]>
Commit | Line | Data |
---|---|---|
1 | // -*- mode: cpp; mode: fold -*- | |
2 | // Description /*{{{*/ | |
3 | // $Id: contents.cc,v 1.4 2003/02/10 07:34:41 doogie Exp $ | |
4 | /* ###################################################################### | |
5 | ||
6 | contents - Archive contents generator | |
7 | ||
8 | The GenContents class is a back end for an archive contents generator. | |
9 | It takes a list of per-deb file name and merges it into a memory | |
10 | database of all previous output. This database is stored as a set | |
11 | of binary trees linked across directories to form a tree of all files+dirs | |
12 | given to it. The tree will also be sorted as it is built up thus | |
13 | removing the massive sort time overhead. | |
14 | ||
15 | By breaking all the pathnames into components and storing them | |
16 | separately a space saving is realized by not duplicating the string | |
17 | over and over again. Ultimately this saving is sacrificed to storage of | |
18 | the tree structure itself but the tree structure yields a speed gain | |
19 | in the sorting and processing. Ultimately it takes about 5 seconds to | |
20 | do 141000 nodes and about 5 meg of ram. | |
21 | ||
22 | The tree looks something like: | |
23 | ||
24 | usr/ | |
25 | / \ / libslang | |
26 | bin/ lib/ --> libc6 | |
27 | / \ \ libfoo | |
28 | games/ sbin/ | |
29 | ||
30 | The ---> is the DirDown link | |
31 | ||
32 | ||
33 | ##################################################################### */ | |
34 | /*}}}*/ | |
35 | // Include Files /*{{{*/ | |
36 | #include "contents.h" | |
37 | ||
38 | #include <apti18n.h> | |
39 | #include <apt-pkg/debfile.h> | |
40 | #include <apt-pkg/extracttar.h> | |
41 | #include <apt-pkg/error.h> | |
42 | #include <stdio.h> | |
43 | #include <stdlib.h> | |
44 | #include <string.h> | |
45 | #include <malloc.h> | |
46 | /*}}}*/ | |
47 | ||
48 | // GenContents::~GenContents - Free allocated memory /*{{{*/ | |
49 | // --------------------------------------------------------------------- | |
50 | /* Since all our allocations are static big-block allocations all that is | |
51 | needed is to free all of them. */ | |
52 | GenContents::~GenContents() | |
53 | { | |
54 | while (BlockList != 0) | |
55 | { | |
56 | BigBlock *Old = BlockList; | |
57 | BlockList = Old->Next; | |
58 | free(Old->Block); | |
59 | delete Old; | |
60 | } | |
61 | } | |
62 | /*}}}*/ | |
63 | // GenContents::Mystrdup - Custom strdup /*{{{*/ | |
64 | // --------------------------------------------------------------------- | |
65 | /* This strdup also uses a large block allocator to eliminate glibc | |
66 | overhead */ | |
67 | char *GenContents::Mystrdup(const char *From) | |
68 | { | |
69 | unsigned int Len = strlen(From) + 1; | |
70 | if (StrLeft <= Len) | |
71 | { | |
72 | StrLeft = 4096*10; | |
73 | StrPool = (char *)malloc(StrLeft); | |
74 | ||
75 | BigBlock *Block = new BigBlock; | |
76 | Block->Block = StrPool; | |
77 | Block->Next = BlockList; | |
78 | BlockList = Block; | |
79 | } | |
80 | ||
81 | memcpy(StrPool,From,Len); | |
82 | StrLeft -= Len; | |
83 | ||
84 | char *Res = StrPool; | |
85 | StrPool += Len; | |
86 | return Res; | |
87 | } | |
88 | /*}}}*/ | |
89 | // GenContents::Node::operator new - Big block allocator /*{{{*/ | |
90 | // --------------------------------------------------------------------- | |
