]> git.saurik.com Git - apt.git/blame - apt-pkg/tagfile.cc
* debian/rules:
[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 /*{{{*/
094a497d
AL
14#include <apt-pkg/tagfile.h>
15#include <apt-pkg/error.h>
cdcc6d34 16#include <apt-pkg/strutl.h>
578bfd0a 17
b2e465d6
AL
18#include <apti18n.h>
19
578bfd0a
AL
20#include <string>
21#include <stdio.h>
851a45a8 22#include <ctype.h>
578bfd0a
AL
23 /*}}}*/
24
851a45a8
AL
25using std::string;
26
578bfd0a
AL
27// TagFile::pkgTagFile - Constructor /*{{{*/
28// ---------------------------------------------------------------------
29/* */
b3d44315
MV
30pkgTagFile::pkgTagFile(FileFd *pFd,unsigned long Size) :
31 Fd(*pFd),
2ca99a0d 32 Size(Size)
578bfd0a 33{
99c2e5ac 34 if (Fd.IsOpen() == false)
0e72dd52 35 {
2ca99a0d 36 Buffer = 0;
0e72dd52 37 Start = End = Buffer = 0;
99c2e5ac 38 Done = true;
0e72dd52
AL
39 iOffset = 0;
40 return;
41 }
42
2ca99a0d
MV
43 Buffer = new char[Size];
44 Start = End = Buffer;
45 Done = false;
dcb79bae 46 iOffset = 0;
2ca99a0d 47 Fill();
578bfd0a
AL
48}
49 /*}}}*/
b2e465d6 50// TagFile::~pkgTagFile - Destructor /*{{{*/
29f7b36c
AL
51// ---------------------------------------------------------------------
52/* */
53pkgTagFile::~pkgTagFile()
54{
2ca99a0d 55 delete [] Buffer;
29f7b36c
AL
56}
57 /*}}}*/
75c541fd 58// TagFile::Resize - Resize the internal buffer /*{{{*/
578bfd0a 59// ---------------------------------------------------------------------
75c541fd
MV
60/* Resize the internal buffer (double it in size). Fail if a maximum size
61 * size is reached.
62 */
63bool pkgTagFile::Resize()
578bfd0a 64{
75c541fd
MV
65 char *tmp;
66 unsigned long EndSize = End - Start;
67
68 // fail is the buffer grows too big
69 if(Size > 1024*1024+1)
432b168c
MV
70 return false;
71
75c541fd
MV
72 // get new buffer and use it
73 tmp = new char[2*Size];
74 memcpy(tmp, Buffer, Size);
75 Size = Size*2;
76 delete [] Buffer;
77 Buffer = tmp;
99c2e5ac 78
75c541fd
MV
79 // update the start/end pointers to the new buffer
80 Start = Buffer;
81 End = Start + EndSize;
82 return true;
83}
81e9789b 84 /*}}}*/
578bfd0a
AL
85// TagFile::Step - Advance to the next section /*{{{*/
86// ---------------------------------------------------------------------
75c541fd
MV
87/* If the Section Scanner fails we refill the buffer and try again.
88 * If that fails too, double the buffer size and try again until a
89 * maximum buffer is reached.
