Sun Nov 9 14:44:02 EST 2003 Gabriele Bartolini * Tagged release htdig-3-2-0b5 Sat Nov 8 2003 Lachlan Andrew * htcommon/default.cc, htsearch/parser.cc: Fix bug #825877 Reduce backlink_factor to comparable with other factors, and interpret multimatch_factor as the *bonus* given for multiple matches. Sat Nov 1 2003 Lachlan Andrew * htsearch/parser.cc: Fix bug #806419. Ignore bad words at start of phrase. Tue Oct 28 11:58:06 EST 2003 Gabriele Bartolini * htdig/htdig.cc: set the debug level when we are importing a cookie file. Fix bug #831478. Mon Oct 27 17:13:02 2003 Gilles Detillieux * htdig/Server.cc: Fix bug #831407. Make sure time properly reset after delay completed, so that it doesn't allow 2 connections per delay. Mon Oct 27 15:57:38 2003 Gilles Detillieux * htdoc/THANKS.html: Added Lachlan, Jim and Neal to the active developers list. Sun Oct 26 2003 Lachlan Andrew * htdoc/hts_templates.html: Clarify that PREV/NEXTPAGE template variables are empty if there is only one page, ignoring no_{prev,next}_page_text. Sun Oct 26 2003 Lachlan Andrew * htcommon/defaults.cc: Fixed documentation to close bug #829767 Clarified that noindex_start/end do not get replaced by whitespace. Also removed spurious '>' from start of boolean_syntax_errors, and added missing '#' to many local tags. Sun Oct 26 12:42:27 EST 2003 Gabriele Bartolini * htcommon/defaults.cc: Fixed description of 'head_before_get' after Lachlan fixes. * htdoc/attrs.html: rerun cf_generate.pl Sat Oct 25 2003 Lachlan Andrew * htsearch/Display.cc: Fix #829761. If last component of the URL is used as a title, URL-decode it. Sat Oct 25 2003 Lachlan Andrew * htdig/Server.cc: Fix #829754. Avoid calculations with negative time Fri Oct 24 17:17:15 2003 Gilles Detillieux * htdoc/htdig.html, htdoc/meta.html, htdoc/require.html: Update URL for the Standard for Robot Exclusion. * htdoc/htmerge.html: Added two clarifications to -m option description. * htdoc/cf_types.html: Make clear distinction between String List and Quoted String List. Fri Oct 24 15:30:08 2003 Gilles Detillieux * htsearch/Display.cc: Fix bug #829746. Applied Niel Kohl's fix for this, to check if words input given before trying to use it, to avoid NULL argument to syslog(). Fri Oct 24 15:15:53 2003 Gilles Detillieux * htsearch/Display.cc: Fix bug #578570. The enddate handling now works correctly for a large, negative startday value. Fri Oct 24 12:47:51 2003 Gilles Detillieux * htdig/HTML.cc (ctor): Fix obvious typo in metadatetags.Pattern setting. Thu Oct 23 10:27:18 2003 Lachlan Andrew * htcommon/default.cc: Fix bug #828808. Default startyear to empty Document "startyear defaults to 1970 if a start/end date set". Thu Oct 23 12:14:30 EST 2003 Gabriele Bartolini * htdig/htdig.cc: restored the code before Oct 21 (fixes ##828628) Thu Oct 23 11:41:15 EST 2003 Gabriele Bartolini * htdig/Retriever.[h,cc]: removed 'head_before_get' overriding by restoring the code before Oct 21. * htdig/Document.[h,cc]: ditto, with the exception of detaching the HEAD before GET mechanism from the persistent connections'. * htcommon/defaults.cc: improved documentation (even though it needs corrections by an english-speaking developer). * These changes fix bug #828628 Wed Oct 22 2003 Lachlan Andrew * htsearch/parser.cc: Applied Neal's patch to fix bug #823403 Documents only added to search list if they were successfully dug. Lines 237-238 of htsearch/Display.cc if (!ref || ref->DocState() != Reference_normal) continue; should now be redundant. (Left in to be defensive.) Tue Oct 21 11:04:56 EST 2003 Gabriele Bartolini * htdig/Retriever.h: added the 'RetrieverType' enum and an object variable for storing the type of dig we are performing (default initial); * htdig/Retriever.cc: changed constructor in order to handle the type, added some debugging explanation regarding the override of the 'head_before_get' attribute, added checks regarding an empty database of URLs to be updated (set the type to initial). * htdig/Document.h: added the attribute 'is_initial' which stores the information regarding the type of indexing (initial or incremental) we are currently performing. Added access methods (get-and-set-like) * htdig/Document.cc: modified the logic of the HeadBeforeGet settings during the retrieval phase, in order to always override user's settings in an incremental dig and automatically set the 'HEAD' call in this case. * htcommon/defaults.cc: modified the default value of 'head_before_get' and a bit of its explanation. * htnet/HtHTTP.cc: detached the HEAD before GET mechanism to the persistent connections one * htdig/Server.cc: added one level of debugging to the display of the server settings in the server constructor Fri Oct 17 2003 Lachlan Andrew * htword/WordType.cc, htcommon/defaults.cc: Patched to fix bug #823083 Don't assume IsStrictChar returns false for digits. Clarify behaviour of allow_numbers in the documentation. Fri Oct 17 2003 Lachlan Andrew * htcommon/defaults.cc: Patched to fix bug #823455 Escaped "$" in valid_punctuation, and add warnings about $, \ and `. Wed Oct 15 11:12:52 2003 Gilles Detillieux * htdig/Server.cc (robotstxt): Patched to fix bug #765726. Don't block paths with subpaths excluded by robots.txt, and make sure any regex meta characters are properly escaped. Tue Oct 14 11:54:07 EST 2003 Gabriele Bartolini * htnet/HtHTTP.cc: add an empty Accept-Encoding header - this inform the server that htdig is only able to manage documents that are not encoded (if no Accept-Encoding is sent, the server assumes that the client is capable of handling every content encoding - i.e. zipped documents with Apache's mod_gzip module). Partial fix of bug #594790 (which now becomes a feature request) Mon Oct 13 2003 Lachlan Andrew * htfuzzy/Regex.cc: Search for regular expression. (Used to ignore it!) * htfuzzy/Speling.cc, htword/{WordList.cc,WordList.h,WordKey.cc,WordKey.h}: When looking in word database for misspelt words, don't ask to match trailing numeric fields in database key. * htcommon/defaults.cc, htdoc/htfuzzy.cc: Update docs. Sun Oct 12 2003 Lachlan Andrew * htsearch/htsearch.cc: Fix bug if fuzzy algorithms produced no search words. Send all debugging output to cerr not cout. More debugging output. Sun Oct 12 2003 Lachlan Andrew * htdig/{Retriever,Server}.cc: Back out the previous. Gilles pointed out inconsistency with Retriever::IsValidURL(). Sun Oct 5 2003 Lachlan Andrew * htdig/{Retriever,Server}.cc: Jim Cole's patch to bug #765726. Don't block paths with subpaths excluded by robots.txt. Sun Oct 5 2003 Lachlan Andrew * htsearch/htsearch.cc: Highlight phrases containing stop words * test/t_htsearch, test/conf/htdig.conf.in: Tests for the above Sat Sep 27 2003 Lachlan Andrew * test/{test_functions.in,t_htdig,t_htdig_local,t_htnet}: Don't assume shell "." command passes arguments. (Doesn't on FreeBSD.) Sat Sep 27 2003 Lachlan Andrew * htlib/HtDateTime.h, htnet/HtCookie.cc: Avoid ambiguous function call on systems (HP-UX) where time_t=int Fri Aug 29 09:35:46 MDT 2003 Neal Richter * removed references to CDB___mp_dirty_level ,CDB_set_mp_diry_level() & CDB_get_mp_diry_level() * The config verb 'wordlist_cache_dirty_level' was left for possible use in the future. Thu Aug 28 15:11:21 MDT 2003 Neal Richter * Changed db/LICENSE file to new LGPL compatible license from Sleepycat Software -- Thanks Sleepycat! * Reverted to Revision 1.2 or db/mp_alloc.c The recent changed cuased large DB growth. Strangely the files contained no 'new' data, they were just much larger. Looks like the pages were being flushed too often???? Thu Aug 28 12:41:22 EST 2003 Gabriele Bartolini * global: updated with 'autoreconf -if' (autoconf 2.57, libtool 1.5.0a and automake 1.7.6) * 'make check' successful on: AMD64 Linux 2.4, Alpha Linux 2.2, RedHat Linux 7.3 (2.4), SPARC Ultra60 Linux 2.4, Sparc R220 Sun Solaris (5.8). * README.developer: added further info Thu Aug 28 12:00:10 EST 2003 Gabriele Bartolini * db/[config.guess,config.sub,install-sh,ltmain.sh,missing]: added in the database directory (this way 'make dist' goes on); I have not been able to tell the db/configure script to get the 'top_srcdir' ones (which should be the default behaviour). Maybe in the future we'll look for this. Thu Aug 28 11:53:48 EST 2003 Gabriele Bartolini * db/configure.in: changed AC_PROG_INSTALL() to AC_PROG_INSTALL and removed AC_CONFIG_AUX_DIR; this implies that autotools copies will be made for the db directory as well. Thu Aug 28 11:36:42 EST 2003 Gabriele Bartolini * [htcommon,htdb,htdig,htfuzzy,htlib,htnet,htsearch,httools,htword,test]/Makefile.am: added the option above to every *_LDFLAGS Thu Aug 28 11:30:39 EST 2003 Gabriele Bartolini * Makefile.am: removed acconfig.h from the EXTRA_DIST list Thu Aug 28 11:25:07 EST 2003 Gabriele Bartolini * configure.in: removed portability checks for error, stat and lstat that caused a compile errors on Solaris. Added the '-mimpure-text' extra ld flag for GCC on solaris systems (a linkage error occurs when libstdc++ is not shared) Thu Aug 28 11:22:57 EST 2003 Gabriele Bartolini * include/Makefile.am: changed htconfig.h.in into config.h.in Thu Aug 28 11:16:19 EST 2003 Gabriele Bartolini * htlib/error.