scmdiff.cpp@ 53429

Last change on this file since 53429 was 52836, checked in by vboxsync, 10 years ago
scmDiffCompare: Use slow diff when ignoring trailing or leading spaces, not trailing or trailing.
Property svn:eol-style set to `native` Property svn:keywords set to `Author Date Id Revision`
File size: 15.7 KB

Line
1	/* $Id: scmdiff.cpp 52836 2014-09-24 13:02:36Z vboxsync $ */
2	/** @file
3	* IPRT Testcase / Tool - Source Code Massager.
4	*/
5
6	/*
7	* Copyright (C) 2010-2012 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	* Header Files *
20	*******************************************************************************/
21	#include <iprt/assert.h>
22	#include <iprt/ctype.h>
23	#include <iprt/message.h>
24	#include <iprt/stream.h>
25	#include <iprt/string.h>
26
27	#include "scmdiff.h"
28
29
30	/*******************************************************************************
31	* Global Variables *
32	*******************************************************************************/
33	static const char g_szTabSpaces[16+1] = " ";
34
35
36
37	/**
38	* Prints a range of lines with a prefix.
39	*
40	* @param pState The diff state.
41	* @param chPrefix The prefix.
42	* @param pStream The stream to get the lines from.
43	* @param iLine The first line.
44	* @param cLines The number of lines.
45	*/
46	static void scmDiffPrintLines(PSCMDIFFSTATE pState, char chPrefix, PSCMSTREAM pStream, size_t iLine, size_t cLines)
47	{
48	while (cLines-- > 0)
49	{
50	SCMEOL enmEol;
51	size_t cchLine;
52	const char *pchLine = ScmStreamGetLineByNo(pStream, iLine, &cchLine, &enmEol);
53
54	RTStrmPutCh(pState->pDiff, chPrefix);
55	if (pchLine && cchLine)
56	{
57	if (!pState->fSpecialChars)
58	RTStrmWrite(pState->pDiff, pchLine, cchLine);
59	else
60	{
61	size_t offVir = 0;
62	const char *pchStart = pchLine;
63	const char pchTab = (const char )memchr(pchLine, '\t', cchLine);
64	while (pchTab)
65	{
66	RTStrmWrite(pState->pDiff, pchStart, pchTab - pchStart);
67	offVir += pchTab - pchStart;
68
69	size_t cchTab = pState->cchTab - offVir % pState->cchTab;
70	switch (cchTab)
71	{
72	case 1: RTStrmPutStr(pState->pDiff, "."); break;
73	case 2: RTStrmPutStr(pState->pDiff, ".."); break;
74	case 3: RTStrmPutStr(pState->pDiff, "[T]"); break;
75	case 4: RTStrmPutStr(pState->pDiff, "[TA]"); break;
76	case 5: RTStrmPutStr(pState->pDiff, "[TAB]"); break;
77	default: RTStrmPrintf(pState->pDiff, "[TAB%.*s]", cchTab - 5, g_szTabSpaces); break;
78	}
79	offVir += cchTab;
80
81	/* next */
82	pchStart = pchTab + 1;
83	pchTab = (const char *)memchr(pchStart, '\t', cchLine - (pchStart - pchLine));
84	}
85	size_t cchLeft = cchLine - (pchStart - pchLine);
86	if (cchLeft)
87	RTStrmWrite(pState->pDiff, pchStart, cchLeft);
88	}
89	}
90
91	if (!pState->fSpecialChars)
92	RTStrmPutCh(pState->pDiff, '\n');
93	else if (enmEol == SCMEOL_LF)
94	RTStrmPutStr(pState->pDiff, "[LF]\n");
95	else if (enmEol == SCMEOL_CRLF)
96	RTStrmPutStr(pState->pDiff, "[CRLF]\n");
97	else
98	RTStrmPutStr(pState->pDiff, "[NONE]\n");
99
100	iLine++;
101	}
102	}
103
104
105	/**
106	* Reports a difference and propels the streams to the lines following the
107	* resync.
108	*
109	*
110	* @returns New pState->cDiff value (just to return something).
111	* @param pState The diff state. The cDiffs member will be
112	* incremented.
113	* @param cMatches The resync length.
114	* @param iLeft Where the difference starts on the left side.
115	* @param cLeft How long it is on this side. ~(size_t)0 is used
116	* to indicate that it goes all the way to the end.
117	* @param iRight Where the difference starts on the right side.
118	* @param cRight How long it is.
