Annotation of libwww/ChangeLog, revision 1.43
1.41 kahan 1: Changes with libwww 5.3.2
2:
3: 2002-06-05 Jose Kahan <jose@w3.org>
4:
1.42 kahan 5: * The Library/User/Extrnals.html file (a list of all exported
6: functions and variables) wasn't being generated with the
7: consequence that make failed.
8:
9: 2002-06-05 Jose Kahan <jose@w3.org>
10:
1.41 kahan 11: * Updated the auto-tools files in ~/libwww and ~/libwww/config
1.43 ! kahan 12: to the latest versions (autoconf 2.13, libtool 1.4.2a, automake
! 13: 1.4-p4). If when compiling you see warning messages saying:
1.42 kahan 14: libtool: test =: unary operator expected
15: it means you're running an old version of libtool. If you
16: cannot update this tool, then, go to the libwww directory and type:
17: libtoolize -c -f
18: then bootstrap the library again. This will fix your problem.
1.41 kahan 19:
20: 2002-06-05 Richard Atterer <libwww@list02.atterer.net>
21:
22: * Patched configure.in and wwwsys.html for building libwww
23: on mingw32 (a version of GCC that doesn't make produced
24: binaries depend on the Cygwin DLL).
25: See http://lists.w3.org/Archives/Public/www-lib/2002AprJun/0053.html.
1.40 kahan 26:
27: 2002-06-04 Thurog <thurog@gmx.de>
28:
29: * Patched configure.in to avoid a broken configure error in
30: autoconf 2.5.3.
31: See http://lists.w3.org/Archives/Public/www-lib/2002AprJun/0034.html.
32:
33: 2002-05-29 Manuele Kirsch <mkirsch@terra.com.br>
1.39 kirschpi 34:
35: * Fixes for windows plataform concerning WebDAV and Extension
36: methods. In HTMethod and HTRequest, functions defined for
37: the Extension Methods are now defined always, but return
38: fail values when HT_EXT is not defined. In addition, the
39: files "Library/src/WWWDAV.html" and "Library/src/windows/wwwdav.files"
40: have been added. These files and modifications were needed
41: to produce the correct "*.def" files, for windows plataform.
42:
1.40 kahan 43: 2002-05-29 Jose Kahan <jose@w3.org>
1.39 kirschpi 44:
45: * Created the necessary makefiles and changed the windows
46: project files for WebDAV and Extension methods.
47:
1.40 kahan 48: 2002-05-28 Manuele Kirsch <mkirsch@terra.com.br>
1.38 kirschpi 49:
50: * Minor fixes concerning WebDAV module and Extension methods. At
51: src/HTTP.c source file, the 207 status codes has now the result
1.39 kirschpi 52: HT_MULTI_STATUS (line 310), and in Examples/davsample.c,
1.38 kirschpi 53: options "-l", "-c" and "-m" can not anymore produce "Segment dump"
1.39 kirschpi 54: error when an command line argument is missing.
55:
56: * Concerning to Extension Methods, the definitions of
57: HTMethod_hasEntity macro into HTMethods.html have been changed
58: to improve the code readability.
1.38 kirschpi 59:
1.37 kirschpi 60: 2002-03-19 Manuele Kirsch <Manuele.Kirsch_Pinheiro@inrialpes.fr>
61:
62: * The WebDAV module has been changed to use HTRequest entity
63: and PostCallback to send all message bodies in its requests.
64:
65: * Extension methods have been changed - now the application
66: may chose, when register an extension method, between use
67: "Expect: 100-continue" heander and HTRequest Entity, or the
68: new message body present in HTRequest. When using the first
69: one, the application must set a PostCallback, and when using
70: the second, it may set a small message body to be sent with
71: the request. This last may be used to other protocols, other
72: than HTTP.
73:
74: * Added a new example (myext2.c) that shows how to set the
75: extension methods and how to use it with "Expect: 100-contine"
76: and with message body.