[h,c]: removed for now, until replacement functions will be correctly performed. Thu Aug 28 11:11:32 EST 2003 Gabriele Bartolini * htdoc/cf_generate.pl: fixed an error when opening tail and head files * Makefile.am: enabled rebuild from a different directory (it is used my 'make dist') Thu Aug 28 10:46:35 EST 2003 Gabriele Bartolini * htlib/malloc.c: modified according to autoconf specifications as far as replacement functions are regarded * htlib/[lstat, stat].c: removed for now Thu Aug 28 10:40:58 EST 2003 Gabriele Bartolini * htdoc/cf_generate.pl: accept an optional parameter (top source directory) * htcommon/defaults.cc: fixed some broken lines which prevented cf_generate.pl from correctly working * htdoc/Makefile.am: modified the automake file for passing the top source directory to cf_generate.pl * htdoc/attrs.html, htdoc/cf_byname.html, htdoc/cf_byprog.html: Regenerated using cf_generate.pl. Tue Aug 26 12:25:40 EST 2003 Gabriele Bartolini * configure.in: removed AC_FUNC_MKTIME because it may not work properly and added default replacement directory (htlib) for future uses * htlib/Makefile.am: back-step with re-inclusion of mktime.c in the list of files to be always compiled (caused linking errors for the __mktime_internal function) * global: updated with 'autoreconf -if' Sun Aug 24 12:44:29 EST 2003 Gabriele Bartolini * updated with 'autoreconf -if': autoconf 2.57, automake 1.7.6 and libtool 1.5.0a (autotools that come with Debian SID) Sun Aug 24 12:39:34 EST 2003 Gabriele Bartolini * configure.in: moved AC_PROG_LEX to AM_PROG_LEX * db/configure.in: enabled AM_MAINTAINER_MODE which prevented users without autotools to configure and compile the program (relatively to the db directory) * include/htconfig.h: previously excluded from the branch (severe error!) Mon Jul 21 20:54:47 CEST 2003 Gabriele Bartolini * htlib/(malloc|error|lstat|stat|realloc).c: added for cross-compiling reasons (as suggested by automake) * htlib/error.h: ditto * db/acconfig.h: removed as suggested by autotools' new versions * configure.in: removed AC_PROG_RANLIB (overriden by AC_PROG_LIBTOOL) * updated as of rerun 'autoreconf -if' Mon Jul 21 10:08:24 CEST 2003 Gabriele Bartolini * Patch provided by Marco Nenciarini has been completely applied; the patch adds support for detection of standard C++ library * all sources using : modified to use standard ISO C++ library, if present * db/configure scripts: modified for autoconf 2.57 Mon Jul 21 09:59:16 CEST 2003 Gabriele Bartolini * [.,*]/Makefile.in: regenerated by new automake against new configure.in * Makefile.config: now looking for the global configuration file in the source directory Mon Jul 21 09:49:22 CEST 2003 Gabriele Bartolini * configure.in: completely rewritten, deprecated directives have been removed and now version 2.57 is a prerequisite. * acinclude.m4: moved all the macros here * aclocal.m4, configure: regenerated by aclocal and autoconf * acconfig.h: removed as now it is deprecated * include/htconfig.h.in: removed, as 'config.h.in' is preferred and auto-generated * config.[guess,sub]: updated with newer versions Tue Jul 8 16:29:44 2003 Gilles Detillieux * htsearch/parser.cc (checkSyntax): Fixed boolean_syntax_errors handling to work over multiple config files. Mon Jul 7 00:41:55 CEST 2003 Gabriele Bartolini * Updated to autoconf 2.57, libtool 1.5 and automake 1.7.5 * removed acconfig.h files * autoconf include file is now include/config.h (for autoheader) * include/htconfig.h.in renamed in include/htconfig.h: now includes config.h and redefines the bool types * htlib/HtRegexList.cc, htdig/(Document.cc|ExternalParser.cc): removed TRUE and FALSE and converted to C++ standard values Sat Jul 5 2003 Lachlan Andrew * test/test_functions.in: Fix bugs starting/killing apache Sat Jul 5 2003 Lachlan Andrew * htcommon/defaults.cc: Disable cache flushing to avoid "page leak". Tue Jun 24 2003 Neal Richter * Update Copyright Notices in code & documentation to 2003 * Changed License Notice GPL -> LGPL License change (Decided by HtDig Board & Membership October 2002 Mon Jun 23 2003 Neal Richter * Raft of changes. Most todo with Native Win32 support * TODO: ExternalTranport & ExternalParser are effectively dissabled with #ifdefs for Native WIN32 * remove global CDB___mp_dirty_level variable and subsitute functions to set/get variable * Added local copies of GNU LGPL regex, POSIX-like dirent routines, getopt library and filecopy routines - mainly for Native WIN32 support * improve IsValidURL with return codes (htdig/Retriever.cc) * lots of improvements/new-features to libhtdig Sun Jun 22 2003 Lachlan Andrew * db/mp_cmpr.c (CDB___memp_cmpr_open): Make weak compression database standalone to avoid recursion This *should* fix all of the recent problems with dirty cache etc. * test/search.cc: Don't take sizeof zero sized array Fri Jun 20 2003 Lachlan Andrew * configure,aclocal.m4,acinclude.m4: --with-ssl set CPPFLAGS, not CFLAGS Fri Jun 20 2003 Lachlan Andrew * db/configure: Hack which should allow select to be detected on HP/UX * db/db.c: Replace HAVE_ZLIB with HAVE_LIBZ (as set by configure) * htword/wordKey.cc: More descriptive error message (Changes to compile with Sun's C++) * htnet/{HtCookie.cc,HtFTP.cc,Transport.cc}: Assign substring of const string to const pointer. * htsearch/ResultMatch.h: Allow use of SortType in ResultMatch::setSortType() * test/search.cc: Don't take sizeof(variable size array) * htdb/htdb_stat.cc: avoid name clash for global var internal * htcommon/URL.h, htlib/HtTime.h, htlib/htString.h, htnet/Connection.h, htword/WordBitCompress.h: Cast default args of type string literal to type (char*) * htdocs/require.html: Remove email address. * htlib/gregex.h: Avoid warning if __restrict_arr already defined Sun Jun 14 2003 Lachlan Andrew * htcommon/defaults.cc: Set wordlist_cache_dirty_level to 1 (it most conservative value). Miscellaneous reformatting. * htdoc/attrs.html, htdoc/cf_byname.html, htdoc/cf_byprog.html: Regenerated using cf_generate.pl. * htdoc/{require.html,meta.html,all.html,meta.html}: Update disk usage for phrase searching. Updated list of supported platforms. More hyperlinks. Fri Jun 13 2003 Lachlan Andrew * htsearch/Display.cc (setVariables), htdocs/hts_template.html: Set MATCH_MESSAGE from method_names (for internationalisability). Removed all trace of hack for config attribute... Thu Jun 12 14:16:05 2003 Gilles Detillieux * htsearch/htsearch.cc (main): Fixed boolean_keywords handling to work over multiple config files (must destroy old list before creating new one). * htcommon/defaults.cc, htsearch/Display.cc (setVariables): Removed incorrect default value for "config" attribute, and removed hack that attempted to correct it. * htdoc/attrs.html: Regenerated using cf_generate.pl. Thu Jun 12 13:28:01 2003 Gilles Detillieux * htcommon/defaults.cc, htcommon/HtSGMLCodec.cc (ctor): Added translate_latin1 option to allow disable Latin 1 specific SGML translations. * htdoc/attrs.html, htdoc/cf_byname.html, htdoc/cf_byprog.html: Regenerated using cf_generate.pl. Mon Jun 9 2003 Lachlan Andrew * htsearch/htsearch.cc: Fixed setupWords loop for junk at end of query Mon Jun 9 2003 Lachlan Andrew * htsearch/Display.cc: Set CONFIG template variable to the base name of the config file (no directory or .conf), as expected by htsearch Mon Jun 9 2003 Lachlan Andrew * test/test_functions.in: avoid trying killing apache multiple times * configure,configure.in: Reformat --help output * htdoc/FAQ.html: Brought up-to-date with main docs * htdoc/hts_templates.html: added hyperlinks. * installdir/search.html: Display version Sun Jun 8 2003 Lachlan Andrew * configure: Hack to set --disable-bigfile for Solaris (with Sun cc) and --disable-shared --enable-static for Mac OS X * test/{test_functions.in,t_htdig,t_htdig_local,t_htnet}: Only start Apache for tests which need it, and kill it after the test * contrib/parse_doc.pl: Allow file names containing spaces (from .deb) Mon Jun 2 2003 Lachlan Andrew * db/mp_cmpr.c: Add default zlib setting to default_cmpr_info * htcommon/defaults.cc, htword/WordDBCompress.cc: Fix docs to say default compression by 8 (not by 3, which I had "fixed" it to...) * htcommon/conf_lexer.{cxx,lxx}: Avoid warnings, and document hack. Thu May 29 2003 Lachlan Andrew * db/mp_cmpr.c: Fix comparison of -1 and unsigned which broke SunOS cc * htdoc/install.html: Warn SunOS cc users to --disable-bigfile * htcommon/conf_lexer.cxx: Suppress warnings of unused identifiers * test/con/htdig.conf2.in: Disable testing of content_classifier attribute, as didn't work until after installation Tue May 27 2003 Lachlan Andrew * db/configure, db/ac{local,include}.m4: Stop test for zlib from adding -I/default/path (*this* time...) * htword/DBPage.h: Fix bug introduce in previous patch * test/Makefile.