91 | /* This eliminates glibc's malloc overhead by allocating large blocks and | |
92 | having a continuous set of Nodes. This takes about 8 bytes off each nodes | |
93 | space needs. Freeing is not supported. */ | |
94 | void *GenContents::Node::operator new(size_t Amount,GenContents *Owner) | |
95 | { | |
96 | if (Owner->NodeLeft == 0) | |
97 | { | |
98 | Owner->NodeLeft = 10000; | |
99 | Owner->NodePool = (Node *)malloc(Amount*Owner->NodeLeft); | |
100 | BigBlock *Block = new BigBlock; | |
101 | Block->Block = Owner->NodePool; | |
102 | Block->Next = Owner->BlockList; | |
103 | Owner->BlockList = Block; | |
104 | } | |
105 | ||
106 | Owner->NodeLeft--; | |
107 | return Owner->NodePool++; | |
108 | } | |
109 | /*}}}*/ | |
110 | // GenContents::Grab - Grab a new node representing Name under Top /*{{{*/ | |
111 | // --------------------------------------------------------------------- | |
112 | /* This grabs a new node representing the pathname component Name under | |
113 | the node Top. The node is given the name Package. It is assumed that Name | |
114 | is inside of top. If a duplicate already entered name is found then | |
115 | a note is made on the Dup list and the previous in-tree node is returned. */ | |
116 | GenContents::Node *GenContents::Grab(GenContents::Node *Top,const char *Name, | |
117 | const char *Package) | |
118 | { | |
119 | /* We drop down to the next dir level each call. This simplifies | |
120 | the calling routine */ | |
121 | if (Top->DirDown == 0) | |
122 | { | |
123 | Node *Item = new(this) Node; | |
124 | Item->Path = Mystrdup(Name); | |
125 | Item->Package = Package; | |
126 | Top->DirDown = Item; | |
127 | return Item; | |
128 | } | |
129 | Top = Top->DirDown; | |
130 | ||
131 | int Res; | |
132 | while (1) | |
133 | { | |
134 | Res = strcmp(Name,Top->Path); | |
135 | ||
136 | // Collision! | |
137 | if (Res == 0) | |
138 | { | |
139 | // See if this the the same package (multi-version dup) | |
140 | if (Top->Package == Package || | |
141 | strcasecmp(Top->Package,Package) == 0) | |
142 | return Top; | |
143 | ||
144 | // Look for an already existing Dup | |
145 | for (Node *I = Top->Dups; I != 0; I = I->Dups) | |
146 | if (I->Package == Package || | |
147 | strcasecmp(I->Package,Package) == 0) | |
148 | return Top; | |
149 | ||
150 | // Add the dup in | |
151 | Node *Item = new(this) Node; | |
152 | Item->Path = Top->Path; | |
153 | Item->Package = Package; | |
154 | Item->Dups = Top->Dups; | |
155 | Top->Dups = Item; | |
156 | return Top; | |
157 | } | |
158 | ||
159 | // Continue to traverse the tree | |
160 | if (Res < 0) | |
161 | { | |
162 | if (Top->BTreeLeft == 0) | |
163 | break; | |
164 | Top = Top->BTreeLeft; | |
165 | } | |
166 | else | |
167 | { | |
168 | if (Top->BTreeRight == 0) | |
169 | break; | |
170 | Top = Top->BTreeRight; | |
171 | } | |
172 | } | |
173 | ||
174 | // The item was not found in the tree | |
175 | Node *Item = new(this) Node; | |
176 | Item->Path = Mystrdup(Name); | |
177 | Item->Package = Package; | |
178 | ||
179 | // Link it into the tree | |
180 | if (Res < 0) | |
181 | { | |
182 | Item->BTreeLeft = Top->BTreeLeft; | |
183 | Top->BTreeLeft = Item; | |
184 | } | |
185 | else | |
186 | { | |
187 | Item->BTreeRight = Top->BTreeRight; | |
188 | Top->BTreeRight = Item; | |
189 | } | |
190 | ||
191 | return Item; | |
192 | } | |
193 | /*}}}*/ | |
194 | // GenContents::Add - Add a path to the tree /*{{{*/ | |