77:
78: 2002-03-08 Manuele Kirsch <Manuele.Kirsch_Pinheiro@inrialpes.fr>
79:
80: * Added new methods in HTMethod enumeration. There are the 7
81: methods defined in WebDAV protocol (LOCK,UNLOCK,PROPFIND,
82: PROPPATCH,MKCOL,COPY and MOVE), and also 7 extension methods
83: (METHOD_EXT_0 to METHOD_EXT_6). These methods are undefined
84: methods that can be set/unset for the applications using two
85: new functions introduced to HTMethod module
86: ("HTMethod_setExtensionMethod" and "HTMethod_deleteExtensionMethod").
87: Those extension methods are available only if configure used
88: --with-extension option (which define HT_EXT).
89:
90: * Added new fields to HTRequest structure: a message body array,
91: its length and format. HTTPGenMake functions in HTTPGen.c has
92: been changed to include Content-Type and Content-Lenght headers
93: if messageBody field and messageBodyType/messageBodyFormat are
94: set. It also send this message body only if it is set. This
95: message body may be use for XML message bodies (indeed, it is
96: used by some DAV methods in HTDAV module). These modifications
97: are activated only if HT_DAV or HT_EXT are defined.
98:
99: * Added a new module HTDAV with high level functions for WebDAV
100: protocol. It is defined in HTDAV.h/html and implemented in
101: HTDAV.c. This is only enabled with "--with-dav" option in
102: configure procedure (this option will define HT_DAV). An example
103: of use (davsample.c) have been included in Library/Examples.
104:
105: * Added new status codes (in HTUtils.h/html) and error codes (in
106: HTError.h/html) defined by WebDAV protocol (102,207,422,423,424
107: and 507). This codes are activated only if HT_DAV is defined.
1.36 kahan 108:
109: 2002-03-21 Jose Kahan <jose@w3.org>
110:
111: * Added the HTDNS.c configure patches for the reentrant call to
112: gethostbynamer function. (sorry, I can't find the name of
113: the person who contributed this patch right now).
114:
1.32 kahan 115: 2001-08-30 Jose Kahan <jose@w3.org>
116:
117: * Added a new field in the HTRequest structure for specifying
118: a user defined Content-Location. This allows to use the library
119: functions to access a resource, while in fact the request sent
120: to the server is made to the value of the user defined
121: Content-Location. This sound silly, but it allows to keep the
122: libwww state coherent and it doesn't break the requests like
123: the previous (and banned since long time) real content-location
124: handling. To avoid confusion, this field is called
125: default_put_name. See HTReqMan.h for the name of the functions
126: that handle this field.
127:
1.37 kirschpi 128: * New public function in HTCache.c to reset the value of a cache
1.32 kahan 129: entry and delete the cached file. This is interesting when
130: we want to keep a cache entry, but force the cache entry to be
131: refreshed. HTCache_resetMeta ().
132:
1.31 kahan 133: 2001-08-14 Jose Kahan <jose@w3.org>
134:
135: * HTProfile_newPreemptiveClient wasn't intializing the
136: default event loop in all cases (bug reported by
137: Olivier Steinau).
138:
1.30 kahan 139: 2001-03-14 Jose Kahan <jose@w3.org>
1.29 kahan 140:
1.30 kahan 141: * Digest authentication didn't work well with pipelining requests.
142: Only one of such requests got the stale nonce. The other 401
143: were being re-interpreted as if the user had given a wrong
144: login name or password. To fix this, we try to infer a stale
145: nonce if the previous nonce and the new one sent by the server
146: are different and a wrong login name/password if they are the
147: same. The patch is in HTAABrow.c.
148: I also added a second bug fix in the same patch as not all
149: pipelined requests seem to go thru the first authentication filter.
150:
1.29 kahan 151: 2001-03-09 Jose Kahan <jose@w3.org>
1.20 kahan 152:
1.29 kahan 153: * Digest authentication wasn't being compiled by default in Win32
154: because windows/config.h was missing the HT_MD5 define.
155:
1.28 kahan 156: 2001-03-06 Serge Adda <sAdda@infovista.com>
157:
158: * HTAnchor bug. HTAnchor_delete wasn't removing the reference to
159: the deleted anchor. This caused a bug whenever requesting another
160: anchor for the same URL. Thanks to Anton Belov for having
161: reported the problem and to Serge for his patch. See
162: http://lists.w3.org/Archives/Public/www-lib/2001JanMar/0133.html.