{in,am}: Replace non-portable make -C X by cd X; make Tue May 27 2003 Lachlan Andrew * {,db/}configure, {,db/}ac{local,include}.m4: Stop test for zlib from adding -I/default/path (broke SunOS cc) Fix -Wall test if CCC is g++ but CC is not gcc * test/dbbench.cc: #include later, to avoid #define open causing problems * includedir/synonyms: Remove trailing blank line which caused warning * htnet/HtCookieInFileJar.cc,htfuzzy/Synonym.cc: .get() to stop warnings * htlib/mhash_md5.c: char -> unsigned char to stop warnings * test/search.cc, htword/WordDBPage.h: Casts to (int) to stop printf warnings. ALLIGN -> ALIGN Sat May 24 2003 Lachlan Andrew * htcommon/defaults.cc: Keep more wordlist cache pages clean * {,db/}configure{,.in}, {,db/}ac{local,include}.m4: Patch by Richard Munroe to test if -Wno-deprecated needed. Many bug fixes / extra search paths added. * include/htconfig.h.in, db/db_config.h.in: Only '#define const' if not C++ (htword/WordDB.cc uses db_config.h) * test/dbbench.cc: check for alloca even if gcc * test/t_url: used grep -C instead of grep -c (for portability) * db/mp_{alloc,cmpr}.c: Removed/replaced C++ style comments * htdoc/require.html: Revised list of supported platforms Thu May 22 2003 Lachlan Andrew * htnet/HtFile.cc: Fix previous .get() patch... Thu May 22 2003 Lachlan Andrew * htlib/DB_2.cc: Set wordlist_cache_dirty_level before opening database, to avoid database memory allocation problem. * db/db_err.c: Make 'fatal' errors actually exit. * htdig/Document.cc, htsearch/parser.cc, htdig/htdig.cc, * htnet/Ht{HTTP,File}.cc: Add .get() to use of strings to avoid compiler warnings (FreeBSD). Thu May 22 2003 Lachlan Andrew * ltmain.sh, test/Makefile.in: Hack to list library dependencies multiple times in g++ command, to get MacOS X to 'make check'. * test/{search,word}.cc: cast sizeof() to (int) to avoid warnings. * htdoc/install.html: Documented MacOS X's shared libraries problem. Sun May 18 2003 Lachlan Andrew * db/mp_alloc.c: Hopefully the *last* fix for this morning's patch... * configure, aclocal.m4, acinclude.m4: Look for httpd modules in .../libexec/httpd for OS X * test/conf/httpd.conf: Disabled mod_auth_db, mod_log{agent,referer}. Sun May 18 2003 Lachlan Andrew * db/db.h.in: Declare variable introduced in db/mp_cmpr.c patch Sun May 18 2003 Lachlan Andrew * db/mp.h, db/mp_{alloc,bh,cmpr,region}.c, * htword/WordDB.cc, htdig/htdig.cc: Avoid infinite loop if memp_alloc has only dirty, "weakly compressed" (i.e. overflow) pages. * htcommon/defaults.cc: Document the above, plus misc updates. * htword/WordDBPage.h: Cast sizeof() to (int) in printf()s to avoid compiler warnings. Sun APR 20 2003 Lachlan Andrew * htdig/htdig.cc: delete db.words.db_weakcmpr if -i specified. Wed Feb 26 22:10:40 CET 2003 Gabriele Bartolini * htnet/HtHTTP.cc: fixed colon (':') problem with HTTP header parsing, as Frank Passek, Gilles and others suggested, as space is not mandatory between the field declaration and the field value returned by the server Sun Feb 23 10:20:58 CET 2003 Gabriele Bartolini * htcommon/defaults.[cc,xml]: added the 'cookies_input_file' configuration attribute for pre-loading cookies in memory * htdig/htdig.cc: added the feature above; the code automatically loads the cookies from the input file into the 'jar' that will be used during the crawl. Sun Feb 23 10:16:08 CET 2003 Gabriele Bartolini * htnet/HtHTTP.h: removed the NULL pointer check before assigning a new jar to the HTTP code Tue Feb 11 2003 Lachlan Andrew * htcommon/defaults.cc: Set default compression_level to 6, which enables Neal's wordlist_compression_zlib flag. Tue Feb 11 2003 Lachlan Andrew * htcommon/{DocumentRef.h, HtWordReference.h}, htsearch/WeightWord.{cc,h}, htsearch/parser.{cc,h}, htsearch/htsearch.cc: Added field-restricted searching, by title:word or author:word * htdig/ExternalParser.cc, htdig/HTML.{cc,h}, htdig/Parsable.{cc,h}, htdig/Retriever.{cc,h}: Parse author from tags. Also moved some common functionality from HTML/ExternalParser into Parsable. * test/t_htsearch, htcommon/defaults.cc, htdoc/{TODO.html,hts_general.html,hts_method.html}: Test and document the above Sun Feb 9 2003 Lachlan Andrew * htdig/HTML.cc: fix bug in detection of deprecated noindex_start/end * htsearch/Display.cc: try harder to find value for DBL_MAX #680836 * htcommon/defaults.cc: fixed typos. Sat Feb 1 13:57:17 CET 2003 Gabriele Bartolini * htnet/HtCookie.[h,cc]: allowed printDebug to be passed an ostream object * htnet/HtCookieMemJar.cc: removed a debug call Thu Jan 30 19:28:32 CET 2003 Gabriele Bartolini * configure.in: used AC_LIBOBJ instead of deprecated LTLIBOBJS's workaround * ltconfig: removed as not needed anymore since libtool 1.4 * db/configure.in: added AC_CONFIG_AUX_DIR(../) for letting automake know to use the main ltmain.sh file * configure, aclocal.m4, Makefile.in, */Makefile.in, config.guess, config.sub, install-sh, ltmain.sh, missing, mkinstalldirs: re-generated by autotools: aclocal, autoconf 2.57, automake 1.6.3 and libtool 1.4.3 * db/aclocal.m4, db/configure, db/mkinstalldirs: ditto Thu Jan 30 00:16:51 CET 2003 Gabriele Bartolini * htsearch/htsearch.cc: removed a warning due to a not-initialized pointer Wed Jan 29 22:53:25 CET 2003 Gabriele Bartolini * acinclude.m4: included the function for checking against SSL, as found in the ac-archive. Tue Jan 28 12:23:16 CET 2003 Gabriele Bartolini * htnet/Makefile.am: added HtCookieInFileJar.[h,cc] files * installdir/cookies.txt: example file for pre-loading HTTP cookies * installdir/Makefile.am: added cookies.txt Tue Jan 28 12:16:28 CET 2003 Gabriele Bartolini * htnet/HtCookieMemJar.[h,cc]: performed deep copy of the jar in the copy constructor Tue Jan 28 12:13:44 CET 2003 Gabriele Bartolini * htnet/HtCookie.[h,cc]: added the constructor of a cookie object from a line of a cookie input file (Netscape's way): if an expiration value of '0' is set through the cookies input file, the cookie is managed as a session cookie. Improved copy constructor, solving a bug related to the expires field. Tue Jan 28 12:11:27 CET 2003 Gabriele Bartolini * htnet/HtCookieInFileJar.[h,cc]: class for importing cookies from a text file Tue Jan 28 12:08:20 CET 2003 Gabriele Bartolini * htlib/HtDateTime.h: added the constructor HtDateTime(const int) Sat Jan 25 2003 Lachlan Andrew * htsearch/Display.cc: Convert "
\n" in $(DESCRIPTION) to "
" so it can be used in Javascript (feature request #529926). Tue Jan 21 2003 Lachlan Andrew * HTML.cc (HTML, parse): Handle noindex_start/end as string lists. * test/{t_htsearch,htdocs/set1/script}: Test the above * htcomon/defaults.cc: Add "> (istream&,String&) ): Exit loop when getline fails for reasons other than a full buffer. * htnet/HtFile.cc (File2Mime), installdir/HtFileType: Allow file names containing spaces. Sat Jan 11 2003 Lachlan Andrew * htnet/HtFile.cc (Request), htdig/Document.cc (RetrieveLocal), htcommon/URL.h htcommon/URLTrans.cc: Decode URL paths before use as local filenames (file:/// & local_urls). * test/{t_htdig,t_htdig_local,t_htsearch}, test/conf/htdig.conf2.in, test/htdocs/set1/{index.html,site 1,sub%20dir/empty file.html}: Tests for the above. * htcommon/HtConfiguration.cc: brackets around assignment in 'if'. * test/search.cc (LocationCompare): Only specify default arg once. Fri Jan 10 2003 Lachlan Andrew * htlib/String.cc (operator>> (istream&,String&) ): Check status of stream, no return value of get(). Fixes bug (for some C++ libs) where reading stops at a blank line. Fri Jan 1 2003 Lachlan Andrew * htnet/HtFile.cc(Ext2Mime,Request), htdig/Document.cc(RetrieveLocal): Determine local files' MIME types from mime.types, not hard-coded. URLs matching attribute "bad_local_extensions" must use their true transport protocol (HTTP for http://, filesystem for file:///). * htnet/HtFile.cc (File2Mime, Request): For file:/// URLs only, files without (or with unrecognised) extensions are checked by the program specfied by the "content_classifier" attribute. * htnet/htFile.cc (Request): Symbolic links are treated as redirects, to avoid problems with relative references. * htcommon/defaults.cc: Documented the above (and added crossrefs). * test/t_ht{dig,dig_local,search}, test/htdocs/set1/*, test/conf/htdig.conf2.in: Add tests for bad_local_extensions. Mon Dec 31 2002 Lachlan Andrew * configure.in,htfuzzy/EndingsDB.cc,htlib/{HtR,r}egex.h,Makefile.am: Renamed regex.h to gregex.h and allow use of rx instead. * htcommon/defaults.cc,htdocs/{attrs,cf_byprog,cf_byname}.html: Fixed typo in cross-references to restrict and limit_urls_to. * test/t_htmerge: Re-enabled htmerge command (discarding output). * test/Makefile,test/conf/htdig.conf3.in: Added conf3 and fixed db path. Mon Dec 30 2002 Lachlan Andrew * contrib/doc2html/*: Incorporated David Adams' latest version, 3.0.1. Mon Dec 30 2002 Lachlan Andrew Forward-ported several patches from 3.1.6: * htdig/ExternalParser.cc: Added "description_meta_tag_names" attrib. Added "dc.date|dc.date.created|dc.date.modified" synonyms for "date". Allow spaces between "url" and "=" in refresh. Fixed bug in flag positions. Added "use_doc_date" attribute. * htdig/HTML.cc: Added "description \_meta_tag_names" attribute. Added "dc.date|..." synonyms. Added "ignore_alt_text" attribute. * htdig/Retriever.cc: Added "ignore_dead_servers" attribute. Added call to "url.rewrite() in got_href(). * htdig/FAQ.html: Latest version now 3.1.6. Mention old security hole. Describe external converters for PostScript etc. Mention pdf_parser not supported in 3.2. * htdoc/{attrs,cf_byname,cf_byprog}.html: New attributes added (automatically from defaults.cc). * htdoc/htmerge.html: Update for multiple database support. * htdoc/hts_form.html: Describe relative/incomplete dates. * htdoc/require.html: Describe phrase searching, external parsers, external transports. Added some new supported systems. (Commented out as testing incomplete.) * htfuzzy/Synonym.cc: Protect against "synonym" entries with one word. * htlib/String.cc: Protect against negative string lengths. * htsearch/Display.{cc,h}: Added "search_result_contenttype" attribute, and corresponding displayHTTPheaders() function. Rewrite URLs. Remove old "ANCHOR" variable. Handle relative dates. Added "max_excerpts" attribute and buildExcerpts() function. Added "anchor_target" attribute. * htsearch/DocMatch.h: Added "orMatches" * htsearch/htsearch.cc: Added "boolean_keywords" attribute. Rewrite URLs. * htsearch/parser.cc: Added "boolean_syntax_errors" attribute. Added wildcard search. Fixed bug in perform_phrase() so it now handles "bad words" and short words properly. Added "multimatch_factor" to give greater weight to documents matching multiple "OR" terms. * htsearch/htparser.h: Added boolean_keywords support. * htcommon/defaults.