195 | // --------------------------------------------------------------------- | |
196 | /* This takes a full pathname and adds it into the tree. We split the | |
197 | pathname into directory fragments adding each one as we go. Technically | |
198 | in output from tar this should result in hitting previous items. */ | |
199 | void GenContents::Add(const char *Dir,const char *Package) | |
200 | { | |
201 | Node *Root = &this->Root; | |
202 | ||
203 | // Drop leading slashes | |
204 | while (*Dir == '/' && *Dir != 0) | |
205 | Dir++; | |
206 | ||
207 | // Run over the string and grab out each bit up to and including a / | |
208 | const char *Start = Dir; | |
209 | const char *I = Dir; | |
210 | while (*I != 0) | |
211 | { | |
212 | if (*I != '/' || I - Start <= 1) | |
213 | { | |
214 | I++; | |
215 | continue; | |
216 | } | |
217 | I++; | |
218 | ||
219 | // Copy the path fragment over | |
220 | char Tmp[1024]; | |
221 | strncpy(Tmp,Start,I - Start); | |
222 | Tmp[I - Start] = 0; | |
223 | ||
224 | // Grab a node for it | |
225 | Root = Grab(Root,Tmp,Package); | |
226 | ||
227 | Start = I; | |
228 | } | |
229 | ||
230 | // The final component if it does not have a trailing / | |
231 | if (I - Start >= 1) | |
232 | Root = Grab(Root,Start,Package); | |
233 | } | |
234 | /*}}}*/ | |
235 | // GenContents::WriteSpace - Write a given number of white space chars /*{{{*/ | |
236 | // --------------------------------------------------------------------- | |
237 | /* We mod 8 it and write tabs where possible. */ | |
238 | void GenContents::WriteSpace(FILE *Out,unsigned int Current,unsigned int Target) | |
239 | { | |
240 | if (Target <= Current) | |
241 | Target = Current + 1; | |
242 | ||
243 | /* Now we write tabs so long as the next tab stop would not pass | |
244 | the target */ | |
245 | for (; (Current/8 + 1)*8 < Target; Current = (Current/8 + 1)*8) | |
246 | fputc('\t',Out); | |
247 | ||
248 | // Fill the last bit with spaces | |
249 | for (; Current < Target; Current++) | |
250 | fputc(' ',Out); | |
251 | } | |
252 | /*}}}*/ | |
253 | // GenContents::Print - Display the tree /*{{{*/ | |
254 | // --------------------------------------------------------------------- | |
255 | /* This is the final result function. It takes the tree and recursively | |
256 | calls itself and runs over each section of the tree printing out | |
257 | the pathname and the hit packages. We use Buf to build the pathname | |
258 | summed over all the directory parents of this node. */ | |
259 | void GenContents::Print(FILE *Out) | |
260 | { | |
261 | char Buffer[1024]; | |
262 | Buffer[0] = 0; | |
263 | DoPrint(Out,&Root,Buffer); | |
264 | } | |
265 | void GenContents::DoPrint(FILE *Out,GenContents::Node *Top, char *Buf) | |
266 | { | |
267 | if (Top == 0) | |
268 | return; | |
269 | ||
270 | // Go left | |
271 | DoPrint(Out,Top->BTreeLeft,Buf); | |
272 | ||
273 | // Print the current dir location and then descend to lower dirs | |
274 | char *OldEnd = Buf + strlen(Buf); | |
275 | if (Top->Path != 0) | |
276 | { | |
277 | strcat(Buf,Top->Path); | |
278 | ||
279 | // Do not show the item if it is a directory with dups | |
280 | if (Top->Path[strlen(Top->Path)-1] != '/' /*|| Top->Dups == 0*/) | |
281 | { | |
282 | fputs(Buf,Out); | |
283 | WriteSpace(Out,strlen(Buf),60); | |
284 | for (Node *I = Top; I != 0; I = I->Dups) | |
285 | { | |
286 | if (I != Top) | |
287 | fputc(',',Out); | |
288 | fputs(I->Package,Out); | |
289 | } | |
290 | fputc('\n',Out); | |
291 | } | |
292 | } | |
293 | ||
294 | // Go along the directory link | |
295 | DoPrint(Out,Top->DirDown,Buf); | |
296 | *OldEnd = 0; | |
297 | ||
298 | // Go right | |
299 | DoPrint(Out,Top->BTreeRight,Buf); | |
300 | } | |
301 | /*}}}*/ | |
302 | ||
303 | // ContentsExtract::Read - Read the archive /*{{{*/ | |
304 | // --------------------------------------------------------------------- | |
305 | /* */ | |
306 | bool ContentsExtract::Read(debDebFile &Deb) | |
307 | { | |
308 | Reset(); | |
309 | return Deb.ExtractArchive(*this); | |
310 | } | |
311 | /*}}}*/ | |
312 | // ContentsExtract::DoItem - Extract an item /*{{{*/ | |
313 | // --------------------------------------------------------------------- | |
314 | /* This just tacks the name onto the end of our memory buffer */ | |
315 | bool ContentsExtract::DoItem(Item &Itm,int &Fd) | |
316 | { | |
317 | unsigned long Len = strlen(Itm.Name); | |
318 | ||
319 | // Strip leading ./'s | |
320 | if (Itm.Name[0] == '.' && Itm.Name[1] == '/') | |
321 | { | |
322 | // == './' | |
323 | if (Len == 2) | |
324 | return true; | |
325 | ||
326 | Len -= 2; | |
327 | Itm.Name += 2; | |
328 | } | |
329 | ||
330 | // Allocate more storage for the string list | |
331 | if (CurSize + Len + 2 >= MaxSize || Data == 0) | |
332 | { | |
333 | if (MaxSize == 0) | |
334 | MaxSize = 512*1024/2; | |
335 | char *NewData = (char *)realloc(Data,MaxSize*2); | |
336 | if (NewData == 0) | |
337 | return _error->Error(_("realloc - Failed to allocate memory")); | |
338 | Data = NewData; | |
339 | MaxSize *= 2; | |
340 | } | |
341 | ||
342 | strcpy(Data+CurSize,Itm.Name); | |
343 | CurSize += Len + 1; | |
344 | return true; | |
345 | } | |
346 | /*}}}*/ | |
347 | // ContentsExtract::TakeContents - Load the contents data /*{{{*/ | |
348 | // --------------------------------------------------------------------- | |
349 | /* */ | |
350 | bool ContentsExtract::TakeContents(const void *NewData,unsigned long Length) | |
351 | { | |
352 | if (Length == 0) | |
353 | { | |
354 | CurSize = 0; | |
355 | return true; | |
356 | } | |
357 | ||
358 | // Allocate more storage for the string list | |
359 | if (Length + 2 >= MaxSize || Data == 0) | |
360 | { | |
361 | if (MaxSize == 0) | |
362 | MaxSize = 512*1024/2; | |
363 | while (MaxSize*2 <= Length) | |
364 | MaxSize *= 2; | |
365 | ||
366 | char *NewData = (char *)realloc(Data,MaxSize*2); | |
367 | if (NewData == 0) | |
368 | return _error->Error(_("realloc - Failed to allocate memory")); | |
369 | Data = NewData; | |
370 | MaxSize *= 2; | |
371 | } | |
372 | memcpy(Data,NewData,Length); | |
373 | CurSize = Length; | |
374 | ||
375 | return Data[CurSize-1] == 0; | |
376 | } | |
377 | /*}}}*/ | |
378 | // ContentsExtract::Add - Read the contents data into the sorter /*{{{*/ | |
379 | // --------------------------------------------------------------------- | |
380 | /* */ | |
381 | void ContentsExtract::Add(GenContents &Contents,string const &Package) | |
382 | { | |
383 | const char *Start = Data; | |
384 | char *Pkg = Contents.Mystrdup(Package.c_str()); | |
385 | for (const char *I = Data; I < Data + CurSize; I++) | |
386 | { | |
387 | if (*I == 0) | |
388 | { | |
389 | Contents.Add(Start,Pkg); | |
390 | Start = ++I; | |
391 | } | |
392 | } | |
393 | } | |
394 | /*}}}*/ |