90 */
578bfd0a
AL
91bool pkgTagFile::Step(pkgTagSection &Tag)
92{
75c541fd 93 while (Tag.Scan(Start,End - Start) == false)
0852eaef 94 {
99c2e5ac
MV
95 if (Fill() == false)
96 return false;
97
75c541fd
MV
98 if(Tag.Scan(Start,End - Start))
99 break;
100
101 if (Resize() == false)
99c2e5ac 102 return _error->Error(_("Unable to parse package file %s (1)"),
75c541fd 103 Fd.Name().c_str());
613f9499 104 }
dcb79bae
AL
105 Start += Tag.size();
106 iOffset += Tag.size();
b2e465d6
AL
107
108 Tag.Trim();
99c2e5ac
MV
109 return true;
110}
111 /*}}}*/
112// TagFile::Fill - Top up the buffer /*{{{*/
113// ---------------------------------------------------------------------
114/* This takes the bit at the end of the buffer and puts it at the start
115 then fills the rest from the file */
116bool pkgTagFile::Fill()
117{
118 unsigned long EndSize = End - Start;
119 unsigned long Actual = 0;
120
121 memmove(Buffer,Start,EndSize);
122 Start = Buffer;
123 End = Buffer + EndSize;
124
125 if (Done == false)
126 {
127 // See if only a bit of the file is left
128 if (Fd.Read(End,Size - (End - Buffer),&Actual) == false)
129 return false;
130 if (Actual != Size - (End - Buffer))
131 Done = true;
132 End += Actual;
133 }
134
135 if (Done == true)
136 {
137 if (EndSize <= 3 && Actual == 0)
138 return false;
139 if (Size - (End - Buffer) < 4)
140 return true;
141
142 // Append a double new line if one does not exist
143 unsigned int LineCount = 0;
144 for (const char *E = End - 1; E - End < 6 && (*E == '\n' || *E == '\r'); E--)
145 if (*E == '\n')
146 LineCount++;
147 for (; LineCount < 2; LineCount++)
148 *End++ = '\n';
149
150 return true;
151 }
152
578bfd0a
AL
153 return true;
154}
155 /*}}}*/
ad00ae81
AL
156// TagFile::Jump - Jump to a pre-recorded location in the file /*{{{*/
157// ---------------------------------------------------------------------
03e39e59
AL
158/* This jumps to a pre-recorded file location and reads the record
159 that is there */
ad00ae81
AL
160bool pkgTagFile::Jump(pkgTagSection &Tag,unsigned long Offset)
161{
b2e465d6
AL
162 // We are within a buffer space of the next hit..
163 if (Offset >= iOffset && iOffset + (End - Start) > Offset)
164 {
165 unsigned long Dist = Offset - iOffset;
166 Start += Dist;
167 iOffset += Dist;
168 return Step(Tag);
169 }
170
2ca99a0d 171 // Reposition and reload..
ad00ae81 172 iOffset = Offset;
2ca99a0d
MV
173 Done = false;
174 if (Fd.Seek(Offset) == false)
175 return false;
176 End = Start = Buffer;
99c2e5ac 177
2ca99a0d
MV
178 if (Fill() == false)
179 return false;
99c2e5ac 180
2ca99a0d
MV
181 if (Tag.Scan(Start,End - Start) == true)
182 return true;
183
184 // This appends a double new line (for the real eof handling)
185 if (Fill() == false)
186 return false;
0852eaef 187
99c2e5ac 188 if (Tag.Scan(Start,End - Start) == false)
b2e465d6 189 return _error->Error(_("Unable to parse package file %s (2)"),Fd.Name().c_str());
06bba740 190
ad00ae81
AL
191 return true;
192}
193 /*}}}*/
578bfd0a
AL
194// TagSection::Scan - Scan for the end of the header information /*{{{*/
195// ---------------------------------------------------------------------
c176c4d0
DK
196/* This looks for the first double new line in the data stream.