{cc,xml}: New attributes added, and enhanced descriptions Cleaned code to remove some compiler warnings/errors: * htcommon/HtConfiguration.cc: Brackets around assignment 'path=' inside 'if' * htdig/Server.cc, htsearch/Display.cc: Added ".get()" when strings passed as arguments. * htlib/StringMatch.h, htword/WordBitCompress.h: Explicit cast of NULL to (char*)NULL for broken C++ compilers. Also: * STATUS: Removed "not all htsearch input parameters handled properly", "Return all URLs", "Turn on URL parser test", "htsearch phrase support tests". Reduced list of things to do for "require.html". * test/t_htsearch, test/conf/htdig.conf3.in: Added testing of phrases and boolean_keywords / boolean_syntax_errors. Thu Nov 28 09:02:46 2002 Gilles Detillieux * installdir/english.0: Removed S flag from birth, because it doesn't do what we want (birthes, not births). Tue Nov 26 23:16:08 2002 Gilles Detillieux * htdoc/hts_form.html: Fixed typo in link & description for restrict. Tue Nov 26 22:30:06 2002 Gilles Detillieux * installdir/english.0: Patched with Lachlan Andrew's changes, fixing lots of dubious uses of suffixes to get more appropriate and correct fuzzy endings expansions. * installdir/synonyms: Updated with the version contributed by David Adams, with minor changes. Kept old one as synonyms.original. Mon Nov 4 10:44:35 CET 2002 Gabriele Bartolini * htcommon/URL.[h,cc]: added the assignment operator Sun Oct 27 09:29:02 2002 Geoffrey Hutchison Merge in word DB zlib patch from Neal Richter. * db/db.h.in, db/mp_cmpr.c, htword/WordList.cc, htword/WordDBCompress.h, htword/WordDBCompress.cc: Add support for using the zlib compression (and compression level) if specified by the new wordlist_compress_zlib, which is "true" by default. * htcommon/defaults.cc: Add attribute wordlist_compress_zlib as above. * htdoc/attrs.html, htdoc/cf_byname.html, htdoc/cf_byprog.html: Update using cf_generate.pl. Sat Oct 26 21:59:01 2002 Geoffrey Hutchison Merge in fixes from Lachlan Andrew * test/Makefile.am, test/Makefile.in, test/t_url, test/url.cc, test/url.children, test/url.parents, test/url.output: Add URL tests to the automatic test suite (rather than requiring them to be run manually). * */Makefile.in: Regenerate using automake-1.4p6. * htcommon/URL.cc, htcommon/URL.h: Add new configuration attribute allow_double_slash to only remove // marks when requested (since some server-side code uses them), handle initial protocols without double slashes, and only remove the default doc string from appropriate protocol URLs (e.g. not file), treat ".//" as a relative path, and collapse /../ *after* // and /./ handling. * htcommon/defaults.cc: Add documentation for allow_double_slash, as well as various documentation cleanups. * htdig/ExternalTransport.cc: Fix minor bug--recognize service specified as https:// rather than https. * htdoc/hts_form.html, htdoc/hts_templates.html: Documentation fixes. * htsearch/htsearch.cc: Create valid boolean query if "exact" not specified in search_algorithms by adding the exact word with low weight. Solves PR#405294. Fri Oct 4 17:05:06 2002 Geoff Hutchison * htcommon/defaults.xml: Added first-draft XML version of defaults file. This will eventually be used to generate defaults.cc and documentation automatically. (As pointed out by Brian White, this will make the binaries smaller.) Wed Sep 25 13:56:31 2002 Gilles Detillieux * htdig/HTML.cc (parse): Fixed handling of JavaScript skipping so it doesn't get confused by "<" in code. Thu Sep 19 09:04:50 CEST 2002 Gabriele Bartolini * htnet/HtHTTP.cc : another check for cookie jar's null pointer Tue Sep 17 17:41:51 2002 Gilles Detillieux * htcommon/defaults.cc (external_protocols): Fixed table formatting as suggested by Lachlan Andrew. Thu Aug 29 21:21:34 CEST 2002 Soeren Vejrup Carlsen * htdig/Document.[h,cc]: first steps in FTP handling. HtFTP.h included and we now test for the 'ftp' protocol in the Document::Retrieve function. Has not yet been tested! * htnet/HtFTP.[h,cc]: added class to handle the FTP-protocol. Very experimental (has not been tested yet). Fri Aug 9 13:01:05 2002 Gilles Detillieux * httools/htnotify.cc (readPreAndPostamble): Check for empty strings in file names, not just NULL, as suggested by Martin Kraemer. Wed Aug 7 12:11:31 2002 Gilles Detillieux * htdig/ExternalParser.cc (parse): Fixed to impose max_doc_size restriction on external converter output which it reads in. Tue Aug 6 18:21:11 CEST 2002 Gabriele Bartolini * these changes were suggested by David Reed (thanks) * htdig/Document.cc: manage cookies via SSL * htnet/HtCookie.[h,cc]: features both RFC2109 and Netscape version * htnet/HtCookieJar.cc: ditto Tue Aug 6 17:12:22 CEST 2002 Gabriele Bartolini * htcommon/defaults.cc: added the 'http_proxy_authorization' attribute. Needs revision due to my usual *spaghetti* english. :-) * htdig/Document.[h,cc]: proxy authorization is now enabled Tue Aug 6 09:28:39 CEST 2002 Gabriele Bartolini * htnet/Connection.[h,cc]: IP address storing as string (sync with ht://Check) * htnet/Transport.[h,cc]: HTTP Proxy and Basic credentials handling moved here (ditto) through the use of a protected static method * htnet/HtHTTP.h: SetCredentials declared to be virtual (unnecessary because inherited, but gives better understanding); new method SetProxyCredentials for proxy authorization. * htnet/HtHTTP.cc: HTTP header Proxy-Authorization is now handled. The SetCredentials and SetProxyCredentials methods now make use of the Transport::SetHTTPBasicAccessAuthorizationString method, in order to write the string for negotiating the access. Fri Aug 2 15:40:18 2002 Gilles Detillieux * htdig/Document.cc (Retrieve): Allow redirects from HTTPSConnect. Tue Jul 30 12:46:56 2002 Gilles Detillieux * htlib/md5.cc: Added missing include of stdlib.h, as Geoff suggested. Sat Jul 27 11:57:25 2002 Geoff Hutchison * htnet/SSLConnection.cc: Add fix for segfault on SSL connections noticed by several users. Fix contributed by Andy Bach . Tue Jun 18 10:22:01 2002 Geoff Hutchison * htdig/Retriever.cc (got_word): Check that the word length meets the minimum word length before doing any processing. Fri Jun 14 17:26:21 2002 Gilles Detillieux * htsearch/Display.cc (buildMatchList), htsearch/HtURLSeedScore.cc (Match), htsearch/SplitMatches.cc (Match): Added Jim Cole's fix to bugs in handling of search_results_order. Wed May 15 09:45:40 CEST 2002 Gabriele Bartolini * htnet/Retriever.cc: fixed the bug regarding the server_wait_time feature after the maximum number of requests per connection has been reached. Tue Apr 9 16:41:33 CEST 2002 Gabriele Bartolini * htnet/HtCookie*.[h,cc]: RFC2109 compliant. * htlib/HtDateTime.[h,cc]: Add const-ness to the DiffTime static method Tue Apr 9 12:52:30 CEST 2002 Gabriele Bartolini * htnet/HtCookie.cc: fixed a bug regarding expiry date recognition Fri Apr 5 14:08:39 2002 Gilles Detillieux * htdig/ExternalTransport.cc (Request): Fixed to strip CR from header lines, output header lines with -vvv. Tue Mar 19 08:40:54 CET 2002 Gabriele Bartolini * htnet/HtCookie.cc: enhanced controls regarding the expires setting when no expires is returned. Prevents NULL pointer exceptions to be arisen. Mon Mar 18 11:28:02 CET 2002 Gabriele Bartolini * htlib/HtDateTime.h: added the copy constructor * htnet/HtCookie.cc: fixed a NULL pointer bug regarding 'datestring' management and HtDateTime copy constructor is now used Tue Mar 12 18:19:49 2002 Gilles Detillieux * htlib/HtDateTime.cc (Parse, SetFTime): Added Parse method for more flexible parsing of LOOSE/SHORT formats, use it in SetFTime. Also skip unexpected leading spaces in SetFTime, as these frequently cause problems with some strptime() implementations. Mon Feb 11 23:28:37 2002 Geoff Hutchison * htdig/Retriever.h (got_redirect): Add referer to properly handle broken links through a redirect as reported by Joe Jah. * htdig/Retriever.cc: As above. * htdig/Document.cc (Retrieve): Fix bug that prevented external transport methods from reporting redirects as reported by Jamie Anstice . * htlib/Dictionary.cc (hashCode): Trial of hash function suggested by Jamie Anstice. Sat Feb 9 18:06:29 2002 Geoff Hutchison * htsearch/DocMatch.[h,cc]: Add scoring code for the new htsearch framework. Thu Feb 7 11:32:14 2002 Gabriele Bartolini * htnet/HtHTTP.cc (ReadChunkedBody): gets control of Read_Line methods (return error when they fail). Fri Feb 1 17:12:31 2002 Geoff Hutchison * Merged htdig-3-2-x branch back into CVS mainline. * ChangeLog.0: Update with current 3.1.6 ChangeLog. Thu Jan 24 18:06:04 2002 Geoff Hutchison * configure.in, aclocal.m4: Use new CHECK_SSL macro from the autoconf archive. * configure: Generate via autoconf. Fri Jan 18 11:15:29 2002 Geoff Hutchison * htnet/Transport.h (class Transport): Add const to SetCredentials method declaration as pointed out by Roman Maeder. Wed Jan 16 13:35:26 2002 Geoff Hutchison * db/db.h.in: Add #include which seems to help problems of stat64 conflicts on Solaris as suggested by Gilles. Sat Jan 12 16:19:55 2002 Gilles Detillieux * htcommon/defaults.cc: A few changes to the wording and formatting of the 'accept_language' attribute description. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Fri Jan 11 21:18:00 CET 2002 Gabriele Bartolini * htcommon/defaults.cc: added the 'accept_language' attribute Fri Jan 11 20:53:36 CET 2002 Gabriele Bartolini * htnet/HtHTTP.[h,cc]: management of the accept-language directive added * htcommon/URL.[h,cc]: const-ness in copy constructor and other cosmetic changes * htlib/Server.[h,cc]: management of the 'accept_language' attribute as a server block configuration directive. * htlib/Document.cc: set of the attribute above for the HTTP layer Fri Jan 11 13:25:49 2002 Gilles Detillieux * htdig/ExternalTransport.cc (Request): Fixed to allocate access_time object before setting it. Fri Jan 4 12:31:34 2002 Gilles Detillieux * htnet/HtCookie.cc, htword/WordKeyInfo.cc, htword/WordMonitor.cc, test/search.cc: changed all uses of strcasecmp to mystrcasecmp for consistency and portability. Fri Jan 4 12:17:10 2002 Gilles Detillieux * htnet/HtHTTP.cc (HTTPRequest): make the second comparison of the transfer-encoding header the same as the first, i.e. case insensitive and limited to 7 characters. Fri Jan 4 15:13:13 CET 2002 Gabriele Bartolini * htnet/HtHTTP.cc: parse the transfer-encoding header as case insens. [fix htdig-Bugs-499388 by Matthias Emmert ] Sun Dec 30 15:47:35 CET 2001 Gabriele Bartolini * HtHTTP.[h,cc]: management of the Content-Language directive for the response Sat Dec 29 13:07:08 CET 2001 Gabriele Bartolini * htnet/HtCookie.[h,cc]: new fields (srcURL and isDomainValid) and a more robust class with initialization list and copy constructor * htnet/HtCookieJar.[h,cc]: method for calculating the minimum number of periods that a domain specification of a cookie must have. Depending on what the Netscape cookies specification says. * htnet/HtCookieMemJar.cc: Management of the domain field of the cookie Mon Dec 17 06:45:02 CET 2001 Gabriele Bartolini * htdig/htdig.cc: fixed bug about cookie jar creation. It is done in here, because there is only one jar for the whole process. However it can be moved anywhere else. :-) Mon Dec 17 06:40:25 CET 2001 Gabriele Bartolini * htnet/HtHTTP.cc: check for null pointer of cookie jar Sun Dec 16 19:55:07 CET 2001 Gabriele Bartolini * htnet/Connection.[h,cc]: default constructor is changed and accepts a socket value (by default is -1) * htnet/HtCookieJar.[h,cc]: added a simple iterator * htnet/HtCookieMemJar.[h,cc]: ditto * htnet/HtFile: removed the management of modification_time (constructor) * htnet/HtHTTP.[h,cc]: constructor with initilization list and without a default constructor (the construction is now forced to pass a valid connection object). Removed any memory deletion from the destructor. The class is now abstract (see the virtual pure destructor). * htnet/HtHTTPBasic.cc: creates a Connection object in the initialization and the destructor has no responsability * htnet/HtHTTPSecure.cc: creates an SSLConnection object in the initialization and the destructor has no responsability * htnet/HtNNTP.cc: creates a Connection object in the initialization and the destructor has no responsability * htnet/Transport.[h,cc]: default constructor accepts a pointer to a Connection object and the destructor carries out the deletion of it Thu Dec 6 13:24:30 2001 Gilles Detillieux * contrib/examples/rundig.sh: Fixed to make use of DBDIR variable, and to test for and copy db.words.db.work_weakcmpr if it's there. Fri Oct 19 11:07:33 2001 Gilles Detillieux * htdig/Retriever.cc (IsValidURL): Fixed discrepancies in debug levels for messages giving cause of rejection, inadvertantly changed when regex support added. Wed Oct 17 15:48:23 2001 Gilles Detillieux * htdig/ExternalTransport.h: Added missing class keyword on friend declaration. Tue Oct 16 14:35:16 2001 Gilles Detillieux * htcommon/default.cc (external_parsers): Documented external converter chaining to same content-type, e.g. text/html->text/html-internal. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Mon Oct 15 22:25:55 2001 Geoff Hutchison * htdig/Document.cc, htdig/htdig.cc, htdig/Retriever.cc: Make sure setEscaped is called with the current value of case_sensitive. Fixes bug pointed out by Phil Glatz. Fri Oct 12 17:14:08 2001 Gilles Detillieux * htdoc/htdump.html, htdoc/htload.html: Fixed 3 little typos. Fri Oct 12 15:11:45 2001 Gilles Detillieux * htnet/HtHTTP.cc (ParseHeader): Show header lines in debugging output at verbosity level 3, not 4, for consistency with 3.1.x. * htcommon/URL.cc (removeIndex): Fixed to make sure the matched file name is at the end of the URL. Fri Oct 12 10:39:54 2001 Gilles Detillieux * htlib/HtRegexList.cc (setEscaped): Fixed to set compiled flag to FALSE when there's no pattern, so match() can detect this condition. Fixes handling of empty lists in bad_querystr, exclude_urls, etc. * htdig/Retriever.cc (IsValidURL): Fixed bad_querystr matching to look at right part of URL, not whole URL. Mon Sep 24 11:47:15 2001 Gilles Detillieux * htnet/HtHTTP.cc (SetRequestCommand): Put If-Modified-Since header out in GMT, not local time, and only put it out if existing document time > 0. * htsearch/parser.cc (perform_phrase): Optimized phrase search handling to use linear algorithm with Dictionary lookups instead of n**2 alg., as suggested by Toivo Pedaste. Tue Sep 18 10:50:40 2001 Gilles Detillieux * htdoc/running.html: New documentation on how to run after configuring. * htdoc/rundig.html: New manual page for rundig script. * htdoc/install.html: Added link to running.html. * htdoc/contents.html: Added link to running.html, rundig.html, related projects. Updated links to contrib and developer site. Fri Sep 14 22:12:56 2001 Gilles Detillieux * htcommon/URL.h: Moved DefaultPort() from private to public for use in HtHTTP.cc. Fri Sep 14 09:25:20 2001 Gilles Detillieux * htnet/HtHTTP.cc (SetRequestCommand): Add port to Host: header when port is not default, as per RFC2616(14.23). Fixes bug #459969. Sat Sep 8 22:15:33 2001 Geoff Hutchison * acconfig.h, include/htconfig.h.in: Add undef for ALLOW_INSECURE_CGI_CONFIG, which if defined does about what you'd expect. (This is for any wrapper authors who don't want to rewrite but are willing to run insecure.) * htsearch/htsearch.cc: Only allow the -c flag to work when REQUEST_METHOD is undefined. Fixes PR#458013. Tue Sep 4 18:58:31 2001 Geoff Hutchison * htsearch/DocMatch.cc: Add scoring for Quim's new parser framework. Only the normal word scoring is currently done, not backlink_factor or other "Document" methods. Fri Aug 31 15:34:28 2001 Gilles Detillieux * htdig/HTML.h, htdig/HTML.cc (ctor, parse, do_tag): Fixed buggy handling of nested tags that independently turn off indexing, so doesn't cancel tag. Add handling of tag. Added <> delim. to tag debugging output. Fixed a few typos. Wed Aug 29 10:33:01 2001 Gilles Detillieux * htcommon/defaults.cc (url_part_aliases): Added clarification explaining how to use example. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Mon Aug 27 15:05:09 2001 Gilles Detillieux * installdir/search.html: Add DTD tag for HTML 4 compliance. * installdir/htdig.conf: Added .css to bad_extensions default, added missing closing ">". * htdoc/config.html: Updated with sample of latest htdig.conf and installdir/*.html. Wed Jul 25 22:16:06 2001 Gilles Detillieux * htcommon/defaults.cc: Put new htnotify_* entries in alphabetical order. Removed superfluous quotes from htnotify_webmaster example (htnotify.cc adds in the quotes). * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Tue Jul 24 16:07:01 2001 Gilles Detillieux * htcommon/defaults.cc: Changed references in (no_)page_number_text entries from maximum_pages to maximum_page_buttons. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Tue Jul 24 14:38:22 2001 Gilles Detillieux * htdoc/hts_templates.html: Document Quim Sanmarti's URL decoding feature for template variables. Thu Jul 12 14:12:02 2001 Gilles Detillieux * htnet/HtFile.cc (Request): Fixed so it doesn't remove newlines from documents, and so it only tries to open mime.types once even if the open fails. Thu Jul 12 11:40:07 2001 Gilles Detillieux * contrib/conv_doc.pl, contrib/parse_doc.pl: Fixed EOF handling in dehyphenation, fixed to handle %xx codes in title made from URL. * contrib/doc2html/doc2html.pl, contrib/doc2html/pdf2html.pl, contrib/doc2html/swf2html.pl: Fixed to handle %xx codes in URL title. Wed Jul 11 15:05:47 2001 Gilles Detillieux * htsearch/Display.cc (readFile): Added missing fclose() call, and debugging message for when file can't be opened. Wed Jul 11 14:26:28 2001 Gilles Detillieux * htsearch/Display.cc (displayParsedFile): Added debugging message when file can't be opened. * htseach/Display.cc (buildMatchList): Fixed while loop to avoid warning. * htsearch/htsearch.cc (main): Fixed handling of syntax error message to use String class instead of strdup(). * htsearch/parser.cc (setError): Added debugging message when error is set. * htsearch/parser.cc (parse): Fixed not to clear error message after it's set. Sat Jul 7 22:19:18 2001 Geoff Hutchison * */Makefile.in: Update using current production automake (1.4-p4). * htfuzzy/Regexp.[cc,h]: Change class name to Regexp to prevent further namespace clashes. * htfuzzy/Fuzzy.c: #include "Regexp.h" now and make sure we create the right class when needed. * htlib/mktime.c: Change included mktime declaration to mymktime to avoid conflict on Mac OS X. (For some reason, autoconf's AC_FUNC_MKTIME doesn't work for Mac OS X. So this is a hack in the meantime.) * htfuzzy/Makefile.am: Rename Regex files. Oops! Fri Jul 6 18:38:58 2001 Geoff Hutchison * htfuzzy/Regexp.cc, htfuzzy/Regexp.h: Rename Regex class to prevent problems on case-insensitive systems. * htlib/HtRegexReplaceList.cc, htlib/String.cc, htdig/htdig.cc: Change #include of to modern standard of iostream.h. * htlib/Configuration.cc (Read): Make sure we never reference a negative position when trimming off whitespace. * config.guess, config.sub: Update with new versions from GNU to recognize various flavors of Mac OS X/Rhapsody. * htlib/strptime.cc: Make sure len is initialized. Fri Jul 6 12:04:52 2001 Gilles Detillieux * htlib/HtRegexList.cc (setEscaped): Fixed a potential problem with list building. When we go back a step, we still have to compile the new pattern in case it's the last one. Wed Jul 4 23:39:19 2001 Gilles Detillieux * htcommon/URL.cc (parse, ServerAlias): Fixed two problems that caused incorrect signatures to be generated. Wed Jul 4 13:52:54 2001 Gilles Detillieux * test/document.cc (dodoc), test/url.cc (dourl), test/testnet.cc (Retrieve): Fixed up handling of config to match David Graff's changes of May 16, and handling of HtHTTPBasic class to match Joshua Gerth's changes of Mar 17. Tue Jul 3 16:20:56 2001 Gilles Detillieux * htdig/Retriever.cc (GetLocal): Fixed to use URL class on given URL, so that default port numbers are stripped off. This was needed to allow local fetching of robots.txt. * htnet/Connection.cc (ctors, dtor, Assign_Server, Get_Peername), htnet/Connection.h: Got rid of strdup stuff, used String class for peer & server_name. * htnet/Connection.cc (Get_PeerIP): Used unambiguous name for structure. * htnet/HtHTTP.cc (ctor, dtor): Don't allocate a 2nd Connection, as child classes already do this, and set pointer to null when connection is deleted, so we don't try to delete it twice. This was messing up the heap and causing segfaults. Call Transport::CloseConnection before deleting connection. * htnet/HtHTTPBasic.cc (dtor), htnet/HtHTTPSecure.cc (dtor), * htnet/HtNNTP.cc (dtor): Only delete connection if non-null, & set to null after deleting. Call Transport::CloseConnection before deleting connection. * htnet/Transport.cc (CloseConnection): Don't exit if connection pointer is null, as this may be normal when called from destructor. Fri Jun 29 11:14:36 2001 Gilles Detillieux * htfuzzy/Endings.cc (getWords): Undid change introduced in 3.1.3, in part. It now gets permutations of word whether or not it has a root, but it also gets permutations of one or more roots that the word has, based on a suggestion by Alexander Lebedev. * htfuzzy/EndingsDB.cc (createRoot): Fixed to handle words that have more than one root. * installdir/english.0: Removed P flag from wit, like and high, so they're not treated as roots of witness, likeness and highness, which are already in the dictionary. Mon Jun 25 12:50:47 2001 Gilles Detillieux * htsearch/htsearch.cc (main): Got rid of last remnants of 'urllist' and used the 'l' StringList as was used in the code before, to make restrict and exclude handling work properly. Mon Jun 25 15:52:19 CEST 2001 Gabriele Bartolini * htsearch/htsearch.cc: defined 'urllist' in order to remove the compilation error (as Jesse suggested). Fri Jun 22 16:28:13 2001 Gilles Detillieux * htsearch/Display.cc (buildMatchList): Fix date_factor calculation to avoid 32-bit int overflow after multiplication by 1000, and avoid repetitive time(0) call, as contributed by Marc Pohl. Also move the localtime() call up before gmtime() call, to avoid clobbering gmtime's returned static structure (my thinko). * htdig/htdig.cc (main): Use .work file for md5_db, if -a given, as contributed by Marc Pohl. * htcommon/URL.cc (constructURL): Ensure that the _host is set if we are constructing non-file urls, as contributed by Marc Pohl. * htdoc/THANKS.html: Credit Marc Pohl for patches. Tue Jun 19 17:14:05 2001 Gilles Detillieux * README: Bump up to 3.2.0b4, fix note about bug report submissions. Tue Jun 19 17:01:16 2001 Gilles Detillieux * htsearch/Display.cc (setVariables): Fixed handling of build_select_lists attribute, to deal with new restrict & exclude attributes. Mon Jun 18 12:16:27 2001 Gilles Detillieux * configure.in, configure: Fix "hdig" typo in help. Fri Jun 15 17:57:19 2001 Gilles Detillieux * htcommon/defaults.cc: Noted effect of locale setting on floating point numbers in search_algorithm and locale descriptions. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Fri Jun 15 15:36:51 2001 Gilles Detillieux * htdoc/cf_generate.pl: Fixed to handle new defaults.cc format with trailing backslashes. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Fri Jun 15 14:57:21 2001 Gilles Detillieux * htdb/htdb_dump.cc, htdb/htdb_load.cc, htdb/htdb_stat.cc: Added a conditional include of if HAVE_GETOPT_H is defined. Fri Jun 15 11:25:24 2001 Gilles Detillieux * htsearch/htsearch.cc (main), htcommon/defaults.cc, htdoc/hts_form.html: two new attributes, used by htsearch, have been added: restrict and exclude. They can now give more control to template customisation through configuration files, allowing to restrict or exclude URLs from search without passing any CGI variables (although this specification overrides the configuration one). Fri Jun 15 09:34:23 2001 Gilles Detillieux * htsearch/htsearch.cc (main): Changed ridiculously outdated question "Did you run htmerge?" to "Did you run htdig?". Fri Jun 8 11:07:04 2001 Geoff Hutchison * htsearch/Display.cc: Add header, now needed for RH 7.1. Thu Jun 7 12:05:09 2001 Gilles Detillieux * contrib/htdig-3.2.0.spec: Updated to 3.2.0b4. * contrib/README: Mention acroconv.pl script. Thu Jun 7 10:46:19 2001 Gilles Detillieux * htsearch/Display.cc (expandVariables): Use isalnum() instead of isalpha() to allow digits in variable names, allow '-' in variable names too for consistency with attribute name handling. Wed Jun 6 16:14:06 2001 Gilles Detillieux * httools/htpurge.cc (main): Added missing "u:" declaration in getopt() call. Wed Jun 6 15:24:04 2001 Gilles Detillieux * contrib/doc2html/DETAILS, contrib/doc2html/README, contrib/doc2html/doc2html.pl, contrib/doc2html/pdf2html.pl, contrib/doc2html/swf2html.pl: Update to version 3.0 of doc2html, contributed by David Adams . Wed May 16 11:23:04 2001 Geoff Hutchison Added a pile of changes contributed by David Graff fixing compilation problems with non-gcc/g++ compilers (i.e. Sun's compiler). * Makefile.config, db/Makefile.am: Added no-dependencies to AUTOMAKE_OPTIONS for those not on GNU C/C++ * configure.in: Changed AM_PROG_YACC to AC_PROG_YACC as autoconf and autoreconf both complain that AM_PROG_YACC is not in the library. * htcommon/DocumentDB.cc: Removed default parameters as they are already declared in the header * htcommon/HtConfiguration.cc: Changed some of the loop declarations so that Sparc C 4.2 is happy. Removed default parameters as they are already declared in the header Moved inline ParseString to header where it belongs. Added initialization for HtConfiguration::_config static member variable. Added implementation of HtConfiguration::config() static class member. * htcommon/HtConfiguration.h: Added include for ParsedString.h. Added declaration of static member function ::config(). Added private static member variable _config;. Added inline ParseString from implementation. * htcommon/HtURLCodec.cc, htcommon/HtURLRewriter.cc, htcommon/HtZlibCodec.cc, htcommon/URL.cc, htcommon/conf_lexer.lxx, htdig/Document.cc, htdig/ExternalParser.cc, htdig/ExternalTransport.cc, htdig/HTML.cc, htdig/Parsable.cc, htdig/Plaintext.cc, htdig/Retriever.cc, : Changed to use new global configuration semantics. * htcommon/conf_parser.yxx: Added a return to yyerror to quiet Sparc C 4.2. Should really return a value here. Is it normal to return a YY_something or just -1, 0, ? * htcommon/defaults.cc: Added line continuation characters at the end of all the string lines that did not completed by a quote. * htcommon/defaults.h, htdig/htdig.h: Removed extern HtConfiguation config in favor of HtConfiguration::config(). * htdig/ExternalTransport.h Changed return type of GetResponse to match superclass. * htdig/Server.cc, htdig/htdig.cc, htfuzzy/htfuzzy.cc, htnet/HtFile.cc, htsearch/Display.cc, htsearch/QueryLexer.cc, htsearch/WordSearcher.cc, htsearch/htsearch.cc, htsearch/parser.cc, htsearch/qtest.cc, httools/htdump.cc, httools/htload.cc, httools/htmerge.cc, httools/htnotify.cc, httools/htpurge.cc, httools/htstat.cc htlib/Configuration.cc, htlib/HtRegex.cc: Changed constructor to use initializers * htlib/HtDateTime.cc: Moved inlines to header * htlib/HtDateTime.h: Added inlines from implementation * htlib/HtHeap.cc, htlib/HtHeap.h, htlib/HtVector.cc, htlib/HtVector.h, htlib/HtVectorGeneric.h, htlib/HtVectorGenericCode.h: Changed Copy member to return same type as superclass * htlib/HtRegexReplace.cc, htlib/HtRegexReplaceList.cc: Removed default parameters as they are declared already in the header * htlib/myqsort.h: Changed comment in header to use C-style comments as it's compiled using a C. * htlib/regex.h: Changed #if __STDC__ to #if defined(__STDC__) * htword/WordKey.h: Corrected const'ness Wed May 9 07:50:19 CEST 2001 Gabriele Bartolini * htnet/HtCookieJar.h: ShowSummary makes the class abstract Sat May 5 20:51:00 2001 Geoff Hutchison * htdoc/cf_blocks.html: Add colon in example and description of blocks to match code for the moment. The parser can be changed later if we like. Sat May 5 20:38:44 2001 Geoff Hutchison * htlib/ParsedString.cc (get): Use isalnum() instead of isalpha() for looking up--allows names that contain digits too. Sat May 5 20:36:29 2001 Geoff Hutchison * htlib/htString.h (class String): Remove now-obsolete and confusing int() casting operator. This was previously used to make a string of a certain length. Use String(int) as a ctor instead. Sat May 5 20:30:18 2001 Geoff Hutchison * htword/WordContext.[h,cc]: Change Initialize to supply a config that can be modified (i.e. if we don't have ZLIB_H). Sat May 5 23:30:55 CEST 2001 Gabriele Bartolini * htnet/HtCookieJar.h: ShowSummary, printing cookies (to be derived) * htnet/HtCookieMemJar.[h,cc]: ShowSummary, printing cookies Thu May 3 23:14:14 CEST 2001 Gabriele Bartolini * htnet/HtHTTP[h,cc]: connection object is now created and destroyed. NULL pointers converted to C++ standard (0). * htnet/Transport[h,cc]: NULL pointers converted to C++ standard (0). * htnet/Connection[h,cc]: ditto Thu May 3 23:09:33 CEST 2001 Gabriele Bartolini * htlib/HtDateTime.[h,cc]: Timestamp format added (used by ht://Check for MySQL interfacing) - keeping them equal helps me maintaining both of them! Thu May 3 10:28:56 2001 Gilles Detillieux * htsearch/parser.cc (perform_and): Add missing return statement, as suggested by Quim Sanmarti. Fri Mar 30 15:50:42 2001 Gilles Detillieux * htsearch/ResultMatch.h, htsearch/ResultMatch.cc (setTitle): Changed argument type to char * to fix problem with sort by title not working, as reported by Adam Lewenberg. Fri Mar 30 14:08:51 2001 Gilles Detillieux * htdig/Document.h, htdig/Retriever.cc (parse_url): Define and use Document::StoredLength() method to get actual length of data retrieved and given to md5(), which may be less than original length. Fixes bug reported by Michael Haggerty. Wed Mar 21 22:22:55 2001 Geoff Hutchison * htsearch/Display.cc (generateStars): Add NSTARS variable for template output as suggested by Caleb Crome (except here precision is 0). Fixes feature request #405787. * htdoc/hts_templates.html: Add description of NSTARS variable above. * htlib/HtRegex.cc (set): Make sure we free memory if we've already compiled a pattern. * htdig/Retriever.cc (got_href): Fix bug pointed out by Gilles with hopcounts and don't bother to update the DocURL unless we have a new doc. Mon Mar 19 18:00:18 2001 Geoff Hutchison * htcommon/URL.cc (URL): Make sure even absolute relative URLs are run through normalizePath() as pointed out by Gilles. Allows backout of previous fix of #408586, which does extra re-parsing of URL. * htdig/Retriever.cc (Need2Get): Back out change of Mar. 17 for above. * htcommon/conf_lexer.[cxx, lxx]: Apply change suggested by Jesse to remove empty statements. Mon Mar 19 11:33:25 2001 Geoff Hutchison * htlib/HtRegexList.cc (setEscaped): Fix assorted bugs, including obvious segfault, incorrect creation of limits, and failure to set "compiled" flag before return(). * htdig/Retriever.cc (IsValidURL): Make sure the tmpList is cleared before attempting to parse the bad_querystr config--otherwise we'll just Add to the end of the list. Sun Mar 18 14:01:56 CET 2001 Gabriele Bartolini * htnet/Transport.[h,cc], htnet/HtHTTP.cc: In order to modularize the net code the default parser string for the content-type has been added to the Transport class. * htdig/Document.cc: modified for the changes above. Sat Mar 17 16:38:27 2001 Geoff Hutchison * configure.in, configure, include/htconfig.h.in: Add tests for libssl, libcrypto, and ssl.h. * htnet/SSLConnection.[cc,h], htnet/HtHTTPBasic.[cc,h], htnet/HTTPSecure.[cc,h]: New files. Contributed by Joshua Gerth . * htnet/Transport.[cc,h], htnet/HtNTTP.cc, htnet/HtHTTP.cc, htnet/Connection.h: Changes needed to support SSLConnection class. * htdig/Document.cc, htdig/Document.h: Ditto. * htnet/Makefile.am, htnet/Makefile.in: Add above for compilation. * htdoc/THANKS.html: Updated with new contributors. Sat Mar 17 15:28:20 2001 Geoff Hutchison * htword/WordContext.cc (Initialize): If HAVE_LIBZ or HAVE_ZLIB_H are not defined, make sure wordlist_compress is set to false. This semi-hack will not be necessary with new mifluz code which does not necessary need zlib. Fixes bug #405761. Sat Mar 17 14:39:17 2001 Geoff Hutchison * htdig/HTML.cc (do_tag): Fixed problems with META descriptions containing newlines, returns or tabs. They are now replaced with spaces. Fixes bug #405771. Sat Mar 17 14:26:55 2001 Geoff Hutchison * htdig/HTML.cc (do_tag): Improve handling of whitespace in META refresh handling. Fixes bug #406244. * htlib/HtRegexList.cc (setEscaped): Make this more efficient by building up larger and larger patterns--when we fail, go back a step and add the pattern in the next loop. This ensures we have a list of the maximum allowable length regexp. * htdig/Retriever.cc (Need2Get): Add change suggested by Yariv Tal to run URLs through the URL parser for cleanup before comparing to the visited list. Fixes bug #408586. Mon Mar 12 13:28:56 2001 Michael Haggerty * htdig/Retriever.cc, htdig/Retriever.h: Fixed two off-by-one errors related to Retriever::factor table. Mon Mar 12 11:25:31 2001 Geoff Hutchison * htlib/Dictionary.cc (Add): Fix comments about add method--it will replace existing keys. Fixes report #407940. Thu Mar 8 15:31:45 2001 Gabriele Bartolini * htnet/HtHTTP.cc: removed an unuseful Tue Mar 6 11:42:10 2001 Geoff Hutchison * htlib/regex.[c,h]: Update with versions from glibc 2.2.2. Mon Mar 5 13:47:30 2001 Geoff Hutchison * ltconfig (host_os): Add test to solve problems building C++ shared libraries on some platforms. Currently should only make --enable-shared the default on Linux and *BSD* unless specified explicitly by the user. Mon Mar 5 12:52:57 2001 Geoff Hutchison * htlib/String.cc (operator =): Add fix contributed by Yariv Tal , fixed bug #406075. Mon Mar 5 12:06:26 2001 Geoff Hutchison * htlib/HtRegexList.cc (match): Ignore rearrangement code for the moment--may or may not be the culprit for bug #405277, but is a start to debugging the problem. * htlib/List.[cc,h]: Remove *prev pointer from listnode structure and add a *prev pointer to the cursor structure. Saves one pointer per item in the list, plus overhead. Mon Mar 5 11:56:16 2001 Geoff Hutchison * htcommon/defaults.cc (bad_extensions): Add .css to ignore CSS docs. * htdig/Document.cc (getParsable): Ignore CSS documents -- they aren't very useful to parse. Solves bug report #405772. Sun Mar 04 11:32:43 2001 Gabriele Bartolini * htnet/HtHTTP.cc: fixed a bug regarding with persistent connections enabled, but head call before the get one disabled. Sourceforge.net's bug reference: 405275 - fixed. Sat Mar 3 21:09:55 2001 Geoff Hutchison * .version: Bump to 3.2.0b4 so snapshots have right versioning. Thu Mar 1 16:51:09 2001 Geoff Hutchison * configure.in: Added test for alloca.h, which is needed for the regex.c code. Wed Feb 28 12:54:43 CEST 2001 Gabriele Bartolini * htcommon/defaults.cc: 'disable_cookies' option has been added, with a 'server' scope. By default it is set to 'false'. * htdig/Server.h, cc: management of the option above has been enhanced. * htnet/HtHTTP.h, cc: now an HTTP connection can disable/enable cookies through the configuration attribute 'disable_cookies'. * htdig/Document.cc: management of cookies enabling/disabling is here. * Cookies classes: now support the expiration time. Need only the subdomain treatment. Mon Feb 26 16:37:30 2001 Geoff Hutchison * htcommon/conf_lexer.lxx: Don't directly call exit(1) on an error condition! Seems a harsh problem for an unknown character. * htcommon/conf_parser.yxx: Ditto. (Running out of memory is a much more fatal condition, of course.) * htcommon/conf_lexer.cxx: Regenerate using flex 2.5.4. * htcommon/conf_parser.cxx: Regenerate using bison 1.28. Sun Feb 25 19:46:01 CEST 2001 Gabriele Bartolini * htnet/HtHTTP.h, cc: support for cookies enabled * htnet/Makefile.am: files for cookies have been added to make. Sun Feb 25 19:27:18 CEST 2001 Gabriele Bartolini * htnet/HtCookie.h,cc: class HTTP cookie * htnet/HtCookieJar.h,cc: abstract class for managing the 'jar' of cookies. In this way, we can use different methods for the storage of them. * htnet/HtCookieMemJar.h,cc: class for managing the 'jar' of cookies in memory, without persistent storage (no db or file). * Many thanks to Robert LaFerla for his coding on this! Yeah, really really thanks Robert! Thu Feb 22 16:43:18 2001 Geoff Hutchison * htdoc/ChangeLog, htdig/RELEASE.html, README: Update to roll the release of 3.2.0b3. Thu Feb 22 16:22:05 2001 Gilles Detillieux * htsearch/htsearch.cc (main), htsearch/Display.cc (setVariables, createURL, buildMatchList), htdoc/hts_form.html, htdoc/hts_templates.html: Add Mike Grommet's date range search feature. Mon Feb 19 18:24:42 2001 Geoff Hutchison * htfuzzy/Synonym.cc (createDB): Create database in a temporary directory before we move it into place, much like the endings code. This should prevent problems when we just append to the DB instead of making a new one. * htdig/htdig.cc (main): Fix bug discovered by Gilles--htword should be initialized *after* we are finished modifying config attributes based on flags and unlink with -i. * installdir/rundig: Fix bug with calling htpurge with -s option. Thu Feb 15 11:03:42 2001 Geoff Hutchison * htdoc/*.html: Update with 2001 copyrights and various changes with the website move for the pending 3.2.0b3 release. Thu Feb 15 10:41:47 2001 Geoff Hutchison * htlib/HtRegexList.cc (match): Fix thinko with logic for matching and add code to rearrange matching nodes for hopefully better performance. Sun Feb 11 16:42:11 2001 Geoff Hutchison * htlib/HtRegexList.h, htlib/HtRegexList.cc (class HtRegexList): Simple List(HtRegex) object with similar calling conventions to HtRegex class. This version is not as sophisticated as it could be, but it's not likely to drop objects when reorganizing. * htlib/Makefile.[in,am]: Add HtRegexList files to list for compilation. * htdig/htdig.h, htdig/htdig.cc, htdig/Retriever.cc: Use HtRegexList instead of HtRegex for setting escaped values--should never fail (since each String item is short). * htlib/HtDateTime.cc: Put back timezone specs into the output formats so we give everything even if we ignore it when reading input. Mon Feb 5 11:47:07 2001 Geoff Hutchison * htlib/HtDateTime.cc: Remove the timezone specs in the date formats--these are not required in the RFCs because many dates are in GMT anyway. Wed Jan 17 08:48:30 2001 Gilles Detillieux * htdig/ExternalTransport.cc (Request): Oops, fixed a holdover from code borrowed from ExternalParser.cc's fork handling. Mon Jan 15 23:09:37 2001 Geoff Hutchison * htnet/Connection.cc: Back out previous change--this should not in any way be needed since the configure script should set FD_SET_T. * configure.in, configure: Add more lenient prototyping for select() test--now allows "const struct timeval" for compilation on BSDI. * htdoc/RELEASE.html: Update with Gilles's changes. * htdoc/cf_blocks.html: New file describing and blocks. * htdoc/cf_general.html, htdoc/confmenu.html: Refer to the above. Mon Jan 15 17:46:07 2001 Gilles Detillieux * htsearch/TemplateList.cc (createFromString), htcommon/defaults.cc: Treat template_map as a _quoted_ string list. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Mon Jan 15 17:40:45 2001 Gilles Detillieux * htdoc/hts_templates.html: Add METADESCRIPTION variable. * htsearch/Display.cc (displayMatch): Add METADESCRIPTION variable. * htdig/ExternalParser.cc (parse): Fix up handling of arguments. * htdig/ExternalTransport.cc (Request): Fix up handling of fork/exec and command arguments, add wait() call. Wed Jan 10 19:23:36 2001 Gilles Detillieux * installdir/rundig: Fix -a handling to move db.words.db.work_weakcmpr into place if it exists Sat Jan 6 21:50:58 2001 Geoff Hutchison * configure.in: Add checks for and for ExternalParser. * include/htconfig.h.in: Regenerate using autoheader. * configure: Regenerate using configure. * htnet/Connection.cc: Add definition for FD_SET_T to fix problems compiling on BSDI mentioned by Joe. * htdig/ExternalParser.cc: Use or as appropriate. Should fix problems with compiliation mentioned by Jesse on HP/UX. * README, htdoc/RELEASE.html: Adjust dates for the new year. * htdoc/upgrade.html: A few "remaining features" have been implemented. Sun Dec 06 19:46:15 CEST 2000 Gabriele Bartolini * htnet/HtHTTP.cc: Fixed bug for Read_Line function call in ReadChunkedBody method. Many thanks to Robert LaFerla. ;-) Tue Dec 12 13:24:49 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Fixed to properly handle binary output from an external converter. Fixed some compilation errors. Tue Dec 12 12:52:14 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Handle parser command string as a string list again to allow arguments, build up argv and use execv instead of execl. Tue Dec 12 12:25:04 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Add call to wait for child process, to avoid zombie buildup. Mon Dec 11 23:57:43 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Fix up handling of fds in child process, more fault-tolerant handling of pipe or fork errors. Mon Dec 11 23:30:55 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Fix up handling of creation of temporary file, check for proper return code, give error if appropriate. Mon Dec 11 23:19:28 2000 Gilles Detillieux * htdig/ExternalParser.cc (parse): Lowercase content-types and strip off any trailing semicolons, at one last spot. This reinserts code added Sep 11, which was dropped Oct 9, probably inadvertantly during mifluz back-out. Sun Dec 10 15:28:44 2000 Geoff Hutchison * htdig/ExternalTransport.cc: Use fork/exec instead of calling popen, which bypasses any shell escape problems. * htdig/ExternalParser.cc: Ditto, plus use of mkstemp where available to pick the filename. * configure, configure.in: Check for mkstemp where available. * include/htconfig.h.in: Define it as above. * htlib/Makefile.am: Omit regex.c from SOURCES--this is included when necessary by the configure script. Otherwise this produces duplicate declarations, etc. * htlib/Makefile.in: Regenerate using automake --foreign. * htcommon/URL.cc: Fix bug with ports of 0 showing up in URLs like mailto: or other less-common protocols. Fri Dec 1 14:45:33 2000 Gilles Detillieux * contrib/htdig-3.2.0.spec: Updated to 3.2.0b3. Fri Dec 1 13:59:09 2000 Gilles Detillieux * htlib/Makefile.am: Fix pkginclude_HEADERS to list missing headers ber.h, libdefs.h, myqsort.h, mhash_md5.h, omit unneeded langinfo.h; fix libht_la_SOURCES to list missing sources regex.c, myqsort.c. * htlib/Makefile.in: Regenerate using automake --foreign * htlib/langinfo.h, htlib/nl_types.h: Removed as they're now unused. Fri Dec 1 13:22:47 2000 Gilles Detillieux * htlib/strptime.cc (mystrptime): make ptr const and use cast on return value to avoid warnings. * htlib/Makefile.am: Fix pkginclude_HEADERS to list HtRegexReplace*.h rather than .cc. * htlib/Makefile.in: Regenerate using automake --foreign Fri Dec 1 11:58:21 2000 Gilles Detillieux * Makefile.in, [hit]*/Makefile.in: Regenerate using automake --foreign after fixing bug with cp -pr in automake. Thu Nov 30 14:41:58 2000 Gilles Detillieux * htdoc/Makefile.am: Removed howitworks.html from EXTRA_DIST. * Makefile.in (distdir): Added missing variable name 'd' to cp -pr. Thu Nov 30 14:01:48 2000 Gilles Detillieux * htlib/strptime.cc, htlib/lib.h: make first 2 args to strptime const to avoid warnings, use cast in asizeof to avoid warnings. * htsearch/qtest.cc: Change include from iostream to iostream.h * htsearch/DocMatch.cc: Change include from iostream to iostream.h * htsearch/Display.cc (createURL, buildMatchList, excerpt, hilight): Clean up code to get rid of warnings, especially resulting from NULLs in ternary operators. Thu Nov 30 10:55:09 2000 Gilles Detillieux * htlib/String_fmt.cc (form, vform): Use vsnprintf rather than vsprintf, for buffer overflow prevention if vsnprintf available. * htdig/Retriever.cc: Remove unused strptime declaration. * htlib/HtDateTime.cc: Use mystrptime if HAVE_STRPTIME not set. Wed Nov 29 23:31:10 2000 Geoff Hutchison * htdb/htdb_stat.cc, htdb_load.cc, htdb_dump.cc: Make sure we include htconfig.h to include proper declarations. * htlib/strptime.cc: Change to strptime.cc, from htdig-3.1 series hopefully more portable until I can find a more suitable replacement. * htlib/Makefile.am, htlib/Makefile.in: As above. * htlib/clib.h, htlib/lib.h: Ditto. * htdoc/all.html: Add a first draft of program summaries. Wed Nov 29 18:00:15 2000 Gilles Detillieux * htdig/Retriever.cc (parse_url): Remove undeclared "dup" variable, add missing calls to words.Skip(). Wed Nov 29 17:44:56 2000 Gilles Detillieux * htdig/htdig.html: Add description of -v output. Mon Nov 27 12:03:34 2000 Gilles Detillieux * htlib/md5.cc: Added missing include of time.h Fri Nov 24 00:56:01 2000 Toivo Pedaste * htsearch/Display.cc: Some extra debugging for scoring Sun Nov 19 00:56:01 2000 Geoff Hutchison * htnet/HtFile.cc (Request): Use opendir/readdir instead of scandir for generating directory listings on-the-fly. * htdoc/RELEASE.html: Write up release notes for 3.2.0b3. * htdoc/THANKS.html: Update list of contributors for 3.2.0b3 as current. Fri Nov 17 14:52:37 2000 Gilles Detillieux * contrib/acroconv.pl: Added external converter script to convert PDFs with acroread. Mon Nov 6 12:13:13 2000 Gilles Detillieux * htdig/Retriever.cc (GetLocal, GetLocalUser): move String definition out of while statement for AIX xlC compiler. Mon Oct 30 21:50:02 2000 Geoff Hutchison * htdig/Server.h, htdig/Server.cc (push): Add newDoc paramter that will allow redirects (old docs) to be followed and not count against the maxDoc restrictions. * htdig/Retriever.cc (got_redirect): Use new parameter so we don't count against a server's max documents since it's a redirect. * htlib/nl_types.h: Add for systems missing this header file. Sun Oct 29 21:36:51 2000 Geoff Hutchison * htcommon/defaults.cc: Updated per-server and per-URL fields to match code. I still have a "wish list" of additional attributes that should work this way eventually. * htdoc/attrs.html, cf_by{name,prog}.html: Reran cf_generate.pl. Sun Oct 22 17:13:08 2000 Geoff Hutchison * htcommon/HtWordList.h: Add missing include for stdlib.h needed for abort(). * htsearch/BooleanQueryParser.cc (ParseAnd): Fix problems with RH7 compiler -- shouldn't use "not" as a variable name! Thu Oct 19 22:19:16 2000 Geoff Hutchison * ltmain.sh, ltconfig: Update with versions from libtool 1.3.5. which may fix some problems building libraries. Mon Oct 9 21:59:11 2000 Geoff Hutchison * */* [many, many files]: Backed out mifluz merge by going back on modified files to 091000 snapshot. * configure: Regenerated from configure.in. * */Makefile.in: Regenerated using automake. Fri Oct 6 11:03:14 2000 Gilles Detillieux * htdig/HTML.cc (do_tag): Parse