119	*/
120	static size_t scmDiffReport(PSCMDIFFSTATE pState, size_t cMatches,
121	size_t iLeft, size_t cLeft,
122	size_t iRight, size_t cRight)
123	{
124	/*
125	* Adjust the input.
126	*/
127	if (cLeft == ~(size_t)0)
128	{
129	size_t c = ScmStreamCountLines(pState->pLeft);
130	if (c >= iLeft)
131	cLeft = c - iLeft;
132	else
133	{
134	iLeft = c;
135	cLeft = 0;
136	}
137	}
138
139	if (cRight == ~(size_t)0)
140	{
141	size_t c = ScmStreamCountLines(pState->pRight);
142	if (c >= iRight)
143	cRight = c - iRight;
144	else
145	{
146	iRight = c;
147	cRight = 0;
148	}
149	}
150
151	/*
152	* Print header if it's the first difference
153	*/
154	if (!pState->cDiffs)
155	RTStrmPrintf(pState->pDiff, "diff %s %s\n", pState->pszFilename, pState->pszFilename);
156
157	/*
158	* Emit the change description.
159	*/
160	char ch = cLeft == 0
161	? 'a'
162	: cRight == 0
163	? 'd'
164	: 'c';
165	if (cLeft > 1 && cRight > 1)
166	RTStrmPrintf(pState->pDiff, "%zu,%zu%c%zu,%zu\n", iLeft + 1, iLeft + cLeft, ch, iRight + 1, iRight + cRight);
167	else if (cLeft > 1)
168	RTStrmPrintf(pState->pDiff, "%zu,%zu%c%zu\n", iLeft + 1, iLeft + cLeft, ch, iRight + 1);
169	else if (cRight > 1)
170	RTStrmPrintf(pState->pDiff, "%zu%c%zu,%zu\n", iLeft + 1, ch, iRight + 1, iRight + cRight);
171	else
172	RTStrmPrintf(pState->pDiff, "%zu%c%zu\n", iLeft + 1, ch, iRight + 1);
173
174	/*
175	* And the lines.
176	*/
177	if (cLeft)
178	scmDiffPrintLines(pState, '<', pState->pLeft, iLeft, cLeft);
179	if (cLeft && cRight)
180	RTStrmPrintf(pState->pDiff, "---\n");
181	if (cRight)
182	scmDiffPrintLines(pState, '>', pState->pRight, iRight, cRight);
183
184	/*
185	* Reposition the streams (safely ignores return value).
186	*/
187	ScmStreamSeekByLine(pState->pLeft, iLeft + cLeft + cMatches);
188	ScmStreamSeekByLine(pState->pRight, iRight + cRight + cMatches);
189
190	pState->cDiffs++;
191	return pState->cDiffs;
192	}
193
194	/**
195	* Helper for scmDiffCompare that takes care of trailing spaces and stuff
196	* like that.
197	*/
198	static bool scmDiffCompareSlow(PSCMDIFFSTATE pState,
199	const char *pchLeft, size_t cchLeft, SCMEOL enmEolLeft,
200	const char *pchRight, size_t cchRight, SCMEOL enmEolRight)
201	{
202	if (pState->fIgnoreTrailingWhite)
203	{
204	while (cchLeft > 0 && RT_C_IS_SPACE(pchLeft[cchLeft - 1]))
205	cchLeft--;
206	while (cchRight > 0 && RT_C_IS_SPACE(pchRight[cchRight - 1]))
207	cchRight--;
208	}
209
210	if (pState->fIgnoreLeadingWhite)
211	{
212	while (cchLeft > 0 && RT_C_IS_SPACE(*pchLeft))
213	pchLeft++, cchLeft--;
214	while (cchRight > 0 && RT_C_IS_SPACE(*pchRight))
215	pchRight++, cchRight--;
216	}
217
218	if ( cchLeft != cchRight
219	\|\| (enmEolLeft != enmEolRight && !pState->fIgnoreEol)
220	\|\| memcmp(pchLeft, pchRight, cchLeft))
221	return false;
222	return true;
223	}
224
225	/**
226	* Compare two lines.
227	*
228	* @returns true if the are equal, false if not.