197 It also indexes the tags in the section. */
0852eaef 198bool pkgTagSection::Scan(const char *Start,unsigned long MaxLength)
578bfd0a
AL
199{
200 const char *End = Start + MaxLength;
201 Stop = Section = Start;
c1a22377 202 memset(AlphaIndexes,0,sizeof(AlphaIndexes));
c7b5ce1c 203
2ca99a0d 204 if (Stop == 0)
0852eaef 205 return false;
81e9789b 206
578bfd0a 207 TagCount = 0;
fd71171a 208 while (TagCount+1 < sizeof(Indexes)/sizeof(Indexes[0]) && Stop < End)
578bfd0a 209 {
81e9789b
MV
210 TrimRecord(true,End);
211
90d64280 212 // Start a new index and add it to the hash
c1a22377
AL
213 if (isspace(Stop[0]) == 0)
214 {
215 Indexes[TagCount++] = Stop - Section;
b2e465d6 216 AlphaIndexes[AlphaHash(Stop,End)] = TagCount;
c1a22377 217 }
0a8e3465 218
c1a22377 219 Stop = (const char *)memchr(Stop,'\n',End - Stop);
0a8e3465 220
c1a22377 221 if (Stop == 0)
0852eaef 222 return false;
81e9789b 223
b8b7c37d 224 for (; Stop+1 < End && Stop[1] == '\r'; Stop++);
c1a22377 225
f3bcc383
AL
226 // Double newline marks the end of the record
227 if (Stop+1 < End && Stop[1] == '\n')
578bfd0a 228 {
578bfd0a 229 Indexes[TagCount] = Stop - Section;
81e9789b 230 TrimRecord(false,End);
0852eaef 231 return true;
578bfd0a
AL
232 }
233
c1a22377
AL
234 Stop++;
235 }
138d4b3d 236
0852eaef 237 return false;
578bfd0a
AL
238}
239 /*}}}*/
81e9789b
MV
240// TagSection::TrimRecord - Trim off any garbage before/after a record /*{{{*/
241// ---------------------------------------------------------------------
242/* There should be exactly 2 newline at the end of the record, no more. */
243void pkgTagSection::TrimRecord(bool BeforeRecord, const char*& End)
244{
245 if (BeforeRecord == true)
246 return;
247 for (; Stop < End && (Stop[0] == '\n' || Stop[0] == '\r'); Stop++);
248}
249 /*}}}*/
b2e465d6
AL
250// TagSection::Trim - Trim off any trailing garbage /*{{{*/
251// ---------------------------------------------------------------------
252/* There should be exactly 1 newline at the end of the buffer, no more. */
253void pkgTagSection::Trim()
254{
255 for (; Stop > Section + 2 && (Stop[-2] == '\n' || Stop[-2] == '\r'); Stop--);
256}
257 /*}}}*/
578bfd0a
AL
258// TagSection::Find - Locate a tag /*{{{*/
259// ---------------------------------------------------------------------
260/* This searches the section for a tag that matches the given string. */
b2e465d6 261bool pkgTagSection::Find(const char *Tag,unsigned &Pos) const
578bfd0a
AL
262{
263 unsigned int Length = strlen(Tag);
b2e465d6 264 unsigned int I = AlphaIndexes[AlphaHash(Tag)];
c1a22377
AL
265 if (I == 0)
266 return false;
267 I--;
268
269 for (unsigned int Counter = 0; Counter != TagCount; Counter++,
270 I = (I+1)%TagCount)
578bfd0a 271 {
c1a22377
AL
272 const char *St;
273 St = Section + Indexes[I];
274 if (strncasecmp(Tag,St,Length) != 0)
578bfd0a
AL
275 continue;
276
b2e465d6
AL
277 // Make sure the colon is in the right place
278 const char *C = St + Length;
279 for (; isspace(*C) != 0; C++);
280 if (*C != ':')
281 continue;
282 Pos = I;
283 return true;
284 }
285
286 Pos = 0;
287 return false;
288}
289 /*}}}*/
290// TagSection::Find - Locate a tag /*{{{*/
291// ---------------------------------------------------------------------
292/* This searches the section for a tag that matches the given string. */
293bool pkgTagSection::Find(const char *Tag,const char *&Start,
294 const char *&End) const
295{
296 unsigned int Length = strlen(Tag);
297 unsigned int I = AlphaIndexes[AlphaHash(Tag)];
298 if (I == 0)
299 return false;
300 I--;
301
302 for (unsigned int Counter = 0; Counter != TagCount; Counter++,
303 I = (I+1)%TagCount)
304 {
305 const char *St;
306 St = Section + Indexes[I];
307 if (strncasecmp(Tag,St,Length) != 0)
308 continue;
309
578bfd0a 310 // Make sure the colon is in the right place
c1a22377 311 const char *C = St + Length;
578bfd0a
AL
312 for (; isspace(*C) != 0; C++);
313 if (*C != ':')
314 continue;
315
316 // Strip off the gunk from the start end
317 Start = C;
318 End = Section + Indexes[I+1];
06bba740
AL
319 if (Start >= End)
320 return _error->Error("Internal parsing error");
321
578bfd0a
AL
