2005-12-15 Mikio Hirabayashi
* estraier.c (est_db_open): "ESTDBHUGE" was added.
* estraier.c (est_regex_replace): "&" was to be handled.
* estraier.c (est_cond_set_expander): new function.
* estcmd.c (runput, rungather): "-lg" was renamed to "-xl",
and "-xh" was added.
* estseek.c (showdoc): a bug about calculating similarity
was fixed.
* estseek.c (realmain): "qxpndcmd" was added.
* estseek.c (expandquery): new function.
* estmaster.c (realmain, startup): "-ro" option was added.
* estmaster.c (sendnodecmdsearchui): the output format
modified for OpenSearch 1.1.
- Release: 1.1.2
2005-12-04 Mikio Hirabayashi
* estraier.c (est_keyword_num, est_keyword_rec_size): new
functions.
* estraier.c (est_keyword_iter_init, est_keyword_iter_next):
new functions.
* estraier.c (est_search_union): scoring was to use standard
deviation.
* estraier.c (est_expand_keyword_bw, est_expand_keyword_ew):
new functions.
* estraier.c (est_db_open): "ESTDBLARGE" was added.
* estseek.c (realmain, setsimilarphrase): "smlrtune" was
added.
* estseek.c (realmain, showpage): "relkeynum" was added.
* estmaster.c (startup, setsimilarphrase): "uismlrtune"
was added.
* estmaster.c (sendnodecmdsearchatom, nodecmdopensearch):
conformed to draft 2.
- Release: 1.1.1
2005-11-11 Mikio Hirabayashi
* estraier.c (est_eclipse_scores): accuracy was improved.
* estraier.c (est_narrow_scores): a bug of segmentation
fault was fixed.
* estraier.c (est_regex_new, est_regex_delete, est_regex_match):
new functions.
* estraier.c (est_expand_word_rx): new function.
* estreier.c (est_db_put_keywords, est_db_get_keywords):
new functions.
* estraier.c (est_db_measure_doc): new function.
* estraier.c (est_db_set_vectorizer): abolished.
* estcmd.c (runextkeys, procextkeys): "-no" was added.
* estcmd.c (runpurge, procpurge, runextkeys, procextkeys):
"-attr" was added.
* estcmd.c (rundraft, procdraft): "-kn" was added.
* estcmd.c (procgather): a bug of handling drive characters
on Windows was fixed.
* estseek.c (realmain, makeshownuri): "lprefix" and so on
were abolished.
* estmaster.c (startup, makeshownuri): "uilprefix" and so
on were abolished.
* estmaster.c (startup, sendnodecmdputdoc): "lateextkeys"
was added.
* estmaster.c (searchlocal): a bug of segmentation fault
was fixed.
- Release: 1.0.6
2005-11-08 Mikio Hirabayashi
* estraier.c (est_narrow_scores): speed of attribute search
was improved.
* estcmd.c (doputdoc, procpurge, procextkeys): local paths
was to be printed.
* estcmd.c (doputdoc): "@size" was to be added even with
"-fo" option.
* estcmd.c (procsearch): "_lfile" was to be printed when
human readable mode.
* estcmd.c (procout, procedit, procget, procuriid): retrieving
by local path was added.
* estcmd.c (procpurge): retrieving by local path was added.
* estcmd.c (runiconv, prociconv): new functions.
* estcmd.c (procedit): "-pc" was to convert the encoding
of the attribute value.
* estseek.c (realmain, showform): "clipview" attribute was
added.
- Release: 1.0.5
2005-11-02 Mikio Hirabayashi
* estraier.c (est_vector_cosine): new function.
* estraier.c (est_eclipse_scores): new function.
* estcmd.c (runsearch, procsearch): "-ec" option was added.
* estcmd.c (procgather, procpurge): "-no" option was added.
* estcmd.c (urltopath): a bug about drive letter was fixed.
* estcmd.c (runlist, proclist): "-lp" option was added.
* estseek.cgi (showresult): similar eclipse was implemented.
* estmaster.c (startup, searchlocal): "rateuri" attribute
was added.
* estmaster.c (sendmenudata): new function.
- Release: 1.0.4
bump PKGREVISION where necessary
Move PKG_TEXMFPREFIX and PKG_LOCALTEXMFPREFIX definitions to
teTeX?/buidlink3.mk, so that packages may include
print/teTeX1-bin/buildlink3.mk directly (however, using teTeX/module.mk do not
allow that)
Complete list of changes from 0.60.3:
* Fixed a bug that caused Aspell to crash when checking certain
Russian words, this bug likely affected other languages as well.
* Updated to Gettext 0.14.5 which is required for AMD64, also
updated to to Libtool 1.5.20.
* Fixed an alignment bug which causes mmap to always fail when
reading in dictionaries.
* Added note about how `make clean' will remove the HTML manuals.
* Added manual page for prezip-bin and enhanced word-list-compress
manual page thanks to the work of Jose Da Silva.
* Other minor updates and bug fixes.
rather than PKG_FAIL_REASON, so that they provide useful error
messages in build logs, and so that they continue to work on platforms
where they aren't broken.
1.57
- (revision bumped to match XML::LibXML)
- Some minor fixes to perl extension functions support
- Fixed callbacks support
- Enable xinclude in XSLT files
- Fixed some memory leaks
- Fixed some segfaults
New Features:
Major changes in ICU 3.4 include the following:
Updates to conform to Unicode 4.1, including new characters
properties and values, text segmentation, plus collation
updated for Unicode Technical Standard #10 (UCA) and regex
updated for Unicode Technical Standard #18.
* Updates to conform to the Common Locale Data
Repository (CLDR), Version 1.3 for the latest locale
data. This includes:
* New data to support localization of timezones,
United Nations M.49 regions (including continents
and regions), mappings from language to script
and territory.
* Consolidation of inherited data and improved
resource aliasing for smaller data footprint
* Additional locales, and many other fixes and
additions of locale data.
* POSIX migration support: direct API support for all
POSIX character classes, implemented according to
Unicode recommendations
> program requires Xaw, which requires the Xlib shape routines, which on
> Interix are in Xext. http://www.faqs.org/faqs/x-faq/part6/section-27.html
Move fix from Makefile to options.mk, so it is only used when the x11
option is enabled.
* Support for Atom 1.0.
* Support for iTunes extensions.
* Support for dc:contributor.
* Universal Feed Parser now captures the feed's namespaces. See
Namespace Handling for details.
* Lots of things have been renamed to match Atom 1.0 terminology.
issued is now entries[i].published, modified is now entries[i].updated,
and url is now href everywhere. You can still access these elements
with the old names, so you shouldn't need to change any existing
code, but don't be surprised if you can't find them during
debugging.
* category and categories have been replaced by tags, see feed.tags
and entries[i].tags. The old names still work.
* mode is gone from all detail and content dictionaries. It was
never terribly useful, since Universal Feed Parser unescapes
content automatically.
* entries[i].source is now a dictionary of feed metadata as per
section 4.2.11 of RFC 4287. Universal Feed Parser no longer
supports the RSS 2.0's source element.
* Content in unknown namespaces is no longer discarded (bug 993305)
* Lots of other bug fixes.