163:
164:
1.27 kahan 165: 2001-03-06 Stefan Wiesner <sw@ais-ag.de>
166:
167: * Timer bug. There was a bug in the code of function HTTimer_new
168: if a timer is already expired when HTTimer_new is called. See
169: http://lists.w3.org/Archives/Public/www-lib/2001JanMar/0123.html
170: for the full analysis.
171:
172:
1.26 kahan 173: 2001-02-26 Jose Kahan <jose@w3.org>
174:
175: * Cache bug. An HTResponse object was being freed repetively inside
176: a while loop, when reading a cached metadata in meta_read(). I
177: moved the delete to outside this function.
178:
1.29 kahan 179: Changes with libwww 5.3.2
180:
1.23 kahan 181: 2000-12-19 Jose Kahan <jose@w3.org>
182:
183: * Compilation fix: two prototypes in HTResponse.html were declared
184: as PUBLIC, instead of as external.
1.24 kahan 185: Updated the copyright date of the WinCommander
1.25 kahan 186: Fixed runtime crashes with the following example applications:
187: chunk.c chunkbody.c getheaders.c head.c multichunk.c
1.23 kahan 188:
1.22 kahan 189: 2000-12-18 Jose Kahan <jose@w3.org>
190:
191: * Compilation fix: The HTMIME_anchor2response() function
192: added a dependency between the cache and mime mini-libs.
193: I changed the call to a stream converter (HTCacheCopyHeaders)
194: to make it optional and to avoid the dependency. The converter
195: is initialized in HTInit.c:HTConverterInit.
196:
1.21 kahan 197: 2000-12-14 Jose Kahan <jose@w3.org>
198:
199: * Bug fix: make dist didn't work because of a recurrent rule
200: in the ComLine make file (it tried to use www to create a
201: documentation file).
202:
1.20 kahan 203: 2000-12-13 Jose Kahan <jose@w3.org>
1.19 kahan 204:
205: * Bug fix: When objects were retrieved from the cache, the
206: response object didn't have a copy of the request headers.
207: These headers were only found in the anchor associated to
208: the request. In order to make a coherent behavior, I made
209: a new function HTMIME.c:HTMIME_anchor2response() that copies
210: the headers from one object to the other. Well, more than a
211: copy, it adds a link to the headers. I call this function in
212: HTCache:CacheCleanup when it succesfully retrieved an object
1.37 kirschpi 213: from the cache. Result, you can now use the response object
214: in both cases.
1.19 kahan 215:
1.20 kahan 216: 2000-10-30 Jose Kahan <jose@w3.org>
1.18 kahan 217:
218: * Enhancement, added a new field to the HTResponse object to
219: store the HTTP reason sent by the server. Two new functions to
220: get it/set it: HTResponse_reason(), HTResponse_setReason()
221:
1.16 kahan 222: 2000-09-01 Wayne Davison <wayne@clari.net>
223:
224: * HTAccess.c: Typo fix
1.17 kahan 225: * HTChunk.c, HTChunk.html: new function, HTChunk_fromBuffer for
226: creating a chunk from an existing block of memory, without having
227: to allocate more memory.
1.16 kahan 228:
1.15 kahan 229: 2000-09-01 Peter Stamfest <peter.stamfest@eunet.at>
230:
1.16 kahan 231: * Makefile.am: corrected a DESTDIR problem
1.15 kahan 232:
1.14 kahan 233: 2000-09-01 Jose Kahan <jose@w3.org>
234:
1.16 kahan 235: * config.h: Updated the libwww version
1.14 kahan 236:
1.13 kahan 237: 2000-09-01 Zhu Qun-Ying <qunying@yahoo.com>
238:
1.16 kahan 239: * config.h: BUG FIX: wrong closing comment
1.13 kahan 240:
1.12 barstow 241: 2000-08-30 Arthur Barstow <barstow@w3.org>
242:
243: * HTRDF.c: BUG FIX: nested typedNode constructs with
244: attributes caused re-using a free'd pointer.
245:
246: * HTRDF.c: per a request from the Redland RDF project, added
247: a void * context parameter to HTRDF_parse{File,Buffer}.
248: The application-specific context will be available in the
249: triple callback handler.
250:
1.37 kirschpi 251: * HTRDF.html - updated to API changes mentioned above
1.12 barstow 252:
253: * Library/Examples/ updated rdf_parse_{file,buffer} programs
254: to reflect the above API changes.
255:
1.11 kahan 256: 2000-08-29 Jose Kahan <jose@w3.org>
257:
258: * Temporarily commented the XML structured stream in HTXML.c
259: (HTXMLStructured_new) and its related code, as it creates
260: a dependency with the libwwwxml mini-library. If this code
261: is still useful, it should go into another module, HTSXML.c
262: to avoid the forced dependency.
263:
264: 2000-08-28 Jose Kahan <jose@w3.org>
265:
266: * Updated expat to the test version 19990728, as this version
267: adds some interesting API extensions and is now being widely
268: used by other projects, including the Perl XML modules. The
269: previous version is CVS tagged as expat-1-1.
270:
1.10 barstow 271: 2000-08-18 Arthur Barstow <barstow@w3.org>
272:
273: * HTRDF.c: Added a new public API HTRDF_parseBuffer. It
1.37 kirschpi 274: parses an arbitrary buffer (char *) of RDF.
1.10 barstow 275:
276: * HTRDF.c: Changed the name of HTRDFParseFile to HTRDF_parseFile.
1.37 kirschpi 277: it now returns BOOL instead of char * to be consistent with
278: other RDF parser APIs.
1.10 barstow 279:
1.37 kirschpi 280: * HTRDF.html - updated to HTRDF.c changes mentioned above
1.10 barstow 281:
1.37 kirschpi 282: * Library/Examples/: added two new example programs:
1.10 barstow 283:
1.37 kirschpi 284: 1. rdf_parse_file - demos/tests HTRDF_parseFile()
285: 2. rdf_parse_buffer - demos/tests HTRDF_parseBuffer()
1.10 barstow 286:
1.37 kirschpi 287: * configure.in - added the two new programs above
1.10 barstow 288:
1.37 kirschpi 289: * Library/Examples/Makefile.am - added the two new programs
1.10 barstow 290:
1.9 barstow 291: 2000-08-17 Arthur Barstow <barstow@w3.org>
292:
293: * HTRDF.c: Fixed a memory leak. (HTElement structures for elements
294: within a property with parseType="Literal" were not free'd
295: after the parser was done.)
296:
297: * HTRDF.c: Bug fixes. Added code to circumvent calls to HTSAlloc
298: (via StrAllocCopy) when the src and dest were identical (this
299: causes a free'd pointer to be used) or both NULL.
300:
301: * HTRDF.c: Changed addMarkup{Add,End} to make one call to
302: StrAllocMCat instead of multiple StrAllocCat calls.
303:
1.7 kahan 304: 2000-08-11 Jose Kahan <jose@w3.org>
305:
306: * Memory leak. The HTPlain (HTPlain.c) stream wasn't freeing the
307: HTTextImp (me->text) object in the HTPlain_free and
308: HTPlain_abort. I found the bug while running the robot
309: thru Insure and found the fix by seeing what was being
310: done in the HTML (HTML.c) stream.
1.8 kahan 311:
312: * The make dist command wasn't copying the .c, .files, and .mak
313: files found in the Library/src/windows. As a result, the tar files
314: created with this command were missing those files. The problem
315: came from some wrong assumptions in the Makefile.am of the
316: windows directory.
1.7 kahan 317:
1.6 kahan 318: 2000-08-11 Wayne Davison <wayne@clari.net>
319:
320: * Added two new functions to HTHash.c, HTHashtable_removeObject()
321: and HTHashTable_walk. The first one allows to delete individual
322: elements from a hash-table, the second one allows to walk
323: throug all associated elements in a table without having to ask
324: for a list of all keys and then looking each one up. See
325: http://lists.w3.org/Archives/Public/www-lib/2000JulSep/0146.html
326: for a more detaile description on how to use these functions.
327:
328: * I got rid of a few useless variable initializations in HTHash.c
329: (when the variables were being set to a value immediately
330: following the init value).
331:
332: * I removed some superfluous spaces that I saw.
333:
1.5 kahan 334: 2000-08-10 Wayne Davison <wayne@clari.net>
335:
1.6 kahan 336: * Since "cvs update" was complaining about a bunch of generated
337: files in the Library/src dir, I tweaked it to ignore the *.la
338: files and the .libs dir.
1.5 kahan 339:
1.4 kahan 340: 2000-08-10 Jose Kahan <jose@w3.org>
341:
1.6 kahan 342: * Because (IMO) of a very imbricated protocol stack, the Robot
1.4 kahan 343: could hang when accessing local files, as it become starved
1.6 kahan 344: on its own sockets. I fixed the problem by adding a new
1.4 kahan 345: state in HTFile:FileEvent() to take into account the case when
346: a request goes pending. This is mostly based on a previous
347: (not commited) patch by Kinuko Yasuda, plus some long debugging to
348: find out what was happening and a fix to a memory bug that the
349: other patch had.
350:
1.3 kahan 351: 2000-08-08 Wayne Davison <wayne@clari.net>
352:
353: * New HTChunk_setSize function to allow to set the size of a
354: chunk object upwards or downwards. (HTChunk_truncate() still
1.37 kirschpi 355: exists and still enforces the sizing of the string downwards.)
356: Both these string-sizing functions only zero the relevant bytes
357: in the "size" region (if the string is being shortened). Note
358: that HTChunk_clear() was left unchanged -- it still clears all of
359: the Chunk's allocated memory.
1.3 kahan 360:
1.37 kirschpi 361: * Fixed the HT_OUTOFMEM() calls in HTChunk_ensure() to have the
362: proper location string.
1.3 kahan 363:
1.37 kirschpi 364: * Some code in HTBind.c was optimized to use HTChunk_truncate().
1.3 kahan 365:
1.37 kirschpi 366: * HTFTP.c and HTNews.c was optimized to use HTChunk_setSize()
1.3 kahan 367: rather than using HTChunk_ensure() and writing data past the "size".
368:
1.37 kirschpi 369: * Optimized the code in HTMIME.c and HTML.c to use
1.3 kahan 370: HTChunk_truncate(chunk,0) rather than HTChunk_clear(chunk).
371:
372: * Documented the new HTChunk_setSize() function.
373:
374: * Improved the comments for HTChunk_clear().
375:
376: * Fixed the comments for HTChunk_ensure() (it was describing the
377: extra-size parameter incorrectly).
378:
379: * Fixed the comments for HTChunk_toString() (to indicate that the
380: function destroys the chunk when handing off the CString data).
381:
382: * Fixed a few spelling errors in HTChunk.html.
383:
384: * Note: There are probably other places in the code that could be
385: optimized to use truncate rather than clear, but I didn't have
386: time to check each one to see if the chunk's data was properly
387: respecting the size indicator or not.
388:
1.2 kahan 389: 2000-08-09 Jose Kahan <jose@w3.org>
390:
391: * Updated the "how to bootstrap section" with instructions on what to
392: do if libtool and aclocal are installed in different directories.
393: Thanks to Ken Olum and James Henstrige for the tips.
394:
1.1 kahan 395: 2000-08-08 Jens Meggers <jens@meggers.com>
396:
397: * The cache garbage collector (HTCacheGarbage()) could go into an
398: endless loop when it tried to removed locked entries. The problem
399: was that it wasn't checking the return code of HTCache_remove().
400:
401: 2000-08-07 Wayne Davison <wayne@clari.net>
402:
403: * Changed the macro argument from "data" to "dp" in
404: HTArray_firstObject() and HTArray_nextObject(). These macros
405: are defined such that if the caller doesn't pass in a variable
406: named exactly "data" as the second parameter, there will either be a
407: syntax error generated during the compilation OR the wrong structure
408: element will get referenced. This is because the macro argument
409: "data" was the same name as the HTArray structure element that was
1.11 kahan 410: being referenced in the macro.
1.20 kahan 411:
412: Changes with previous versions
413:
414: * No info available (this log was started with version 5.3.2)
Webmaster