322 for (; (isspace(*Start) != 0 || *Start == ':') && Start < End; Start++);
323 for (; isspace(End[-1]) != 0 && End > Start; End--);
06bba740 324
578bfd0a
AL
325 return true;
326 }
c1a22377 327
578bfd0a
AL
328 Start = End = 0;
329 return false;
330}
331 /*}}}*/
0e66b144 332// TagSection::FindS - Find a string /*{{{*/
a05599f1
AL
333// ---------------------------------------------------------------------
334/* */
b2e465d6 335string pkgTagSection::FindS(const char *Tag) const
a05599f1
AL
336{
337 const char *Start;
338 const char *End;
339 if (Find(Tag,Start,End) == false)
340 return string();
341 return string(Start,End);
342}
343 /*}}}*/
344// TagSection::FindI - Find an integer /*{{{*/
345// ---------------------------------------------------------------------
346/* */
b2e465d6 347signed int pkgTagSection::FindI(const char *Tag,signed long Default) const
a05599f1
AL
348{
349 const char *Start;
b0b4efb9
AL
350 const char *Stop;
351 if (Find(Tag,Start,Stop) == false)
352 return Default;
353
354 // Copy it into a temp buffer so we can use strtol
355 char S[300];
356 if ((unsigned)(Stop - Start) >= sizeof(S))
357 return Default;
358 strncpy(S,Start,Stop-Start);
359 S[Stop - Start] = 0;
360
361 char *End;
362 signed long Result = strtol(S,&End,10);
363 if (S == End)
364 return Default;
365 return Result;
366}
367 /*}}}*/
368// TagSection::FindFlag - Locate a yes/no type flag /*{{{*/
369// ---------------------------------------------------------------------
370/* The bits marked in Flag are masked on/off in Flags */
371bool pkgTagSection::FindFlag(const char *Tag,unsigned long &Flags,
b2e465d6 372 unsigned long Flag) const
b0b4efb9
AL
373{
374 const char *Start;
375 const char *Stop;
376 if (Find(Tag,Start,Stop) == false)
377 return true;
a05599f1 378
b0b4efb9
AL
379 switch (StringToBool(string(Start,Stop)))
380 {
381 case 0:
382 Flags &= ~Flag;
383 return true;
384
385 case 1:
386 Flags |= Flag;
387 return true;
388
389 default:
b2e465d6 390 _error->Warning("Unknown flag value: %s",string(Start,Stop).c_str());
b0b4efb9
AL
391 return true;
392 }
393 return true;
a05599f1
AL
394}
395 /*}}}*/
b2e465d6
AL
396// TFRewrite - Rewrite a control record /*{{{*/
397// ---------------------------------------------------------------------
398/* This writes the control record to stdout rewriting it as necessary. The
399 override map item specificies the rewriting rules to follow. This also
400 takes the time to sort the feild list. */
401
402/* The order of this list is taken from dpkg source lib/parse.c the fieldinfos
403 array. */
404static const char *iTFRewritePackageOrder[] = {
405 "Package",
406 "Essential",
407 "Status",
408 "Priority",
409 "Section",
410 "Installed-Size",
411 "Maintainer",
412 "Architecture",
413 "Source",
414 "Version",
415 "Revision", // Obsolete
416 "Config-Version", // Obsolete
417 "Replaces",
418 "Provides",
419 "Depends",
420 "Pre-Depends",
421 "Recommends",
422 "Suggests",
423 "Conflicts",
308c7d30 424 "Breaks",
b2e465d6
AL
425 "Conffiles",
426 "Filename",
427 "Size",
428 "MD5Sum",
cde41ae8
MV
429 "SHA1",
430 "SHA256",
b2e465d6
AL
431 "MSDOS-Filename", // Obsolete
432 "Description",
433 0};
434static const char *iTFRewriteSourceOrder[] = {"Package",
435 "Source",
436 "Binary",
437 "Version",
438 "Priority",
439 "Section",
440 "Maintainer",
441 "Build-Depends",
442 "Build-Depends-Indep",
443 "Build-Conflicts",
444 "Build-Conflicts-Indep",
445 "Architecture",
446 "Standards-Version",
447 "Format",
448 "Directory",
449 "Files",
450 0};
451
452/* Two levels of initialization are used because gcc will set the symbol
453 size of an array to the length of the array, causing dynamic relinking
454 errors. Doing this makes the symbol size constant */
455const char **TFRewritePackageOrder = iTFRewritePackageOrder;
456const char **TFRewriteSourceOrder = iTFRewriteSourceOrder;
457
458bool TFRewrite(FILE *Output,pkgTagSection const &Tags,const char *Order[],
459 TFRewriteData *Rewrite)
460{
461 unsigned char Visited[256]; // Bit 1 is Order, Bit 2 is Rewrite
462 for (unsigned I = 0; I != 256; I++)
463 Visited[I] = 0;
464
465 // Set new tag up as necessary.
466 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
467 {
468 if (Rewrite[J].NewTag == 0)
469 Rewrite[J].NewTag = Rewrite[J].Tag;
470 }
471
472 // Write all all of the tags, in order.
473 for (unsigned int I = 0; Order[I] != 0; I++)
474 {
475 bool Rewritten = false;
476
477 // See if this is a field that needs to be rewritten
478 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
479 {
480 if (strcasecmp(Rewrite[J].Tag,Order[I]) == 0)
481 {
482 Visited[J] |= 2;
483 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
484 {
485 if (isspace(Rewrite[J].Rewrite[0]))
486 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
487 else
488 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
489 }
490
491 Rewritten = true;
492 break;
493 }
494 }
495
496 // See if it is in the fragment
497 unsigned Pos;
498 if (Tags.Find(Order[I],Pos) == false)
499 continue;
500 Visited[Pos] |= 1;
501
502 if (Rewritten == true)
503 continue;
504
505 /* Write out this element, taking a moment to rewrite the tag
506 in case of changes of case. */
507 const char *Start;
508 const char *Stop;
509 Tags.Get(Start,Stop,Pos);
510
511 if (fputs(Order[I],Output) < 0)
512 return _error->Errno("fputs","IO Error to output");
513 Start += strlen(Order[I]);
514 if (fwrite(Start,Stop - Start,1,Output) != 1)
515 return _error->Errno("fwrite","IO Error to output");
516 if (Stop[-1] != '\n')
517 fprintf(Output,"\n");
518 }
519
520 // Now write all the old tags that were missed.
521 for (unsigned int I = 0; I != Tags.Count(); I++)
522 {
523 if ((Visited[I] & 1) == 1)
524 continue;
525
526 const char *Start;
527 const char *Stop;
528 Tags.Get(Start,Stop,I);
529 const char *End = Start;
530 for (; End < Stop && *End != ':'; End++);
531
532 // See if this is a field that needs to be rewritten
533 bool Rewritten = false;
534 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
535 {
536 if (stringcasecmp(Start,End,Rewrite[J].Tag) == 0)
537 {
538 Visited[J] |= 2;
539 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
540 {
541 if (isspace(Rewrite[J].Rewrite[0]))
542 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
543 else
544 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
545 }
546
547 Rewritten = true;
548 break;
549 }
550 }
551
552 if (Rewritten == true)
553 continue;
554
555 // Write out this element
556 if (fwrite(Start,Stop - Start,1,Output) != 1)
557 return _error->Errno("fwrite","IO Error to output");
558 if (Stop[-1] != '\n')
559 fprintf(Output,"\n");
560 }
561
562 // Now write all the rewrites that were missed
563 for (unsigned int J = 0; Rewrite != 0 && Rewrite[J].Tag != 0; J++)
564 {
565 if ((Visited[J] & 2) == 2)
566 continue;
567
568 if (Rewrite[J].Rewrite != 0 && Rewrite[J].Rewrite[0] != 0)
569 {
570 if (isspace(Rewrite[J].Rewrite[0]))
571 fprintf(Output,"%s:%s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
572 else
573 fprintf(Output,"%s: %s\n",Rewrite[J].NewTag,Rewrite[J].Rewrite);
574 }
575 }
576
577 return true;
578}
579 /*}}}*/