]> git.saurik.com Git - apt.git/blame - apt-pkg/tagfile.cc
test/hash.cc:
[apt.git] / apt-pkg / tagfile.cc
CommitLineData
578bfd0a
AL
1// -*- mode: cpp; mode: fold -*-
2// Description /*{{{*/
b3d44315 3// $Id: tagfile.cc,v 1.37.2.2 2003/12/31 16:02:30 mdz Exp $
578bfd0a
AL
4/* ######################################################################
5
6 Fast scanner for RFC-822 type header information
7
ad00ae81 8 This uses a rotating buffer to load the package information into.
578bfd0a
AL
9 The scanner runs over it and isolates and indexes a single section.
10
11 ##################################################################### */
12 /*}}}*/
13// Include Files /*{{{*/
6c139d6e 14#ifdef __GNUG__
094a497d 15#pragma implementation "apt-pkg/tagfile.h"
6c139d6e
AL
16#endif
17
094a497d
AL
18#include <apt-pkg/tagfile.h>
19#include <apt-pkg/error.h>
cdcc6d34 20#include <apt-pkg/strutl.h>
578bfd0a 21
b2e465d6
AL
22#include <apti18n.h>
23
578bfd0a
AL
24#include <string>
25#include <stdio.h>
851a45a8 26#include <ctype.h>
578bfd0a
AL
27 /*}}}*/
28
851a45a8
AL
29using std::string;
30
578bfd0a
AL
31// TagFile::pkgTagFile - Constructor /*{{{*/
32// ---------------------------------------------------------------------
33/* */
b3d44315
MV
34pkgTagFile::pkgTagFile(FileFd *pFd,unsigned long Size) :
35 Fd(*pFd),
36 Size(Size)
578bfd0a 37{
432b168c 38 if (Fd.IsOpen() == false || Fd.Size() == 0)
0e72dd52
AL
39 {
40 Buffer = 0;
41 Start = End = Buffer = 0;
0e72dd52 42 iOffset = 0;
432b168c 43 Map = NULL;
0e72dd52
AL
44 return;
45 }
46
432b168c
MV
47 Map = new MMap (Fd, MMap::Public | MMap::ReadOnly);
48 Buffer = (char *) Map->Data ();
49 Start = Buffer;
50 End = Buffer + Map->Size ();
dcb79bae 51 iOffset = 0;
578bfd0a
AL
52}
53 /*}}}*/
b2e465d6 54// TagFile::~pkgTagFile - Destructor /*{{{*/
29f7b36c
AL
55// ---------------------------------------------------------------------
56/* */
57pkgTagFile::~pkgTagFile()
58{
432b168c 59 delete Map;
29f7b36c
AL
60}
61 /*}}}*/
578bfd0a
AL
62// TagFile::Step - Advance to the next section /*{{{*/
63// ---------------------------------------------------------------------
64/* If the Section Scanner fails we refill the buffer and try again. */
65bool pkgTagFile::Step(pkgTagSection &Tag)
66{
432b168c
MV
67 if (Start == End)
68 return false;
69
0852eaef
AL
70 if (Tag.Scan(Start,End - Start) == false)
71 {
432b168c
MV
72 return _error->Error(_("Unable to parse package file %s (1)"),
73 Fd.Name().c_str());
613f9499 74 }
dcb79bae
AL
75 Start += Tag.size();
76 iOffset += Tag.size();
b2e465d6
AL
77
78 Tag.Trim();
578bfd0a
AL
79 return true;
80}
81 /*}}}*/
ad00ae81
AL
82// TagFile::Jump - Jump to a pre-recorded location in the file /*{{{*/
83// ---------------------------------------------------------------------
03e39e59
AL
84/* This jumps to a pre-recorded file location and reads the record
85 that is there */
ad00ae81
AL
86bool pkgTagFile::Jump(pkgTagSection &Tag,unsigned long Offset)
87{
b2e465d6
AL
88 // We are within a buffer space of the next hit..
89 if (Offset >= iOffset && iOffset + (End - Start) > Offset)
90 {
91 unsigned long Dist = Offset - iOffset;
92 Start += Dist;
93 iOffset += Dist;
94 return Step(Tag);
95 }
96
97 // Reposition and reload..
ad00ae81 98 iOffset = Offset;
432b168c 99 Start = Buffer + iOffset;
0852eaef 100
b90d7d85 101 // Start != End is a special case to not fail on empty TagFiles
c08447c8 102 if (Start != End && Tag.Scan(Start,End - Start) == false)
b2e465d6 103 return _error->Error(_("Unable to parse package file %s (2)"),Fd.Name().c_str());
06bba740 104
ad00ae81
AL
105 return true;
106}
107 /*}}}*/
578bfd0a
AL
108// TagSection::Scan - Scan for the end of the header information /*{{{*/
109// ---------------------------------------------------------------------
110/* This looks for the first double new line in the data stream. It also
c1a22377
AL
111 indexes the tags in the section. This very simple hash function for the
112 first 3 letters gives very good performance on the debian package files */
b2e465d6
AL
113inline static unsigned long AlphaHash(const char *Text, const char *End = 0)
114{
115 unsigned long Res = 0;
116 for (; Text != End && *Text != ':' && *Text != 0; Text++)
117 Res = (unsigned long)(*Text) ^ (Res << 2);
118 return Res & 0xFF;
119}
120
0852eaef 121bool pkgTagSection::Scan(const char *Start,unsigned long MaxLength)
578bfd0a
AL
122{
123 const char *End = Start + MaxLength;
124 Stop = Section = Start;
c1a22377 125 memset(AlphaIndexes,0,sizeof(AlphaIndexes));
c7b5ce1c 126
432b168c 127 if (Stop == 0 || MaxLength == 0)
0852eaef 128 return false;
578bfd0a
AL
129
130 TagCount = 0;
fd71171a 131 while (TagCount+1 < sizeof(Indexes)/sizeof(Indexes[0]) && Stop < End)
578bfd0a 132 {
90d64280 133 // Start a new index and add it to the hash
c1a22377
AL
134 if (isspace(Stop[0]) == 0)
135 {
136 Indexes[TagCount++] = Stop - Section;
b2e465d6 137 AlphaIndexes[AlphaHash(Stop,End)] = TagCount;
c1a22377 138 }
0a8e3465 139
c1a22377 140 Stop = (const char *)memchr(Stop,'\n',End - Stop);
0a8e3465 141
c1a22377 142 if (Stop == 0)
0852eaef 143 return false;
138d4b3d 144
b8b7c37d 145 for (; Stop+1 < End && Stop[1] == '\r'; Stop++);
c1a22377 146
f3bcc383
AL
147 // Double newline marks the end of the record
148 if (Stop+1 < End && Stop[1] == '\n')
578bfd0a 149 {
578bfd0a 150 Indexes[TagCount] = Stop - Section;
fd71171a 151 for (; Stop < End && (Stop[0] == '\n' || Stop[0] == '\r'); Stop++);
0852eaef 152 return true;
578bfd0a
AL
153 }
154
c1a22377
AL
155 Stop++;
156 }
138d4b3d 157
432b168c
MV
158 if ((Stop+1 >= End) && (End[-1] == '\n' || End[-1] == '\r'))
159 {
160 Indexes[TagCount] = (End - 1) - Section;
161 return true;
162 }
163
0852eaef 164 return false;
578bfd0a
AL
165}
166 /*}}}*/
b2e465d6
AL
167// TagSection::Trim - Trim off any trailing garbage /*{{{*/
168// ---------------------------------------------------------------------
169/* There should be exactly 1 newline at the end of the buffer, no more. */
170void pkgTagSection::Trim()
171{
172 for (; Stop > Section + 2 && (Stop[-2] == '\n' || Stop[-2] == '\r'); Stop--);
173}
174 /*}}}*/
578bfd0a
AL
175// TagSection::Find - Locate a tag /*{{{*/
176// ---------------------------------------------------------------------
177/* This searches the section for a tag that matches the given string. */
b2e465d6 178bool pkgTagSection::Find(const char *Tag,unsigned &Pos) const
578bfd0a
AL
179{
180 unsigned int Length = strlen(Tag);
b2e465d6 181 unsigned int I = AlphaIndexes[AlphaHash(Tag)];
c1a22377
AL
182 if (I == 0)
183 return false;
184 I--;
185
186 for (unsigned int Counter = 0; Counter != TagCount; Counter++,
187 I = (I+1)%TagCount)
578bfd0a 188 {
c1a22377
AL
189 const char *St;
190 St = Section + Indexes[I];
191 if (strncasecmp(Tag,St,Length) != 0)
578bfd0a
AL
192 continue;
193
b2e465d6
AL
194 // Make sure the colon is in the right place
195 const char *C = St + Length;
196 for (; isspace(*C) != 0; C++);
197 if (*C != ':')
198 continue;
199 Pos = I;
200 return true;
201 }
202
203 Pos = 0;
204 return false;
205}
206 /*}}}*/
207// TagSection::Find - Locate a tag /*{{{*/
208// ---------------------------------------------------------------------
209/* This searches the section for a tag that matches the given string. */
210bool pkgTagSection::Find(const char *Tag,const char *&Start,
211 const char *&End) const
212{
213 unsigned int Length = strlen(Tag);
214 unsigned int I = AlphaIndexes[AlphaHash(Tag)];
215 if (I == 0)
216 return false;
217 I--;
218
219 for (unsigned int Counter = 0; Counter != TagCount; Counter++,
220 I = (I+1)%TagCount)
221 {
222 const char *St;
223 St = Section + Indexes[I];
224 if (strncasecmp(Tag,St,Length) != 0)
225 continue;
226
578bfd0a 227 // Make sure the colon is in the right place
c1a22377 228 const char *C = St + Length;
578bfd0a
AL
229 for (; isspace(*C) != 0; C++);
230 if (*C != ':')
231 continue;
232
233 // Strip off the gunk from the start end
234 Start = C;
235 End = Section + Indexes[I+1];
06bba740
AL
236 if (Start >= End)
237 return _error->Error("Internal parsing error");
238
578bfd0a
AL
239 for (; (isspace(*Start) != 0 || *Start == ':') && Start < End; Start++);
240 for (; isspace(End[-1]) != 0 && End > Start; End--);
06bba740 241
578bfd0a
AL
242 return true;
243 }
c1a22377 244
578bfd0a
AL
245 Start = End = 0;
246 return false;
247}
248 /*}}}*/
0e66b144 249// TagSection::FindS - Find a string /*{{{*/
a05599f1
AL
250// ---------------------------------------------------------------------
251/* */
b2e465d6 252string pkgTagSection::FindS(const char *Tag) const
a05599f1
AL
253{
254 const char *Start;
255 const char *End;
256 if (Find(Tag,Start,End) == false)
257 return string();
258 return string(Start,End);
259}
260 /*}}}*/
261// TagSection::FindI - Find an integer /*{{{*/
262// ---------------------------------------------------------------------
263/* */
b2e465d6 264signed int pkgTagSection::FindI(const char *Tag,signed long Default) const
a05599f1
AL
265{
266 const char *Start;
b0b4efb9
AL
267 const char *Stop;
268 if (Find(Tag,Start,Stop) == false)
269 return Default;
270
271 // Copy it into a temp buffer so we can use strtol
272 char S[300];
273 if ((unsigned)(Stop - Start) >= sizeof(S))
274 return Default;
275 strncpy(S,Start,Stop-Start);
276 S[Stop - Start] = 0;
277
278 char *End;
279 signed long Result = strtol(S,&End,10);
280 if (S == End)
281 return Default;
282 return Result;
283}
284 /*}}}*/
285// TagSection::FindFlag - Locate a yes/no type flag /*{{{*/
286// ---------------------------------------------------------------------
287/* The bits marked in Flag are masked on/off in Flags */
288bool pkgTagSection::FindFlag(const char *Tag,unsigned long &Flags,
b2e465d6 289 unsigned long Flag) const
b0b4efb9
AL
290{
291 const char *Start;
292 const char *Stop;
293 if (Find(Tag,Start,Stop) == false)
294 return true;
a05599f1 295
b0b4efb9
AL
296 switch (StringToBool(string(Start,Stop)))
297 {
298 case 0:
299 Flags &= ~Flag;
300 return true;
301
302 case 1:
303 Flags |= Flag;
304 return true;
305
306 default:
b2e465d6 307 _error->Warning("Unknown flag value: %s",string(Start,Stop).c_str());
b0b4efb9
AL
308 return true;
309 }
310 return true;
a05599f1
AL
311}
312 /*}}}*/
b2e465d6
AL
313
314// TFRewrite - Rewrite a control record /*{{{*/
315// ---------------------------------------------------------------------
316/* This writes the control record to stdout rewriting it as necessary. The
317 override map item specificies the rewriting rules to follow. This also
318 takes the time to sort the feild list. */
319
320/* The order of this list is taken from dpkg source lib/parse.c the fieldinfos
321 array. */
322static const char *iTFRewritePackageOrder[] = {
323 "Package",
324 "Essential",
325 "Status",
326 "Priority",
327 "Section",
328 "Installed-Size",
329 "Maintainer",
330 "Architecture",
331 "Source",
332 "Version",
333 "Revision", // Obsolete
334 "Config-Version", // Obsolete
335 "Replaces",
336 "Provides",
337 "Depends",
338 "Pre-Depends",
339 "Recommends",
340 "Suggests",
341 "Conflicts",
342 "Conffiles",
343 "Filename",
344 "Size",
345 "MD5Sum",
cde41ae8
MV
346 "SHA1",
347 "SHA256",
b2e465d6
AL
348 "MSDOS-Filename", // Obsolete
349 "Description",
350 0};
351static const char *iTFRewriteSourceOrder[] = {"Package",
352 "Source",
353 "Binary",
354 "Version",
355 "Priority",
356 "Section",
357 "Maintainer",
358 "Build-Depends",
359 "Build-Depends-Indep",
360 "Build-Conflicts",
361 "Build-Conflicts-Indep",
362 "Architecture",
363 "Standards-Version",
364 "Format",
365 "Directory",
366 "Files",
367 0};
368
369/* Two levels of initialization are used because gcc will set the symbol
370 size of an array to the length of the array, causing dynamic relinking
371 errors. Doing this makes the symbol size constant */
372const char **TFRewritePackageOrder = iTFRewritePackageOrder;
373const char **TFRewriteSourceOrder = iTFRewriteSourceOrder;
374
375bool TFRewrite(FILE *Output,pkgTagSection const &Tags,const char *Order[],
376 TFRewriteData *Rewrite)
377{
378 unsigned char Visited[256]; // Bit 1 is Order, Bit 2 is Rewrite
379 for (unsigned I = 0; I != 256; I++)
380 Visited[I] = 0;
381
382 // Set new tag up as necessary.
383 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
384 {
385 if (Rewrite[J].NewTag == 0)
386 Rewrite[J].NewTag = Rewrite[J].Tag;
387 }
388
389 // Write all all of the tags, in order.
390 for (unsigned int I = 0; Order[I] != 0; I++)
391 {
392 bool Rewritten = false;
393
394 // See if this is a field that needs to be rewritten
395 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
396 {
397 if (strcasecmp(Rewrite[J].Tag,Order[I]) == 0)
398 {
399 Visited[J] |= 2;
400 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
401 {
402 if (isspace(Rewrite[J].Rewrite[0]))
403 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
404 else
405 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
406 }
407
408 Rewritten = true;
409 break;
410 }
411 }
412
413 // See if it is in the fragment
414 unsigned Pos;
415 if (Tags.Find(Order[I],Pos) == false)
416 continue;
417 Visited[Pos] |= 1;
418
419 if (Rewritten == true)
420 continue;
421
422 /* Write out this element, taking a moment to rewrite the tag
423 in case of changes of case. */
424 const char *Start;
425 const char *Stop;
426 Tags.Get(Start,Stop,Pos);
427
428 if (fputs(Order[I],Output) < 0)
429 return _error->Errno("fputs","IO Error to output");
430 Start += strlen(Order[I]);
431 if (fwrite(Start,Stop - Start,1,Output) != 1)
432 return _error->Errno("fwrite","IO Error to output");
433 if (Stop[-1] != '\n')
434 fprintf(Output,"\n");
435 }
436
437 // Now write all the old tags that were missed.
438 for (unsigned int I = 0; I != Tags.Count(); I++)
439 {
440 if ((Visited[I] & 1) == 1)
441 continue;
442
443 const char *Start;
444 const char *Stop;
445 Tags.Get(Start,Stop,I);
446 const char *End = Start;
447 for (; End < Stop && *End != ':'; End++);
448
449 // See if this is a field that needs to be rewritten
450 bool Rewritten = false;
451 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
452 {
453 if (stringcasecmp(Start,End,Rewrite[J].Tag) == 0)
454 {
455 Visited[J] |= 2;
456 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
457 {
458 if (isspace(Rewrite[J].Rewrite[0]))
459 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
460 else
461 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
462 }
463
464 Rewritten = true;
465 break;
466 }
467 }
468
469 if (Rewritten == true)
470 continue;
471
472 // Write out this element
473 if (fwrite(Start,Stop - Start,1,Output) != 1)
474 return _error->Errno("fwrite","IO Error to output");
475 if (Stop[-1] != '\n')
476 fprintf(Output,"\n");
477 }
478
479 // Now write all the rewrites that were missed
480 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
481 {
482 if ((Visited[J] & 2) == 2)
483 continue;
484
485 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
486 {
487 if (isspace(Rewrite[J].Rewrite[0]))
488 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
489 else
490 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
491 }
492 }
493
494 return true;
495}
496 /*}}}*/