229	*/
230	DECLINLINE(bool) scmDiffCompare(PSCMDIFFSTATE pState,
231	const char *pchLeft, size_t cchLeft, SCMEOL enmEolLeft,
232	const char *pchRight, size_t cchRight, SCMEOL enmEolRight)
233	{
234	if ( cchLeft != cchRight
235	\|\| (enmEolLeft != enmEolRight && !pState->fIgnoreEol)
236	\|\| memcmp(pchLeft, pchRight, cchLeft))
237	{
238	if ( pState->fIgnoreTrailingWhite
239	\|\| pState->fIgnoreLeadingWhite)
240	return scmDiffCompareSlow(pState,
241	pchLeft, cchLeft, enmEolLeft,
242	pchRight, cchRight, enmEolRight);
243	return false;
244	}
245	return true;
246	}
247
248	/**
249	* Compares two sets of lines from the two files.
250	*
251	* @returns true if they matches, false if they don't.
252	* @param pState The diff state.
253	* @param iLeft Where to start in the left stream.
254	* @param iRight Where to start in the right stream.
255	* @param cLines How many lines to compare.
256	*/
257	static bool scmDiffCompareLines(PSCMDIFFSTATE pState, size_t iLeft, size_t iRight, size_t cLines)
258	{
259	for (size_t iLine = 0; iLine < cLines; iLine++)
260	{
261	SCMEOL enmEolLeft;
262	size_t cchLeft;
263	const char *pchLeft = ScmStreamGetLineByNo(pState->pLeft, iLeft + iLine, &cchLeft, &enmEolLeft);
264
265	SCMEOL enmEolRight;
266	size_t cchRight;
267	const char *pchRight = ScmStreamGetLineByNo(pState->pRight, iRight + iLine, &cchRight, &enmEolRight);
268
269	if (!scmDiffCompare(pState, pchLeft, cchLeft, enmEolLeft, pchRight, cchRight, enmEolRight))
270	return false;
271	}
272	return true;
273	}
274
275
276	/**
277	* Resynchronize the two streams and reports the difference.
278	*
279	* Upon return, the streams will be positioned after the block of @a cMatches
280	* lines where it resynchronized them.
281	*
282	* @returns pState->cDiffs (just so we can use it in a return statement).
283	* @param pState The state.
284	* @param cMatches The number of lines that needs to match for the
285	* stream to be considered synchronized again.
286	*/
287	static size_t scmDiffSynchronize(PSCMDIFFSTATE pState, size_t cMatches)
288	{
289	size_t const iStartLeft = ScmStreamTellLine(pState->pLeft) - 1;
290	size_t const iStartRight = ScmStreamTellLine(pState->pRight) - 1;
291	Assert(cMatches > 0);
292
293	/*
294	* Compare each new line from each of the streams will all the preceding
295	* ones, including iStartLeft/Right.
296	*/
297	for (size_t iRange = 1; ; iRange++)
298	{
299	/*
300	* Get the next line in the left stream and compare it against all the
301	* preceding lines on the right side.
302	*/
303	SCMEOL enmEol;
304	size_t cchLine;
305	const char *pchLine = ScmStreamGetLineByNo(pState->pLeft, iStartLeft + iRange, &cchLine, &enmEol);
306	if (!pchLine)
307	return scmDiffReport(pState, 0, iStartLeft, ~(size_t)0, iStartRight, ~(size_t)0);
308
309	for (size_t iRight = cMatches - 1; iRight < iRange; iRight++)
310	{
311	SCMEOL enmEolRight;
312	size_t cchRight;
313	const char *pchRight = ScmStreamGetLineByNo(pState->pRight, iStartRight + iRight,
314	&cchRight, &enmEolRight);
315	if ( scmDiffCompare(pState, pchLine, cchLine, enmEol, pchRight, cchRight, enmEolRight)
316	&& scmDiffCompareLines(pState,
317	iStartLeft + iRange + 1 - cMatches,
318	iStartRight + iRight + 1 - cMatches,
319	cMatches - 1)
320	)
321	return scmDiffReport(pState, cMatches,
322	iStartLeft, iRange + 1 - cMatches,
323	iStartRight, iRight + 1 - cMatches);
324	}
325
326	/*
327	* Get the next line in the right stream and compare it against all the
328	* lines on the right side.
329	*/
330	pchLine = ScmStreamGetLineByNo(pState->pRight, iStartRight + iRange, &cchLine, &enmEol);
331	if (!pchLine)
332	return scmDiffReport(pState, 0, iStartLeft, ~(size_t)0, iStartRight, ~(size_t)0);
333
334	for (size_t iLeft = cMatches - 1; iLeft <= iRange; iLeft++)
335	{
336	SCMEOL enmEolLeft;
337	size_t cchLeft;
338	const char *pchLeft = ScmStreamGetLineByNo(pState->pLeft, iStartLeft + iLeft,
339	&cchLeft, &enmEolLeft);
340	if ( scmDiffCompare(pState, pchLeft, cchLeft, enmEolLeft, pchLine, cchLine, enmEol)
341	&& scmDiffCompareLines(pState,
342	iStartLeft + iLeft + 1 - cMatches,
343	iStartRight + iRange + 1 - cMatches,
344	cMatches - 1)
345	)
346	return scmDiffReport(pState, cMatches,
347	iStartLeft, iLeft + 1 - cMatches,
348	iStartRight, iRange + 1 - cMatches);
349	}
350	}
351	}
352
353	/**
354	* Creates a diff of the changes between the streams @a pLeft and @a pRight.
355	*
356	* This currently only implements the simplest diff format, so no contexts.
357	*
358	* Also, note that we won't detect differences in the final newline of the
359	* streams.
360	*
361	* @returns The number of differences.
362	* @param pszFilename The filename.
363	* @param pLeft The left side stream.
364	* @param pRight The right side stream.
365	* @param fIgnoreEol Whether to ignore end of line markers.
366	* @param fIgnoreLeadingWhite Set if leading white space should be ignored.
367	* @param fIgnoreTrailingWhite Set if trailing white space should be ignored.
368	* @param fSpecialChars Whether to print special chars in a human
369	* readable form or not.
370	* @param cchTab The tab size.
371	* @param pDiff Where to write the diff.
372	*/
373	size_t ScmDiffStreams(const char *pszFilename, PSCMSTREAM pLeft, PSCMSTREAM pRight, bool fIgnoreEol,
374	bool fIgnoreLeadingWhite, bool fIgnoreTrailingWhite, bool fSpecialChars,
375	size_t cchTab, PRTSTREAM pDiff)
376	{
377	#ifdef RT_STRICT
378	ScmStreamCheckItegrity(pLeft);
379	ScmStreamCheckItegrity(pRight);
380	#endif
381
382	/*
383	* Set up the diff state.
384	*/
385	SCMDIFFSTATE State;
386	State.cDiffs = 0;
387	State.pszFilename = pszFilename;
388	State.pLeft = pLeft;
389	State.pRight = pRight;
390	State.fIgnoreEol = fIgnoreEol;
391	State.fIgnoreLeadingWhite = fIgnoreLeadingWhite;
392	State.fIgnoreTrailingWhite = fIgnoreTrailingWhite;
393	State.fSpecialChars = fSpecialChars;
394	State.cchTab = cchTab;
395	State.pDiff = pDiff;
396
397	/*
398	* Compare them line by line.
399	*/
400	ScmStreamRewindForReading(pLeft);
401	ScmStreamRewindForReading(pRight);
402	const char *pchLeft;
403	const char *pchRight;
404
405	for (;;)
406	{
407	SCMEOL enmEolLeft;
408	size_t cchLeft;
409	pchLeft = ScmStreamGetLine(pLeft, &cchLeft, &enmEolLeft);
410
411	SCMEOL enmEolRight;
412	size_t cchRight;
413	pchRight = ScmStreamGetLine(pRight, &cchRight, &enmEolRight);
414	if (!pchLeft \|\| !pchRight)
415	break;
416
417	if (!scmDiffCompare(&State, pchLeft, cchLeft, enmEolLeft, pchRight, cchRight, enmEolRight))
418	scmDiffSynchronize(&State, 3);
419	}
420
421	/*
422	* Deal with any remaining differences.
423	*/
424	if (pchLeft)
425	scmDiffReport(&State, 0, ScmStreamTellLine(pLeft) - 1, ~(size_t)0, ScmStreamTellLine(pRight), 0);
426	else if (pchRight)
427	scmDiffReport(&State, 0, ScmStreamTellLine(pLeft), 0, ScmStreamTellLine(pRight) - 1, ~(size_t)0);
428
429	/*
430	* Report any errors.
431	*/
432	if (RT_FAILURE(ScmStreamGetStatus(pLeft)))
433	RTMsgError("Left diff stream error: %Rrc\n", ScmStreamGetStatus(pLeft));
434	if (RT_FAILURE(ScmStreamGetStatus(pRight)))
435	RTMsgError("Right diff stream error: %Rrc\n", ScmStreamGetStatus(pRight));
436
437	return State.cDiffs;
438	}
439

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

source: vbox/trunk/src/bldprogs/scmdiff.cpp@ 53429

Download in other formats: