStyle_C && pInfo->enmType != kScmCommentType_MultiLine)
1623	ScmError(pState->pState, VERR_WRONG_ORDER, "License must appear in a multiline comment (no doxygen stuff)\n");
1624	else if (pState->enmCommentStyle == kScmCommentStyle_Python && pInfo->enmType != kScmCommentType_DocString)
1625	ScmError(pState->pState, VERR_WRONG_ORDER, "License must appear in a doc-string\n");
1626
1627	/* Quit if we've flagged a failure. */
1628	if (RT_FAILURE(pState->pState->rc))
1629	return VERR_CALLBACK_RETURN;
1630
1631	/* Record it. */
1632	pState->iLineLicense = iLine;
1633	pState->cLinesLicense = CountLinesInSubstring(pszBody, pszNext - pszBody) - fExternal;
1634	pState->pCurrentLicense = pCur;
1635	pState->fExternalLicense = fExternal;
1636	pState->fIsCorrectLicense = pState->fOpenSource
1637	? pCur == pState->pExpectedLicense
1638	: pCur->enmType == kScmLicenseType_Confidential;
1639	pState->fWellFormedLicense = memcmp(pszBody, pCur->psz, pCur->cch - 1) == 0;
1640	if (!pState->fWellFormedLicense)
1641	ScmVerbose(pState->pState, 1, "* license text isn't well-formed\n");
1642
1643	/* If there was more than one blank line between the copyright and the
1644	license text, extend the license text area and force a rewrite of it. */
1645	if (cBlankLinesAfterCopyright > 1)
1646	{
1647	ScmVerbose(pState->pState, 1, "* %u blank lines between copyright and license text, instead of 1\n",
1648	cBlankLinesAfterCopyright);
1649	pState->iLineLicense -= cBlankLinesAfterCopyright - 1;
1650	pState->cLinesLicense += cBlankLinesAfterCopyright - 1;
1651	pState->fWellFormedLicense = false;
1652	}
1653
1654	/* If there was more than one blank line after the license, trigger a rewrite. */
1655	if (!fExternal && pInfo->cBlankLinesAfter != 1)
1656	{
1657	ScmVerbose(pState->pState, 1, "* copyright comment is followed by %u blank lines instead of 1\n",
1658	pInfo->cBlankLinesAfter);
1659	pState->fWellFormedLicense = false;
1660	}
1661
1662	/** @todo Check that the last comment line doesn't have any code on it. */
1663	/** @todo Check that column 2 contains '' for C/C++ files. /
1664
1665	ScmVerbose(pState->pState, 3,
1666	"Found license %d/%d at %u..%u: is-correct=%RTbool well-formed=%RTbool external-part=%RTbool open-source=%RTbool\n",
1667	pCur->enmType, pCur->enmOpt, pState->iLineLicense, pState->iLineLicense + pState->cLinesLicense,
1668	pState->fIsCorrectLicense, pState->fWellFormedLicense,
1669	pState->fExternalLicense, pState->fOpenSource);
1670
1671	if (fFoundCopyright)
1672	{
1673	pState->iLineComment = pInfo->iLineStart;
1674	pState->cLinesComment = (fExternal ? pState->iLineLicense + pState->cLinesLicense : pInfo->iLineEnd + 1)
1675	- pInfo->iLineStart;
1676	}
1677	else
1678	ScmError(pState->pState, VERR_WRONG_ORDER, "License should be preceeded by the copyright!\n");
1679	break;
1680	}
1681	}
1682	}
1683	}
1684
1685	if (fFoundCopyright && pState->iLineLicense == UINT32_MAX)
1686	ScmError(pState->pState, VERR_WRONG_ORDER, "Copyright should be followed by the license text!\n");
1687
1688	/*
1689	* Stop looking for stuff after 100 comments.
1690	*/
1691	if (pState->cComments > 100)
1692	return VERR_CALLBACK_RETURN;
1693	return VINF_SUCCESS;
1694	}
1695
1696	/**
1697	* Writes comment body text.
1698	*
1699	* @returns Stream status.
1700	* @param pOut The output stream.
1701	* @param pszText The text to write.
1702	* @param cchText The length of the text.
1703	* @param enmCommentStyle The comment style.
1704	* @param enmEol The EOL style.
1705	*/
1706	static int scmWriteCommentBody(PSCMSTREAM pOut, const char *pszText, size_t cchText,
1707	SCMCOMMENTSTYLE enmCommentStyle, SCMEOL enmEol)
1708	{
1709	Assert(pszText[cchText - 1] == '\n');
1710	Assert(pszText[cchText - 2] != '\n');
1711	NOREF(cchText);
1712	do
1713	{
1714	const char *pszEol = strchr(pszText, '\n');
1715	if (pszEol != pszText)
1716	{
1717	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1718	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1719	ScmStreamWrite(pOut, pszText, pszEol - pszText);
1720	ScmStreamPutEol(pOut, enmEol);
1721	}
1722	else
1723	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1724	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1725	pszText = pszEol + 1;
1726	} while (*pszText != '\0');
1727	return ScmStreamGetStatus(pOut);
1728	}
1729
1730
1731	/**
1732	* Updates the copyright year and/or license text.
1733	*
1734	* @returns true if modifications were made, false if not.
1735	* @param pState The rewriter state.
1736	* @param pIn The input stream.
1737	* @param pOut The output stream.
1738	* @param pSettings The settings.
1739	* @param enmCommentStyle The comment style used by the file.
1740	*/
1741	static bool rewrite_Copyright_Common(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings,
1742	SCMCOMMENTSTYLE enmCommentStyle)
1743	{
1744	if ( !pSettings->fUpdateCopyrightYear
1745	&& pSettings->enmUpdateLicense == kScmLicense_LeaveAlone)
1746	return false;
1747
1748	/*
1749	* Try locate the relevant comments.
1750	*/
1751	SCMCOPYRIGHTINFO Info =
1752	{
1753	/.pState = / pState,
1754	/.enmCommentStyle = / enmCommentStyle,
1755
1756	/.cComments = / 0,
1757
1758	/.pszContributedBy = / NULL,
1759
1760	/.iLineComment = / UINT32_MAX,
1761	/.cLinesComment = / 0,
1762
1763	/.iLineCopyright = / UINT32_MAX,
1764	/.uFirstYear = / UINT32_MAX,
1765	/.uLastYear = / UINT32_MAX,
1766	/.fWellFormedCopyright = / false,
1767	/.fUpToDateCopyright = / false,
1768
1769	/.fOpenSource = / true,
1770	/.pExpectedLicense = / NULL,
1771	/.paLicenses = / pSettings->enmUpdateLicense != kScmLicense_Mit
1772	&& pSettings->enmUpdateLicense != kScmLicense_BasedOnMit
1773	? &g_aLicenses[0] : &g_aLicensesWithMit[0],
1774	/.enmLicenceOpt = / pSettings->enmUpdateLicense,
1775	/.iLineLicense = / UINT32_MAX,
1776	/.cLinesLicense = / 0,
1777	/.pCurrentLicense = / NULL,
1778	/.fIsCorrectLicense = / false,
1779	/.fWellFormedLicense = / false,
1780	/.fExternalLicense = / false,
1781
1782	/.fCheckForLgpl = / true,
1783	/.iLineLgplNotice = / UINT32_MAX,
1784	/.iLineAfterLgplComment = / UINT32_MAX,
1785	/.iLineLgplDisclaimer = / UINT32_MAX,
1786	};
1787
1788	/* Figure Info.fOpenSource and the desired license: */
1789	char *pszSyncProcess;
1790	int rc = ScmSvnQueryProperty(pState, "svn:sync-process", &pszSyncProcess);
1791	if (RT_SUCCESS(rc))
1792	{
1793	Info.fOpenSource = strcmp(RTStrStrip(pszSyncProcess), "export") == 0;
1794	RTStrFree(pszSyncProcess);
1795	}
1796	else if (rc == VERR_NOT_FOUND)
1797	Info.fOpenSource = false;
1798	else
1799	return ScmError(pState, rc, "ScmSvnQueryProperty(svn:sync-process): %Rrc\n", rc);
1800
1801	Info.pExpectedLicense = Info.paLicenses;
1802	if (Info.fOpenSource)
1803	{
1804	if ( pSettings->enmUpdateLicense != kScmLicense_Mit
1805	&& pSettings->enmUpdateLicense != kScmLicense_BasedOnMit)
1806	while (Info.pExpectedLicense->enmOpt != pSettings->enmUpdateLicense)
1807	Info.pExpectedLicense++;
1808	else
1809	Assert(Info.pExpectedLicense->enmOpt == kScmLicense_Mit);
1810	}
1811	else
1812	while (Info.pExpectedLicense->enmType != kScmLicenseType_Confidential)
1813	Info.pExpectedLicense++;
1814
1815	/* Scan the comments. */
1816	rc = ScmEnumerateComments(pIn, enmCommentStyle, rewrite_Copyright_CommentCallback, &Info);
1817	if ( (rc == VERR_CALLBACK_RETURN \|\| RT_SUCCESS(rc))
1818	&& RT_SUCCESS(pState->rc))
1819	{
1820	/*
1821	* Do conformity checks.
1822	*/
1823	bool fAddLgplDisclaimer = false;
1824	if (Info.fCheckforLgpl)
1825	{
1826	if ( Info.iLineLgplNotice != UINT32_MAX
1827	&& Info.iLineLgplDisclaimer == UINT32_MAX)
1828	{
1829	if (!pSettings->fLgplDisclaimer) /** @todo reconcile options with common sense. */
1830	ScmError(pState, VERR_NOT_FOUND, "LGPL licence notice on line %u, but no LGPL disclaimer was found!\n",
1831	Info.iLineLgplNotice + 1);
1832	else
1833	{
1834	ScmVerbose(pState, 1, "* Need to add LGPL disclaimer\n");
1835	fAddLgplDisclaimer = true;
1836	}
1837	}
1838	else if ( Info.iLineLgplNotice == UINT32_MAX
1839	&& Info.iLineLgplDisclaimer != UINT32_MAX)
1840	ScmError(pState, VERR_NOT_FOUND, "LGPL disclaimer on line %u, but no LGPL copyright notice!\n",
1841	Info.iLineLgplDisclaimer + 1);
1842	}
1843
1844	if (!pSettings->fExternalCopyright)
1845	{
1846	if (Info.iLineCopyright == UINT32_MAX)
1847	ScmError(pState, VERR_NOT_FOUND, "Missing copyright!\n");
1848	if (Info.iLineLicense == UINT32_MAX)
1849	ScmError(pState, VERR_NOT_FOUND, "Missing license!\n");
1850	}
1851	else if (Info.iLineCopyright != UINT32_MAX)
1852	ScmError(pState, VERR_NOT_FOUND,
1853	"Marked as external copyright only, but found non-external copyright statement at line %u!\n",
1854	Info.iLineCopyright + 1);
1855
1856
1857	if (RT_SUCCESS(pState->rc))
1858	{
1859	/*
1860	* Do we need to make any changes?
1861	*/
1862	bool fUpdateCopyright = !pSettings->fExternalCopyright
1863	&& ( !Info.fWellFormedCopyright
1864	\|\| (!Info.fUpToDateCopyright && pSettings->fUpdateCopyrightYear));
1865	bool fUpdateLicense = !pSettings->fExternalCopyright
1866	&& Info.enmLicenceOpt != kScmLicense_LeaveAlone
1867	&& ( !Info.fWellFormedLicense
1868	\|\| !Info.fIsCorrectLicense);
1869	if ( fUpdateCopyright
1870	\|\| fUpdateLicense
1871	\|\| fAddLgplDisclaimer)
1872	{
1873	Assert(Info.iLineComment != UINT32_MAX);
1874	Assert(Info.cLinesComment > 0);
1875
1876	/*
1877	* Okay, do the work.
1878	*/
1879	ScmStreamRewindForReading(pIn);
1880
1881	if (pSettings->fUpdateCopyrightYear)
1882	Info.uLastYear = g_uYear;
1883
1884	uint32_t iLine = 0;
1885	SCMEOL enmEol;
1886	size_t cchLine;
1887	const char *pchLine;
1888	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
1889	{
1890	if ( iLine == Info.iLineComment
1891	&& (fUpdateCopyright \|\| fUpdateLicense) )
1892	{
1893	/* Leading blank line. */
1894	ScmStreamPutLine(pOut, g_aCopyrightCommentStart[enmCommentStyle].psz,
1895	g_aCopyrightCommentStart[enmCommentStyle].cch, enmEol);
1896
1897	/* Contributed by someone? */
1898	if (Info.pszContributedBy)
1899	{
1900	const char *psz = Info.pszContributedBy;
1901	for (;;)
1902	{
1903	const char *pszEol = strchr(psz, '\n');
1904	size_t cchContribLine = pszEol ? pszEol - psz : strlen(psz);
1905	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1906	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1907	ScmStreamWrite(pOut, psz, cchContribLine);
1908	ScmStreamPutEol(pOut, enmEol);
1909	if (!pszEol)
1910	break;
1911	psz = pszEol + 1;
1912	}
1913
1914	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1915	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1916	}
1917
1918	/* Write the copyright comment line. */
1919	ScmStreamWrite(pOut, g_aCopyrightCommentPrefix[enmCommentStyle].psz,
1920	g_aCopyrightCommentPrefix[enmCommentStyle].cch);
1921
1922	char szCopyright[256];
1923	size_t cchCopyright;
1924	if (Info.uFirstYear == Info.uLastYear)
1925	cchCopyright = RTStrPrintf(szCopyright, sizeof(szCopyright), "Copyright (C) %u %s",
1926	Info.uFirstYear, g_szCopyrightHolder);
1927	else
1928	cchCopyright = RTStrPrintf(szCopyright, sizeof(szCopyright), "Copyright (C) %u-%u %s",
1929	Info.uFirstYear, Info.uLastYear, g_szCopyrightHolder);
1930
1931	ScmStreamWrite(pOut, szCopyright, cchCopyright);
1932	ScmStreamPutEol(pOut, enmEol);
1933
1934	if (pSettings->enmUpdateLicense != kScmLicense_BasedOnMit)
1935	{
1936	/* Blank line separating the two. */
1937	ScmStreamPutLine(pOut, g_aCopyrightCommentEmpty[enmCommentStyle].psz,
1938	g_aCopyrightCommentEmpty[enmCommentStyle].cch, enmEol);
1939
1940	/* Write the license text. */
1941	scmWriteCommentBody(pOut, Info.pExpectedLicense->psz, Info.pExpectedLicense->cch,
1942	enmCommentStyle, enmEol);
1943
1944	/* Final comment line. */
1945	if (!Info.fExternalLicense)
1946	ScmStreamPutLine(pOut, g_aCopyrightCommentEnd[enmCommentStyle].psz,
1947	g_aCopyrightCommentEnd[enmCommentStyle].cch, enmEol);
1948	}
1949	else
1950	Assert(Info.fExternalLicense);
1951
1952	/* Skip the copyright and license text in the input file. */
1953	rc = ScmStreamGetStatus(pOut);
1954	if (RT_SUCCESS(rc))
1955	{
1956	iLine = Info.iLineComment + Info.cLinesComment;
1957	rc = ScmStreamSeekByLine(pIn, iLine);
1958	}
1959	}
1960	/*
1961	* Add LGPL disclaimer?
1962	*/
1963	else if ( iLine == Info.iLineAfterLgplComment
1964	&& fAddLgplDisclaimer)
1965	{
1966	ScmStreamPutEol(pOut, enmEol);
1967	ScmStreamPutLine(pOut, g_aCopyrightCommentStart[enmCommentStyle].psz,
1968	g_aCopyrightCommentStart[enmCommentStyle].cch, enmEol);
1969	scmWriteCommentBody(pOut, g_szLgplDisclaimer, sizeof(g_szLgplDisclaimer) - 1,
1970	enmCommentStyle, enmEol);
1971	ScmStreamPutLine(pOut, g_aCopyrightCommentEnd[enmCommentStyle].psz,
1972	g_aCopyrightCommentEnd[enmCommentStyle].cch, enmEol);
1973
1974	/* put the actual line */
1975	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
1976	iLine++;
1977	}
1978	else
1979	{
1980	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
1981	iLine++;
1982	}
1983	if (RT_FAILURE(rc))
1984	{
1985	RTStrFree(Info.pszContributedBy);
1986	return false;
1987	}
1988	} /* for each source line */
1989
1990	RTStrFree(Info.pszContributedBy);
1991	return true;
1992	}
1993	}
1994	}
1995	else
1996	ScmError(pState, rc, "ScmEnumerateComments: %Rrc\n", rc);
1997	NOREF(pState); NOREF(pOut);
1998	RTStrFree(Info.pszContributedBy);
1999	return false;
2000	}
2001
2002
2003	/** Copyright updater for C-style comments. */
2004	bool rewrite_Copyright_CstyleComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2005	{
2006	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_C);
2007	}
2008
2009	/** Copyright updater for hash-prefixed comments. */
2010	bool rewrite_Copyright_HashComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2011	{
2012	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Hash);
2013	}
2014
2015	/** Copyright updater for REM-prefixed comments. */
2016	bool rewrite_Copyright_RemComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2017	{
2018	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, determinBatchFileCommentStyle(pIn));
2019	}
2020
2021	/** Copyright updater for python comments. */
2022	bool rewrite_Copyright_PythonComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2023	{
2024	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Python);
2025	}
2026
2027	/** Copyright updater for semicolon-prefixed comments. */
2028	bool rewrite_Copyright_SemicolonComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2029	{
2030	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Semicolon);
2031	}
2032
2033	/** Copyright updater for sql comments. */
2034	bool rewrite_Copyright_SqlComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2035	{
2036	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Sql);
2037	}
2038
2039	/** Copyright updater for tick-prefixed comments. */
2040	bool rewrite_Copyright_TickComment(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2041	{
2042	return rewrite_Copyright_Common(pState, pIn, pOut, pSettings, kScmCommentStyle_Tick);
2043	}
2044
2045
2046	/**
2047	* Makefile.kup are empty files, enforce this.
2048	*
2049	* @returns true if modifications were made, false if not.
2050	* @param pIn The input stream.
2051	* @param pOut The output stream.
2052	* @param pSettings The settings.
2053	*/
2054	bool rewrite_Makefile_kup(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2055	{
2056	RT_NOREF2(pOut, pSettings);
2057
2058	/* These files should be zero bytes. */
2059	if (pIn->cb == 0)
2060	return false;
2061	ScmVerbose(pState, 2, " * Truncated file to zero bytes\n");
2062	return true;
2063	}
2064
2065	/**
2066	* Rewrite a kBuild makefile.
2067	*
2068	* @returns true if modifications were made, false if not.
2069	* @param pIn The input stream.
2070	* @param pOut The output stream.
2071	* @param pSettings The settings.
2072	*
2073	* @todo
2074	*
2075	* Ideas for Makefile.kmk and Config.kmk:
2076	* - sort if1of/ifn1of sets.
2077	* - line continuation slashes should only be preceded by one space.
2078	*/
2079	bool rewrite_Makefile_kmk(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2080	{
2081	RT_NOREF4(pState, pIn, pOut, pSettings);
2082	return false;
2083	}
2084
2085
2086	static bool isFlowerBoxSectionMarker(PSCMSTREAM pIn, const char *pchLine, size_t cchLine, uint32_t cchWidth,
2087	const char *ppchText, size_t pcchText, bool *pfNeedFixing)
2088	{
2089	*ppchText = NULL;
2090	*pcchText = 0;
2091	*pfNeedFixing = false;
2092
2093	/*
2094	* The first line.
2095	*/
2096	if (pchLine[0] != '/')
2097	return false;
2098	size_t offLine = 1;
2099	while (offLine < cchLine && pchLine[offLine] == '*')
2100	offLine++;
2101	if (offLine < 20) /* (Code below depend on a reasonable minimum here.) */
2102	return false;
2103	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2104	offLine++;
2105	if (offLine != cchLine)
2106	return false;
2107
2108	size_t const cchBox = cchLine;
2109	*pfNeedFixing = cchBox != cchWidth;
2110
2111	/*
2112	* The next line, extracting the text.
2113	*/
2114	SCMEOL enmEol;
2115	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2116	if (cchLine < cchBox - 3)
2117	return false;
2118
2119	offLine = 0;
2120	if (RT_C_IS_BLANK(pchLine[0]))
2121	{
2122	*pfNeedFixing = true;
2123	offLine = RT_C_IS_BLANK(pchLine[1]) ? 2 : 1;
2124	}
2125
2126	if (pchLine[offLine] != '*')
2127	return false;
2128	offLine++;
2129
2130	if (!RT_C_IS_BLANK(pchLine[offLine + 1]))
2131	return false;
2132	offLine++;
2133
2134	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2135	offLine++;
2136	if (offLine >= cchLine)
2137	return false;
2138	if (!RT_C_IS_UPPER(pchLine[offLine]))
2139	return false;
2140
2141	if (offLine != 4 \|\| cchLine != cchBox)
2142	*pfNeedFixing = true;
2143
2144	*ppchText = &pchLine[offLine];
2145	size_t const offText = offLine;
2146
2147	/* From the end now. */
2148	offLine = cchLine - 1;
2149	while (RT_C_IS_BLANK(pchLine[offLine]))
2150	offLine--;
2151
2152	if (pchLine[offLine] != '*')
2153	return false;
2154	offLine--;
2155	if (!RT_C_IS_BLANK(pchLine[offLine]))
2156	return false;
2157	offLine--;
2158	while (RT_C_IS_BLANK(pchLine[offLine]))
2159	offLine--;
2160	*pcchText = offLine - offText + 1;
2161
2162	/*
2163	* Third line closes the box.
2164	*/
2165	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2166	if (cchLine < cchBox - 3)
2167	return false;
2168
2169	offLine = 0;
2170	if (RT_C_IS_BLANK(pchLine[0]))
2171	{
2172	*pfNeedFixing = true;
2173	offLine = RT_C_IS_BLANK(pchLine[1]) ? 2 : 1;
2174	}
2175	while (offLine < cchLine && pchLine[offLine] == '*')
2176	offLine++;
2177	if (offLine < cchBox - 4)
2178	return false;
2179
2180	if (pchLine[offLine] != '/')
2181	return false;
2182	offLine++;
2183
2184	if (offLine != cchBox)
2185	*pfNeedFixing = true;
2186
2187	while (offLine < cchLine && RT_C_IS_BLANK(pchLine[offLine]))
2188	offLine++;
2189	if (offLine != cchLine)
2190	return false;
2191
2192	return true;
2193	}
2194
2195
2196	/**
2197	* Flower box marker comments in C and C++ code.
2198	*
2199	* @returns true if modifications were made, false if not.
2200	* @param pIn The input stream.
2201	* @param pOut The output stream.
2202	* @param pSettings The settings.
2203	*/
2204	bool rewrite_FixFlowerBoxMarkers(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2205	{
2206	if (!pSettings->fFixFlowerBoxMarkers)
2207	return false;
2208
2209	/*
2210	* Work thru the file line by line looking for flower box markers.
2211	*/
2212	size_t cChanges = 0;
2213	size_t cBlankLines = 0;
2214	SCMEOL enmEol;
2215	size_t cchLine;
2216	const char *pchLine;
2217	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2218	{
2219	/*
2220	* Get a likely match for a first line.
2221	*/
2222	if ( pchLine[0] == '/'
2223	&& cchLine > 20
2224	&& pchLine[1] == '*'
2225	&& pchLine[2] == '*'
2226	&& pchLine[3] == '*')
2227	{
2228	size_t const offSaved = ScmStreamTell(pIn);
2229	char const *pchText;
2230	size_t cchText;
2231	bool fNeedFixing;
2232	bool fIsFlowerBoxSection = isFlowerBoxSectionMarker(pIn, pchLine, cchLine, pSettings->cchWidth,
2233	&pchText, &cchText, &fNeedFixing);
2234	if ( fIsFlowerBoxSection
2235	&& ( fNeedFixing
2236	\|\| cBlankLines < pSettings->cMinBlankLinesBeforeFlowerBoxMakers) )
2237	{
2238	while (cBlankLines < pSettings->cMinBlankLinesBeforeFlowerBoxMakers)
2239	{
2240	ScmStreamPutEol(pOut, enmEol);
2241	cBlankLines++;
2242	}
2243
2244	ScmStreamPutCh(pOut, '/');
2245	ScmStreamWrite(pOut, g_szAsterisks, pSettings->cchWidth - 1);
2246	ScmStreamPutEol(pOut, enmEol);
2247
2248	static const char s_szLead[] = "* ";
2249	ScmStreamWrite(pOut, s_szLead, sizeof(s_szLead) - 1);
2250	ScmStreamWrite(pOut, pchText, cchText);
2251	size_t offCurPlus1 = sizeof(s_szLead) - 1 + cchText + 1;
2252	ScmStreamWrite(pOut, g_szSpaces, offCurPlus1 < pSettings->cchWidth ? pSettings->cchWidth - offCurPlus1 : 1);
2253	ScmStreamPutCh(pOut, '*');
2254	ScmStreamPutEol(pOut, enmEol);
2255
2256	ScmStreamWrite(pOut, g_szAsterisks, pSettings->cchWidth - 1);
2257	ScmStreamPutCh(pOut, '/');
2258	ScmStreamPutEol(pOut, enmEol);
2259
2260	cChanges++;
2261	cBlankLines = 0;
2262	continue;
2263	}
2264
2265	int rc = ScmStreamSeekAbsolute(pIn, offSaved);
2266	if (RT_FAILURE(rc))
2267	return false;
2268	}
2269
2270	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2271	if (RT_FAILURE(rc))
2272	return false;
2273
2274	/* Do blank line accounting so we can ensure at least two blank lines
2275	before each section marker. */
2276	if (!isBlankLine(pchLine, cchLine))
2277	cBlankLines = 0;
2278	else
2279	cBlankLines++;
2280	}
2281	if (cChanges > 0)
2282	ScmVerbose(pState, 2, " * Converted %zu flower boxer markers\n", cChanges);
2283	return cChanges != 0;
2284	}
2285
2286
2287	/**
2288	* Looks for the start of a todo comment.
2289	*
2290	* @returns Offset into the line of the comment start sequence.
2291	* @param pchLine The line to search.
2292	* @param cchLineBeforeTodo The length of the line before the todo.
2293	* @param pfSameLine Indicates whether it's refering to a statemtn on
2294	* the same line comment (true), or the next
2295	* statement (false).
2296	*/
2297	static size_t findTodoCommentStart(char const pchLine, size_t cchLineBeforeTodo, bool pfSameLine)
2298	{
2299	*pfSameLine = false;
2300
2301	/* Skip one '@' or '\\'. */
2302	char ch;
2303	if ( cchLineBeforeTodo > 2
2304	&& ( ((ch = pchLine[cchLineBeforeTodo - 1]) == '@')
2305	\|\| ch == '\\' ) )
2306	cchLineBeforeTodo--;
2307
2308	/* Skip blanks. */
2309	while ( cchLineBeforeTodo > 2
2310	&& RT_C_IS_BLANK(pchLine[cchLineBeforeTodo - 1]))
2311	cchLineBeforeTodo--;
2312
2313	/* Look for same line indicator. */
2314	if ( cchLineBeforeTodo > 0
2315	&& pchLine[cchLineBeforeTodo - 1] == '<')
2316	{
2317	*pfSameLine = true;
2318	cchLineBeforeTodo--;
2319	}
2320
2321	/* Skip s /
2322	while ( cchLineBeforeTodo > 1
2323	&& pchLine[cchLineBeforeTodo - 1] == '*')
2324	cchLineBeforeTodo--;
2325
2326	/* Do we have a comment opening sequence. */
2327	if ( cchLineBeforeTodo > 0
2328	&& pchLine[cchLineBeforeTodo - 1] == '/'
2329	&& ( ( cchLineBeforeTodo >= 2
2330	&& pchLine[cchLineBeforeTodo - 2] == '/')
2331	\|\| pchLine[cchLineBeforeTodo] == '*'))
2332	{
2333	/* Skip slashes at the start. */
2334	while ( cchLineBeforeTodo > 0
2335	&& pchLine[cchLineBeforeTodo - 1] == '/')
2336	cchLineBeforeTodo--;
2337
2338	return cchLineBeforeTodo;
2339	}
2340
2341	return ~(size_t)0;
2342	}
2343
2344
2345	/**
2346	* Looks for a TODO or todo in the given line.
2347	*
2348	* @returns Offset into the line of found, ~(size_t)0 if not.
2349	* @param pchLine The line to search.
2350	* @param cchLine The length of the line.
2351	*/
2352	static size_t findTodo(char const *pchLine, size_t cchLine)
2353	{
2354	if (cchLine >= 4 + 2)
2355	{
2356	/* We don't search the first to chars because we need the start of a comment.
2357	Also, skip the last three chars since we need at least four for a match. */
2358	size_t const cchLineT = cchLine - 3;
2359	if ( memchr(pchLine + 2, 't', cchLineT - 2) != NULL
2360	\|\| memchr(pchLine + 2, 'T', cchLineT - 2) != NULL)
2361	{
2362	for (size_t off = 2; off < cchLineT; off++)
2363	{
2364	char ch = pchLine[off];
2365	if ( ( ch != 't'
2366	&& ch != 'T')
2367	\|\| ( (ch = pchLine[off + 1]) != 'o'
2368	&& ch != 'O')
2369	\|\| ( (ch = pchLine[off + 2]) != 'd'
2370	&& ch != 'D')
2371	\|\| ( (ch = pchLine[off + 3]) != 'o'
2372	&& ch != 'O')
2373	\|\| ( off + 4 != cchLine
2374	&& (ch = pchLine[off + 4]) != ' '
2375	&& ch != '\t'
2376	&& ch != ':' /** @todo */
2377	&& (ch != '' \|\| off + 5 > cchLine \|\| pchLine[off + 5] != '/') /* @todo */
2378	) )
2379	{ /* not a hit - likely */ }
2380	else
2381	return off;
2382	}
2383	}
2384	}
2385	return ~(size_t)0;
2386	}
2387
2388
2389	/**
2390	* Doxygen todos in C and C++ code.
2391	*
2392	* @returns true if modifications were made, false if not.
2393	* @param pState The rewriter state.
2394	* @param pIn The input stream.
2395	* @param pOut The output stream.
2396	* @param pSettings The settings.
2397	*/
2398	bool rewrite_Fix_C_and_CPP_Todos(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2399	{
2400	if (!pSettings->fFixTodos)
2401	return false;
2402
2403	/*
2404	* Work thru the file line by line looking for the start of todo comments.
2405	*/
2406	size_t cChanges = 0;
2407	SCMEOL enmEol;
2408	size_t cchLine;
2409	const char *pchLine;
2410	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2411	{
2412	/*
2413	* Look for the word 'todo' in the line. We're currently only trying
2414	* to catch comments starting with the word todo and adjust the start of
2415	* the doxygen statement.
2416	*/
2417	size_t offTodo = findTodo(pchLine, cchLine);
2418	if ( offTodo != ~(size_t)0
2419	&& offTodo >= 2)
2420	{
2421	/* Work backwards to find the start of the comment. */
2422	bool fSameLine = false;
2423	size_t offCommentStart = findTodoCommentStart(pchLine, offTodo, &fSameLine);
2424	if (offCommentStart != ~(size_t)0)
2425	{
2426	char szNew[64];
2427	size_t cchNew = 0;
2428	szNew[cchNew++] = '/';
2429	szNew[cchNew++] = pchLine[offCommentStart + 1];
2430	szNew[cchNew++] = pchLine[offCommentStart + 1];
2431	if (fSameLine)
2432	szNew[cchNew++] = '<';
2433	szNew[cchNew++] = ' ';
2434	szNew[cchNew++] = '@';
2435	szNew[cchNew++] = 't';
2436	szNew[cchNew++] = 'o';
2437	szNew[cchNew++] = 'd';
2438	szNew[cchNew++] = 'o';
2439
2440	/* Figure out wheter to continue after the @todo statement opening, we'll strip ':'
2441	but need to take into account that we might be at the end of the line before
2442	adding the space. */
2443	size_t offTodoAfter = offTodo + 4;
2444	if ( offTodoAfter < cchLine
2445	&& pchLine[offTodoAfter] == ':')
2446	offTodoAfter++;
2447	if ( offTodoAfter < cchLine
2448	&& RT_C_IS_BLANK(pchLine[offTodoAfter]))
2449	offTodoAfter++;
2450	if (offTodoAfter < cchLine)
2451	szNew[cchNew++] = ' ';
2452
2453	/* Write it out. */
2454	ScmStreamWrite(pOut, pchLine, offCommentStart);
2455	ScmStreamWrite(pOut, szNew, cchNew);
2456	if (offTodoAfter < cchLine)
2457	ScmStreamWrite(pOut, &pchLine[offTodoAfter], cchLine - offTodoAfter);
2458	ScmStreamPutEol(pOut, enmEol);
2459
2460	/* Check whether we actually made any changes. */
2461	if ( cchNew != offTodoAfter - offCommentStart
2462	\|\| memcmp(szNew, &pchLine[offCommentStart], cchNew))
2463	cChanges++;
2464	continue;
2465	}
2466	}
2467
2468	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2469	if (RT_FAILURE(rc))
2470	return false;
2471	}
2472	if (cChanges > 0)
2473	ScmVerbose(pState, 2, " * Converted %zu todo statements.\n", cChanges);
2474	return cChanges != 0;
2475	}
2476
2477
2478	/**
2479	* Tries to parse a C/C++ preprocessor include directive.
2480	*
2481	* This is resonably forgiving and expects sane input.
2482	*
2483	* @retval kScmIncludeDir_Invalid if not a valid include directive.
2484	* @retval kScmIncludeDir_Quoted
2485	* @retval kScmIncludeDir_Bracketed
2486	* @retval kScmIncludeDir_Macro
2487	*
2488	* @param pState The rewriter state (for repording malformed
2489	* directives).
2490	* @param pchLine The line to try parse as an include statement.
2491	* @param cchLine The line length.
2492	* @param ppchFilename Where to return the pointer to the filename part.
2493	* @param pcchFilename Where to return the length of the filename.
2494	*/
2495	SCMINCLUDEDIR ScmMaybeParseCIncludeLine(PSCMRWSTATE pState, const char *pchLine, size_t cchLine,
2496	const char *ppchFilename, size_t pcchFilename)
2497	{
2498	/* Skip leading spaces: */
2499	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2500	cchLine--, pchLine++;
2501
2502	/* Check for '#': */
2503	if (cchLine > 0 && *pchLine == '#')
2504	{
2505	cchLine--;
2506	pchLine++;
2507
2508	/* Skip spaces after '#' (optional): */
2509	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2510	cchLine--, pchLine++;
2511
2512	/* Check for 'include': */
2513	static char const s_szInclude[] = "include";
2514	if ( cchLine >= sizeof(s_szInclude)
2515	&& memcmp(pchLine, RT_STR_TUPLE(s_szInclude)) == 0)
2516	{
2517	cchLine -= sizeof(s_szInclude) - 1;
2518	pchLine += sizeof(s_szInclude) - 1;
2519
2520	/* Skip spaces after 'include' word (optional): */
2521	while (cchLine > 0 && RT_C_IS_BLANK(*pchLine))
2522	cchLine--, pchLine++;
2523	if (cchLine > 0)
2524	{
2525	/* Quoted or bracketed? */
2526	char const chFirst = *pchLine;
2527	if (chFirst == '"' \|\| chFirst == '<')
2528	{
2529	cchLine--;
2530	pchLine++;
2531	const char pchEnd = (const char )memchr(pchLine, chFirst == '"' ? '"' : '>', cchLine);
2532	if (pchEnd)
2533	{
2534	if (ppchFilename)
2535	*ppchFilename = pchLine;
2536	if (pcchFilename)
2537	*pcchFilename = pchEnd - pchLine;
2538	return chFirst == '"' ? kScmIncludeDir_Quoted : kScmIncludeDir_Bracketed;
2539	}
2540	ScmError(pState, VERR_PARSE_ERROR, "Unbalanced #include filename %s: %.*s\n",
2541	chFirst == '"' ? "quotes" : "brackets" , cchLine, pchLine);
2542	}
2543	/* C prepreprocessor macro? */
2544	else if (ScmIsCIdentifierLeadChar(chFirst))
2545	{
2546	size_t cchFilename = 1;
2547	while ( cchFilename < cchLine
2548	&& ScmIsCIdentifierChar(pchLine[cchFilename]))
2549	cchFilename++;
2550	if (ppchFilename)
2551	*ppchFilename = pchLine;
2552	if (pcchFilename)
2553	*pcchFilename = cchFilename;
2554	return kScmIncludeDir_Macro;
2555	}
2556	else
2557	ScmError(pState, VERR_PARSE_ERROR, "Malformed #include filename part: %.*s\n", cchLine, pchLine);
2558	}
2559	else
2560	ScmError(pState, VERR_PARSE_ERROR, "Missing #include filename!\n");
2561	}
2562	}
2563
2564	if (ppchFilename)
2565	*ppchFilename = NULL;
2566	if (pcchFilename)
2567	*pcchFilename = 0;
2568	return kScmIncludeDir_Invalid;
2569	}
2570
2571
2572	/**
2573	* Fix err.h/errcore.h usage.
2574	*
2575	* @returns true if modifications were made, false if not.
2576	* @param pIn The input stream.
2577	* @param pOut The output stream.
2578	* @param pSettings The settings.
2579	*/
2580	bool rewrite_Fix_Err_H(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2581	{
2582	if (!pSettings->fFixErrH)
2583	return false;
2584
2585	static struct
2586	{
2587	const char *pszHeader;
2588	unsigned cchHeader;
2589	int iLevel;
2590	} const s_aHeaders[] =
2591	{
2592	{ RT_STR_TUPLE("iprt/errcore.h"), 1 },
2593	{ RT_STR_TUPLE("iprt/err.h"), 2 },
2594	{ RT_STR_TUPLE("VBox/err.h"), 3 },
2595	};
2596	static RTSTRTUPLE const g_aLevel1Statuses[] = /* Note! Keep in sync with errcore.h content! */
2597	{
2598	{ RT_STR_TUPLE("VINF_SUCCESS") },
2599	{ RT_STR_TUPLE("VERR_GENERAL_FAILURE") },
2600	{ RT_STR_TUPLE("VERR_INVALID_PARAMETER") },
2601	{ RT_STR_TUPLE("VWRN_INVALID_PARAMETER") },
2602	{ RT_STR_TUPLE("VERR_INVALID_MAGIC") },
2603	{ RT_STR_TUPLE("VWRN_INVALID_MAGIC") },
2604	{ RT_STR_TUPLE("VERR_INVALID_HANDLE") },
2605	{ RT_STR_TUPLE("VWRN_INVALID_HANDLE") },
2606	{ RT_STR_TUPLE("VERR_INVALID_POINTER") },
2607	{ RT_STR_TUPLE("VERR_NO_MEMORY") },
2608	{ RT_STR_TUPLE("VERR_PERMISSION_DENIED") },
2609	{ RT_STR_TUPLE("VINF_PERMISSION_DENIED") },
2610	{ RT_STR_TUPLE("VERR_VERSION_MISMATCH") },
2611	{ RT_STR_TUPLE("VERR_NOT_IMPLEMENTED") },
2612	{ RT_STR_TUPLE("VERR_INVALID_FLAGS") },
2613	{ RT_STR_TUPLE("VERR_WRONG_ORDER") },
2614	{ RT_STR_TUPLE("VERR_INVALID_FUNCTION") },
2615	{ RT_STR_TUPLE("VERR_NOT_SUPPORTED") },
2616	{ RT_STR_TUPLE("VINF_NOT_SUPPORTED") },
2617	{ RT_STR_TUPLE("VERR_ACCESS_DENIED") },
2618	{ RT_STR_TUPLE("VERR_INTERRUPTED") },
2619	{ RT_STR_TUPLE("VINF_INTERRUPTED") },
2620	{ RT_STR_TUPLE("VERR_TIMEOUT") },
2621	{ RT_STR_TUPLE("VINF_TIMEOUT") },
2622	{ RT_STR_TUPLE("VERR_BUFFER_OVERFLOW") },
2623	{ RT_STR_TUPLE("VINF_BUFFER_OVERFLOW") },
2624	{ RT_STR_TUPLE("VERR_TOO_MUCH_DATA") },
2625	{ RT_STR_TUPLE("VERR_TRY_AGAIN") },
2626	{ RT_STR_TUPLE("VINF_TRY_AGAIN") },
2627	{ RT_STR_TUPLE("VERR_PARSE_ERROR") },
2628	{ RT_STR_TUPLE("VERR_OUT_OF_RANGE") },
2629	{ RT_STR_TUPLE("VERR_NUMBER_TOO_BIG") },
2630	{ RT_STR_TUPLE("VWRN_NUMBER_TOO_BIG") },
2631	{ RT_STR_TUPLE("VERR_CANCELLED") },
2632	{ RT_STR_TUPLE("VERR_TRAILING_CHARS") },
2633	{ RT_STR_TUPLE("VWRN_TRAILING_CHARS") },
2634	{ RT_STR_TUPLE("VERR_TRAILING_SPACES") },
2635	{ RT_STR_TUPLE("VWRN_TRAILING_SPACES") },
2636	{ RT_STR_TUPLE("VERR_NOT_FOUND") },
2637	{ RT_STR_TUPLE("VWRN_NOT_FOUND") },
2638	{ RT_STR_TUPLE("VERR_INVALID_STATE") },
2639	{ RT_STR_TUPLE("VWRN_INVALID_STATE") },
2640	{ RT_STR_TUPLE("VERR_OUT_OF_RESOURCES") },
2641	{ RT_STR_TUPLE("VWRN_OUT_OF_RESOURCES") },
2642	{ RT_STR_TUPLE("VERR_END_OF_STRING") },
2643	{ RT_STR_TUPLE("VERR_CALLBACK_RETURN") },
2644	{ RT_STR_TUPLE("VINF_CALLBACK_RETURN") },
2645	{ RT_STR_TUPLE("VERR_DUPLICATE") },
2646	{ RT_STR_TUPLE("VERR_MISSING") },
2647	{ RT_STR_TUPLE("VERR_BUFFER_UNDERFLOW") },
2648	{ RT_STR_TUPLE("VINF_BUFFER_UNDERFLOW") },
2649	{ RT_STR_TUPLE("VERR_NOT_AVAILABLE") },
2650	{ RT_STR_TUPLE("VERR_MISMATCH") },
2651	{ RT_STR_TUPLE("VERR_WRONG_TYPE") },
2652	{ RT_STR_TUPLE("VWRN_WRONG_TYPE") },
2653	{ RT_STR_TUPLE("VERR_WRONG_PARAMETER_COUNT") },
2654	{ RT_STR_TUPLE("VERR_WRONG_PARAMETER_TYPE") },
2655	{ RT_STR_TUPLE("VERR_INVALID_CLIENT_ID") },
2656	{ RT_STR_TUPLE("VERR_INVALID_SESSION_ID") },
2657	{ RT_STR_TUPLE("VERR_INCOMPATIBLE_CONFIG") },
2658	{ RT_STR_TUPLE("VERR_INTERNAL_ERROR") },
2659	{ RT_STR_TUPLE("VINF_GETOPT_NOT_OPTION") },
2660	{ RT_STR_TUPLE("VERR_GETOPT_UNKNOWN_OPTION") },
2661	};
2662
2663	/*
2664	* First pass: Scout #include err.h/errcore.h locations and usage.
2665	*
2666	* Note! This isn't entirely optimal since it's also parsing comments and
2667	* strings, not just code. However it does a decent job for now.
2668	*/
2669	int iIncludeLevel = 0;
2670	int iUsageLevel = 0;
2671	uint32_t iLine = 0;
2672	SCMEOL enmEol;
2673	size_t cchLine;
2674	const char *pchLine;
2675	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2676	{
2677	iLine++;
2678	if (cchLine < 6)
2679	continue;
2680
2681	/*
2682	* Look for #includes.
2683	*/
2684	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2685	if ( pchHash
2686	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
2687	{
2688	const char *pchFilename;
2689	size_t cchFilename;
2690	SCMINCLUDEDIR enmIncDir = ScmMaybeParseCIncludeLine(pState, pchLine, cchLine, &pchFilename, &cchFilename);
2691	if ( enmIncDir == kScmIncludeDir_Bracketed
2692	\|\| enmIncDir == kScmIncludeDir_Quoted)
2693	{
2694	unsigned i = RT_ELEMENTS(s_aHeaders);
2695	while (i-- > 0)
2696	if ( s_aHeaders[i].cchHeader == cchFilename
2697	&& RTStrNICmpAscii(pchFilename, s_aHeaders[i].pszHeader, cchFilename) == 0)
2698	{
2699	if (iIncludeLevel < s_aHeaders[i].iLevel)
2700	iIncludeLevel = s_aHeaders[i].iLevel;
2701	break;
2702	}
2703
2704	/* Special hack for error info. */
2705	if (cchFilename == sizeof("errmsgdata.h") - 1 && memcmp(pchFilename, RT_STR_TUPLE("errmsgdata.h")) == 0)
2706	iUsageLevel = 4;
2707
2708	/* Special hack for code templates. */
2709	if ( cchFilename >= sizeof(".cpp.h")
2710	&& memcmp(&pchFilename[cchFilename - sizeof(".cpp.h") + 1], RT_STR_TUPLE(".cpp.h")) == 0)
2711	iUsageLevel = 4;
2712	continue;
2713	}
2714	}
2715	/*
2716	* Look for VERR_, VWRN_, VINF_ prefixed identifiers in the current line.
2717	*/
2718	const char pchHit = (const char )memchr(pchLine, 'V', cchLine);
2719	if (pchHit)
2720	{
2721	const char *pchLeft = pchLine;
2722	size_t cchLeft = cchLine;
2723	do
2724	{
2725	size_t cchLeftHit = &pchLeft[cchLeft] - pchHit;
2726	if (cchLeftHit < 6)
2727	break;
2728	if ( pchHit[4] == '_'
2729	&& ( pchHit == pchLine
2730	\|\| !ScmIsCIdentifierChar(pchHit[-1]))
2731	&& ( (pchHit[1] == 'E' && pchHit[2] == 'R' && pchHit[3] == 'R')
2732	\|\| (pchHit[1] == 'W' && pchHit[2] == 'R' && pchHit[3] == 'N')
2733	\|\| (pchHit[1] == 'I' && pchHit[2] == 'N' && pchHit[3] == 'F') ) )
2734	{
2735	size_t cchIdentifier = 5;
2736	while (cchIdentifier < cchLeftHit && ScmIsCIdentifierChar(pchHit[cchIdentifier]))
2737	cchIdentifier++;
2738	ScmVerbose(pState, 4, "--- status code at %u col %zu: %.*s\n",
2739	iLine, pchHit - pchLine, cchIdentifier, pchHit);
2740
2741	if (iUsageLevel <= 1)
2742	{
2743	iUsageLevel = 3; /* Cannot distingish between iprt/err.h and VBox/err.h, so pick the latter for now. */
2744	for (unsigned i = 0; i < RT_ELEMENTS(g_aLevel1Statuses); i++)
2745	if ( cchIdentifier == g_aLevel1Statuses[i].cch
2746	&& memcmp(pchHit, g_aLevel1Statuses[i].psz, cchIdentifier) == 0)
2747	{
2748	iUsageLevel = 1;
2749	break;
2750	}
2751	}
2752
2753	pchLeft = pchHit + cchIdentifier;
2754	cchLeft = cchLeftHit - cchIdentifier;
2755	}
2756	else
2757	{
2758	pchLeft = pchHit + 1;
2759	cchLeft = cchLeftHit - 1;
2760	}
2761	pchHit = (const char *)memchr(pchLeft, 'V', cchLeft);
2762	} while (pchHit != NULL);
2763	}
2764	}
2765	ScmVerbose(pState, 3, "--- iIncludeLevel=%d iUsageLevel=%d\n", iIncludeLevel, iUsageLevel);
2766
2767	/*
2768	* Second pass: Change err.h to errcore.h if we detected a need for change.
2769	*/
2770	if ( iIncludeLevel <= iUsageLevel
2771	\|\| iIncludeLevel <= 1 /* we cannot safely eliminate errcore.h includes atm. */)
2772	return false;
2773
2774	unsigned cChanges = 0;
2775	ScmStreamRewindForReading(pIn);
2776	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
2777	{
2778	/*
2779	* Look for #includes to modify.
2780	*/
2781	if (cchLine >= 6)
2782	{
2783	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2784	if ( pchHash
2785	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
2786	{
2787	const char *pchFilename;
2788	size_t cchFilename;
2789	SCMINCLUDEDIR enmIncDir = ScmMaybeParseCIncludeLine(pState, pchLine, cchLine, &pchFilename, &cchFilename);
2790	if ( enmIncDir == kScmIncludeDir_Bracketed
2791	\|\| enmIncDir == kScmIncludeDir_Quoted)
2792	{
2793	unsigned i = RT_ELEMENTS(s_aHeaders);
2794	while (i-- > 0)
2795	if ( s_aHeaders[i].cchHeader == cchFilename
2796	&& RTStrNICmpAscii(pchFilename, s_aHeaders[i].pszHeader, cchFilename) == 0)
2797	{
2798	ScmStreamWrite(pOut, pchLine, pchFilename - pchLine - 1);
2799	ScmStreamWrite(pOut, RT_STR_TUPLE("<iprt/errcore.h>"));
2800	size_t cchTrailing = &pchLine[cchLine] - &pchFilename[cchFilename + 1];
2801	if (cchTrailing > 0)
2802	ScmStreamWrite(pOut, &pchFilename[cchFilename + 1], cchTrailing);
2803	ScmStreamPutEol(pOut, enmEol);
2804	cChanges++;
2805	pchLine = NULL;
2806	break;
2807	}
2808	if (!pchLine)
2809	continue;
2810	}
2811	}
2812	}
2813
2814	int rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
2815	if (RT_FAILURE(rc))
2816	return false;
2817	}
2818	ScmVerbose(pState, 2, " * Converted %zu err.h/errcore.h include statements.\n", cChanges);
2819	return true;
2820	}
2821
2822	typedef struct
2823	{
2824	const char *pch;
2825	uint8_t cch;
2826	uint8_t cchSpaces; /*< Number of expected spaces before the word. /
2827	bool fSpacesBefore : 1; /*< Whether there may be spaces or tabs before the word. /
2828	bool fIdentifier : 1; /*< Whether we're to expect a C/C++ identifier rather than pch/cch. /
2829	} SCMMATCHWORD;
2830
2831
2832	int ScmMatchWords(const char pchLine, size_t cchLine, SCMMATCHWORD const paWords, size_t cWords,
2833	size_t *poffNext, PRTSTRTUPLE paIdentifiers, PRTERRINFO pErrInfo)
2834	{
2835	int rc = VINF_SUCCESS;
2836
2837	size_t offLine = 0;
2838	for (size_t i = 0; i < cWords; i++)
2839	{
2840	SCMMATCHWORD const *pWord = &paWords[i];
2841
2842	/*
2843	* Deal with spaces preceeding the word first:
2844	*/
2845	if (pWord->fSpacesBefore)
2846	{
2847	size_t cchSpaces = 0;
2848	size_t cchTabs = 0;
2849	while (offLine < cchLine)
2850	{
2851	const char ch = pchLine[offLine];
2852	if (ch == ' ')
2853	cchSpaces++;
2854	else if (ch == '\t')
2855	cchTabs++;
2856	else
2857	break;
2858	offLine++;
2859	}
2860
2861	if (cchSpaces == pWord->cchSpaces && cchTabs == 0)
2862	{ /* likely */ }
2863	else if (cchSpaces == 0 && cchTabs == 0)
2864	return RTErrInfoSetF(pErrInfo, VERR_PARSE_ERROR, "expected space at offset %u", offLine);
2865	else
2866	rc = VWRN_TRAILING_SPACES;
2867	}
2868	else
2869	Assert(pWord->cchSpaces == 0);
2870
2871	/*
2872	* C/C++ identifier?
2873	*/
2874	if (pWord->fIdentifier)
2875	{
2876	if (offLine >= cchLine)
2877	return RTErrInfoSetF(pErrInfo, VERR_END_OF_STRING,
2878	"expected '%.*s' (C/C++ identifier) at offset %u, not end of string",
2879	pWord->cch, pWord->pch, offLine);
2880	if (!ScmIsCIdentifierLeadChar(pchLine[offLine]))
2881	return RTErrInfoSetF(pErrInfo, VERR_MISMATCH, "expected '%.*s' (C/C++ identifier) at offset %u",
2882	pWord->cch, pWord->pch, offLine);
2883	size_t const offStart = offLine++;
2884	while (offLine < cchLine && ScmIsCIdentifierChar(pchLine[offLine]))
2885	offLine++;
2886	if (paIdentifiers)
2887	{
2888	paIdentifiers->cch = offLine - offStart;
2889	paIdentifiers->psz = &pchLine[offStart];
2890	paIdentifiers++;
2891	}
2892	}
2893	/*
2894	* Match the exact word.
2895	*/
2896	else if ( pWord->cch == 0
2897	\|\| ( pWord->cch <= cchLine - offLine
2898	&& !memcmp(pWord->pch, &pchLine[offLine], pWord->cch)))
2899	offLine += pWord->cch;
2900	else
2901	return RTErrInfoSetF(pErrInfo, VERR_MISMATCH, "expected '%.*s' at offset %u", pWord->cch, pWord->pch, offLine);
2902	}
2903
2904	/*
2905	* Check for trailing characters/whatnot.
2906	*/
2907	if (poffNext)
2908	*poffNext = offLine;
2909	else if (offLine != cchLine)
2910	rc = RTErrInfoSetF(pErrInfo, VERR_TRAILING_CHARS, "unexpected trailing characters at offset %u", offLine);
2911	return rc;
2912	}
2913
2914
2915	/**
2916	* Fix header file include guards and \#pragma once.
2917	*
2918	* @returns true if modifications were made, false if not.
2919	* @param pIn The input stream.
2920	* @param pOut The output stream.
2921	* @param pSettings The settings.
2922	*/
2923	bool rewrite_FixHeaderGuards(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
2924	{
2925	if (!pSettings->fFixHeaderGuards)
2926	return false;
2927
2928	/* always skip .cpp.h files */
2929	size_t cchFilename = strlen(pState->pszFilename);
2930	if ( cchFilename > sizeof(".cpp.h")
2931	&& RTStrICmpAscii(&pState->pszFilename[cchFilename - sizeof(".cpp.h") + 1], ".cpp.h") == 0)
2932	return false;
2933
2934	RTERRINFOSTATIC ErrInfo;
2935	char szNormalized[168];
2936	size_t cchNormalized = 0;
2937	int rc;
2938	bool fRet = false;
2939
2940	/*
2941	* Calculate the expected guard for this file, if so tasked.
2942	* ASSUMES pState->pszFilename is absolute as is pSettings->pszGuardRelativeToDir.
2943	*/
2944	szNormalized[0] = '\0';
2945	if (pSettings->pszGuardRelativeToDir)
2946	{
2947	rc = RTStrCopy(szNormalized, sizeof(szNormalized), pSettings->pszGuardPrefix);
2948	if (RT_FAILURE(rc))
2949	return ScmError(pState, rc, "Guard prefix too long (or something): %s\n", pSettings->pszGuardPrefix);
2950	cchNormalized = strlen(szNormalized);
2951	if (strcmp(pSettings->pszGuardRelativeToDir, "{dir}") == 0)
2952	rc = RTStrCopy(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized,
2953	RTPathFilename(pState->pszFilename));
2954	else if (strcmp(pSettings->pszGuardRelativeToDir, "{parent}") == 0)
2955	{
2956	const char *pszSrc = RTPathFilename(pState->pszFilename);
2957	if (!pszSrc \|\| (uintptr_t)&pszSrc[-2] < (uintptr_t)pState->pszFilename \|\| !RTPATH_IS_SLASH(pszSrc[-1]))
2958	return ScmError(pState, VERR_INTERNAL_ERROR, "Error calculating {parent} header guard!\n");
2959	pszSrc -= 2;
2960	while ( (uintptr_t)pszSrc > (uintptr_t)pState->pszFilename
2961	&& !RTPATH_IS_SLASH(pszSrc[-1])
2962	&& !RTPATH_IS_VOLSEP(pszSrc[-1]))
2963	pszSrc--;
2964	rc = RTStrCopy(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized, pszSrc);
2965	}
2966	else
2967	rc = RTPathCalcRelative(&szNormalized[cchNormalized], sizeof(szNormalized) - cchNormalized,
2968	pSettings->pszGuardRelativeToDir, false /fFromFile/, pState->pszFilename);
2969	if (RT_FAILURE(rc))
2970	return ScmError(pState, rc, "Error calculating guard prefix (RTPathCalcRelative): %Rrc\n", rc);
2971	char ch;
2972	while ((ch = szNormalized[cchNormalized]) != '\0')
2973	{
2974	if (!ScmIsCIdentifierChar(ch))
2975	szNormalized[cchNormalized] = '_';
2976	cchNormalized++;
2977	}
2978	}
2979
2980	/*
2981	* First part looks for the #ifndef xxxx paired with #define xxxx.
2982	*
2983	* We blindly assume the first preprocessor directive in the file is the guard
2984	* and will be upset if this isn't the case.
2985	*/
2986	RTSTRTUPLE Guard = { NULL, 0 };
2987	uint32_t cBlankLines = 0;
2988	SCMEOL enmEol;
2989	size_t cchLine;
2990	const char *pchLine;
2991	for (;;)
2992	{
2993	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
2994	if (pchLine == NULL)
2995	return ScmError(pState, VERR_PARSE_ERROR, "Did not find any include guards!\n");
2996	if (cchLine >= 2)
2997	{
2998	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
2999	if ( pchHash
3000	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
3001	{
3002	/* #ifndef xxxx */
3003	static const SCMMATCHWORD s_aIfndefGuard[] =
3004	{
3005	{ RT_STR_TUPLE("#"), 0, true, false },
3006	{ RT_STR_TUPLE("ifndef"), 0, true, false },
3007	{ RT_STR_TUPLE("IDENTIFIER"), 1, true, true },
3008	{ RT_STR_TUPLE(""), 0, true, false },
3009	};
3010	rc = ScmMatchWords(pchLine, cchLine, s_aIfndefGuard, RT_ELEMENTS(s_aIfndefGuard),
3011	NULL /poffNext/, &Guard, RTErrInfoInitStatic(&ErrInfo));
3012	if (RT_FAILURE(rc))
3013	return ScmError(pState, rc, "%u: Expected first preprocessor directive to be '#ifndef xxxx'. %s (%.*s)\n",
3014	ScmStreamTellLine(pIn) - 1, ErrInfo.Core.pszMsg, cchLine, pchLine);
3015	fRet \|= rc != VINF_SUCCESS;
3016	ScmVerbose(pState, 3, "line %u in %s: #ifndef %.*s\n",
3017	ScmStreamTellLine(pIn) - 1, pState->pszFilename, Guard.cch, Guard.psz);
3018
3019	/* #define xxxx */
3020	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3021	if (!pchLine)
3022	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef %.*s'\n",
3023	ScmStreamTellLine(pIn) - 1, Guard.cch, Guard.psz);
3024	const SCMMATCHWORD aDefineGuard[] =
3025	{
3026	{ RT_STR_TUPLE("#"), 0, true, false },
3027	{ RT_STR_TUPLE("define"), 0, true, false },
3028	{ Guard.psz, (uint8_t)Guard.cch, 1, true, false },
3029	{ RT_STR_TUPLE(""), 0, true, false },
3030	};
3031	rc = ScmMatchWords(pchLine, cchLine, aDefineGuard, RT_ELEMENTS(aDefineGuard),
3032	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3033	if (RT_FAILURE(rc))
3034	return ScmError(pState, rc, "%u: Expected '#define %.s' to follow '#ifndef %.s'. %s (%.*s)\n",
3035	ScmStreamTellLine(pIn) - 1, Guard.cch, Guard.psz, Guard.cch, Guard.psz,
3036	ErrInfo.Core.pszMsg, cchLine, pchLine);
3037	fRet \|= rc != VINF_SUCCESS;
3038
3039	if (Guard.cch >= sizeof(szNormalized))
3040	return ScmError(pState, VERR_BUFFER_OVERFLOW, "%u: Guard macro too long! %.*s\n",
3041	ScmStreamTellLine(pIn) - 2, Guard.cch, Guard.psz);
3042
3043	if (szNormalized[0] != '\0')
3044	{
3045	if ( Guard.cch != cchNormalized
3046	\|\| memcmp(Guard.psz, szNormalized, cchNormalized) != 0)
3047	{
3048	ScmVerbose(pState, 2, "guard changed from %.*s to %s\n", Guard.cch, Guard.psz, szNormalized);
3049	ScmVerbose(pState, 2, "grep -rw %.*s ${WCROOT} \| grep -Fv %s\n",
3050	Guard.cch, Guard.psz, pState->pszFilename);
3051	fRet = true;
3052	}
3053	Guard.psz = szNormalized;
3054	Guard.cch = cchNormalized;
3055	}
3056
3057	/*
3058	* Write guard, making sure we've got a single blank line preceeding it.
3059	*/
3060	ScmStreamPutEol(pOut, enmEol);
3061	ScmStreamWrite(pOut, RT_STR_TUPLE("#ifndef "));
3062	ScmStreamWrite(pOut, Guard.psz, Guard.cch);
3063	ScmStreamPutEol(pOut, enmEol);
3064	ScmStreamWrite(pOut, RT_STR_TUPLE("#define "));
3065	ScmStreamWrite(pOut, Guard.psz, Guard.cch);
3066	rc = ScmStreamPutEol(pOut, enmEol);
3067	if (RT_FAILURE(rc))
3068	return false;
3069	break;
3070	}
3071	}
3072
3073	if (!isBlankLine(pchLine, cchLine))
3074	{
3075	while (cBlankLines-- > 0)
3076	ScmStreamPutEol(pOut, enmEol);
3077	cBlankLines = 0;
3078	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3079	if (RT_FAILURE(rc))
3080	return false;
3081	}
3082	else
3083	cBlankLines++;
3084	}
3085
3086	/*
3087	* Look for pragma once wrapped in #ifndef RT_WITHOUT_PRAGMA_ONCE.
3088	*/
3089	size_t const iPragmaOnce = ScmStreamTellLine(pIn);
3090	static const SCMMATCHWORD s_aIfndefRtWithoutPragmaOnce[] =
3091	{
3092	{ RT_STR_TUPLE("#"), 0, true, false },
3093	{ RT_STR_TUPLE("ifndef"), 0, true, false },
3094	{ RT_STR_TUPLE("RT_WITHOUT_PRAGMA_ONCE"), 1, true, false },
3095	{ RT_STR_TUPLE(""), 0, true, false },
3096	};
3097	static const SCMMATCHWORD s_aPragmaOnce[] =
3098	{
3099	{ RT_STR_TUPLE("#"), 0, true, false },
3100	{ RT_STR_TUPLE("pragma"), 1, true, false },
3101	{ RT_STR_TUPLE("once"), 1, true, false},
3102	{ RT_STR_TUPLE(""), 0, true, false },
3103	};
3104	static const SCMMATCHWORD s_aEndif[] =
3105	{
3106	{ RT_STR_TUPLE("#"), 0, true, false },
3107	{ RT_STR_TUPLE("endif"), 0, true, false },
3108	{ RT_STR_TUPLE(""), 0, true, false },
3109	};
3110
3111	/* #ifndef RT_WITHOUT_PRAGMA_ONCE */
3112	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3113	if (!pchLine)
3114	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after header guard!\n", iPragmaOnce + 1);
3115	size_t offNext;
3116	rc = ScmMatchWords(pchLine, cchLine, s_aIfndefRtWithoutPragmaOnce, RT_ELEMENTS(s_aIfndefRtWithoutPragmaOnce),
3117	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3118	if (RT_SUCCESS(rc))
3119	{
3120	fRet \|= rc != VINF_SUCCESS;
3121	if (offNext != cchLine)
3122	return ScmError(pState, VERR_PARSE_ERROR, "%u: Characters trailing '#ifndef RT_WITHOUT_PRAGMA_ONCE' (%.*s)\n",
3123	iPragmaOnce + 1, cchLine, pchLine);
3124
3125	/* # pragma once */
3126	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3127	if (!pchLine)
3128	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef RT_WITHOUT_PRAGMA_ONCE'\n",
3129	iPragmaOnce + 2);
3130	rc = ScmMatchWords(pchLine, cchLine, s_aPragmaOnce, RT_ELEMENTS(s_aPragmaOnce),
3131	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3132	if (RT_SUCCESS(rc))
3133	fRet \|= rc != VINF_SUCCESS;
3134	else
3135	return ScmError(pState, rc, "%u: Expected '# pragma once' to follow '#ifndef RT_WITHOUT_PRAGMA_ONCE'! %s (%.*s)\n",
3136	iPragmaOnce + 2, ErrInfo.Core.pszMsg, cchLine, pchLine);
3137
3138	/* #endif */
3139	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3140	if (!pchLine)
3141	return ScmError(pState, VERR_PARSE_ERROR, "%u: Unexpected end of file after '#ifndef RT_WITHOUT_PRAGMA_ONCE' and '#pragma once'\n",
3142	iPragmaOnce + 3);
3143	rc = ScmMatchWords(pchLine, cchLine, s_aEndif, RT_ELEMENTS(s_aEndif),
3144	NULL /poffNext/, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3145	if (RT_SUCCESS(rc))
3146	fRet \|= rc != VINF_SUCCESS;
3147	else
3148	return ScmError(pState, rc,
3149	"%u: Expected '#endif' to follow '#ifndef RT_WITHOUT_PRAGMA_ONCE' and '# pragma once'! %s (%.*s)\n",
3150	iPragmaOnce + 3, ErrInfo.Core.pszMsg, cchLine, pchLine);
3151	ScmVerbose(pState, 3, "Found pragma once\n");
3152	fRet \|= !pSettings->fPragmaOnce;
3153	}
3154	else
3155	{
3156	rc = ScmStreamSeekByLine(pIn, iPragmaOnce);
3157	if (RT_FAILURE(rc))
3158	return ScmError(pState, rc, "seek error\n");
3159	fRet \|= pSettings->fPragmaOnce;
3160	ScmVerbose(pState, 2, "Missing #pragma once\n");
3161	}
3162
3163	/*
3164	* Write the pragma once stuff.
3165	*/
3166	if (pSettings->fPragmaOnce)
3167	{
3168	ScmStreamPutLine(pOut, RT_STR_TUPLE("#ifndef RT_WITHOUT_PRAGMA_ONCE"), enmEol);
3169	ScmStreamPutLine(pOut, RT_STR_TUPLE("# pragma once"), enmEol);
3170	rc = ScmStreamPutLine(pOut, RT_STR_TUPLE("#endif"), enmEol);
3171	if (RT_FAILURE(rc))
3172	return false;
3173	}
3174
3175	/*
3176	* Copy the rest of the file and remove pragma once statements, while
3177	* looking for the last #endif in the file.
3178	*/
3179	size_t iEndIfIn = 0;
3180	size_t iEndIfOut = 0;
3181	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3182	{
3183	if (cchLine > 2)
3184	{
3185	const char pchHash = (const char )memchr(pchLine, '#', cchLine);
3186	if ( pchHash
3187	&& isSpanOfBlanks(pchLine, pchHash - pchLine))
3188	{
3189	size_t off = pchHash - pchLine + 1;
3190	while (off < cchLine && RT_C_IS_BLANK(pchLine[off]))
3191	off++;
3192	/* #pragma once */
3193	if ( off + sizeof("pragma") - 1 <= cchLine
3194	&& !memcmp(&pchLine[off], RT_STR_TUPLE("pragma")))
3195	{
3196	rc = ScmMatchWords(pchLine, cchLine, s_aPragmaOnce, RT_ELEMENTS(s_aPragmaOnce),
3197	&offNext, NULL /paIdentifiers/, RTErrInfoInitStatic(&ErrInfo));
3198	if (RT_SUCCESS(rc))
3199	{
3200	fRet = true;
3201	continue;
3202	}
3203	}
3204	/* #endif */
3205	else if ( off + sizeof("endif") - 1 <= cchLine
3206	&& !memcmp(&pchLine[off], RT_STR_TUPLE("endif")))
3207	{
3208	iEndIfIn = ScmStreamTellLine(pIn) - 1;
3209	iEndIfOut = ScmStreamTellLine(pOut);
3210	}
3211	}
3212	}
3213
3214	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3215	if (RT_FAILURE(rc))
3216	return false;
3217	}
3218
3219	/*
3220	* Check out the last endif, making sure it's well formed and make sure it has the
3221	* right kind of comment following it.
3222	*/
3223	if (pSettings->fFixHeaderGuardEndif)
3224	{
3225	if (iEndIfOut == 0)
3226	return ScmError(pState, VERR_PARSE_ERROR, "Expected '#endif' at the end of the file...\n");
3227	rc = ScmStreamSeekByLine(pIn, iEndIfIn);
3228	if (RT_FAILURE(rc))
3229	return false;
3230	rc = ScmStreamSeekByLine(pOut, iEndIfOut);
3231	if (RT_FAILURE(rc))
3232	return false;
3233
3234	pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol);
3235	if (!pchLine)
3236	return ScmError(pState, VERR_INTERNAL_ERROR, "ScmStreamGetLine failed re-reading #endif!\n");
3237
3238	char szTmp[64 + sizeof(szNormalized)];
3239	size_t cchTmp;
3240	if (pSettings->fEndifGuardComment)
3241	cchTmp = RTStrPrintf(szTmp, sizeof(szTmp), "#endif /* !%.s /", Guard.cch, Guard.psz);
3242	else
3243	cchTmp = RTStrPrintf(szTmp, sizeof(szTmp), "#endif"); /* lazy bird */
3244	fRet \|= cchTmp != cchLine \|\| memcmp(szTmp, pchLine, cchTmp) != 0;
3245	rc = ScmStreamPutLine(pOut, szTmp, cchTmp, enmEol);
3246	if (RT_FAILURE(rc))
3247	return false;
3248
3249	/* Copy out the remaining lines (assumes no #pragma once here). */
3250	while ((pchLine = ScmStreamGetLine(pIn, &cchLine, &enmEol)) != NULL)
3251	{
3252	rc = ScmStreamPutLine(pOut, pchLine, cchLine, enmEol);
3253	if (RT_FAILURE(rc))
3254	return false;
3255	}
3256	}
3257
3258	return fRet;
3259	}
3260
3261
3262	/**
3263	* Rewrite a C/C++ source or header file.
3264	*
3265	* @returns true if modifications were made, false if not.
3266	* @param pIn The input stream.
3267	* @param pOut The output stream.
3268	* @param pSettings The settings.
3269	*
3270	* @todo
3271	*
3272	* Ideas for C/C++:
3273	* - space after if, while, for, switch
3274	* - spaces in for (i=0;i<x;i++)
3275	* - complex conditional, bird style.
3276	* - remove unnecessary parentheses.
3277	* - sort defined RT_OS_*\|\| and RT_ARCH
3278	* - sizeof without parenthesis.
3279	* - defined without parenthesis.
3280	* - trailing spaces.
3281	* - parameter indentation.
3282	* - space after comma.
3283	* - while (x--); -> multi line + comment.
3284	* - else statement;
3285	* - space between function and left parenthesis.
3286	* - TODO, XXX, @todo cleanup.
3287	* - Space before/after '*'.
3288	* - ensure new line at end of file.
3289	* - Indentation of precompiler statements (#ifdef, #defines).
3290	* - space between functions.
3291	* - string.h -> iprt/string.h, stdarg.h -> iprt/stdarg.h, etc.
3292	*/
3293	bool rewrite_C_and_CPP(PSCMRWSTATE pState, PSCMSTREAM pIn, PSCMSTREAM pOut, PCSCMSETTINGSBASE pSettings)
3294	{
3295
3296	RT_NOREF4(pState, pIn, pOut, pSettings);
3297	return false;
3298	}
3299

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

source: vbox/trunk/src/bldprogs/scmrw.cpp@ 92709

Download in other formats: