chiark / gitweb /
disorder
16 years agooops
Richard Kettlewell [Sun, 25 Nov 2007 18:23:35 +0000 (18:23 +0000)]
oops

16 years agotests/queue.py: check that the queue is padded to the right length
Richard Kettlewell [Sun, 25 Nov 2007 18:13:28 +0000 (18:13 +0000)]
tests/queue.py: check that the queue is padded to the right length

16 years agocookie header parser
Richard Kettlewell [Sun, 25 Nov 2007 14:55:00 +0000 (14:55 +0000)]
cookie header parser

16 years agomore README fiddling
Richard Kettlewell [Sat, 24 Nov 2007 18:44:32 +0000 (18:44 +0000)]
more README fiddling

16 years agoexpand streaming instructions
Richard Kettlewell [Sat, 24 Nov 2007 15:58:13 +0000 (15:58 +0000)]
expand streaming instructions

16 years agoupgrade instructions
Richard Kettlewell [Fri, 23 Nov 2007 17:49:24 +0000 (17:49 +0000)]
upgrade instructions

16 years agoupdate README* and CHANGES
Richard Kettlewell [Fri, 23 Nov 2007 17:41:24 +0000 (17:41 +0000)]
update README* and CHANGES

16 years agodebian release 1.5.99+dev9 debian-1_5_99dev9
Richard Kettlewell [Fri, 23 Nov 2007 17:35:23 +0000 (17:35 +0000)]
debian release 1.5.99+dev9

16 years agomerge disorder.unicode branch
rjk@greenend.org.uk [Fri, 23 Nov 2007 16:59:24 +0000 (16:59 +0000)]
merge disorder.unicode branch

16 years agoFixes for disorder-dbpugrade:
rjk@greenend.org.uk [Fri, 23 Nov 2007 16:33:50 +0000 (16:33 +0000)]
Fixes for disorder-dbpugrade:
   - increased verbosity
   - bad (non-UTF-8) keys are now logged but ignored by default
   - ...but they can be deleted or abort the upgrade
   - aliases that sort later than the track they alias no longer
     abort the upgrade

16 years agomerge from disorder.dev
rjk@greenend.org.uk [Fri, 23 Nov 2007 14:39:56 +0000 (14:39 +0000)]
merge from disorder.dev

16 years agoinit script always address local instance of daemon
rjk@greenend.org.uk [Fri, 23 Nov 2007 14:32:18 +0000 (14:32 +0000)]
init script always address local instance of daemon

16 years agosynchronize with disorder.dev
Richard Kettlewell [Fri, 23 Nov 2007 14:22:30 +0000 (14:22 +0000)]
synchronize with disorder.dev

16 years ago'plugins' and other _accum configuration items can now have the list
Richard Kettlewell [Fri, 23 Nov 2007 14:15:43 +0000 (14:15 +0000)]
'plugins' and other _accum configuration items can now have the list
cleared by supplying no arguments.  This is used by the test scripts to
ensure they aren't picking up install versions of the plugins.

This is on disorder.unicode rather than disorder.dev because the tests
that matter are all there...

16 years agomake -j safe clients/Makefile.am
Richard Kettlewell [Fri, 23 Nov 2007 12:26:04 +0000 (12:26 +0000)]
make -j safe clients/Makefile.am

16 years agoRemove post_move_cleanup(). The effect is that new chosen tracks are
Richard Kettlewell [Fri, 23 Nov 2007 12:21:51 +0000 (12:21 +0000)]
Remove post_move_cleanup().  The effect is that new chosen tracks are
always added just before the final block of random tracks, and there
is no longer any invisible state in the queue.

16 years agoRemove combining characters from search strings. The effect is that
Richard Kettlewell [Fri, 23 Nov 2007 11:05:56 +0000 (11:05 +0000)]
Remove combining characters from search strings.  The effect is that
accents are ignored in searching.  This is what English speakers
usually expect; if DisOrder becomes popular among a wider audience
then only trackdb_search() needs to be updated to filter out false
positives, the database can keep the accent-free keys.

16 years agocopyright date
Richard Kettlewell [Thu, 22 Nov 2007 17:52:15 +0000 (17:52 +0000)]
copyright date

16 years agodebian version 1.5.99+dev8 debian-1_5_99dev8
Richard Kettlewell [Thu, 22 Nov 2007 17:22:24 +0000 (17:22 +0000)]
debian version 1.5.99+dev8

16 years agotypo in man disorder-playrtp
Richard Kettlewell [Thu, 22 Nov 2007 17:16:17 +0000 (17:16 +0000)]
typo in man disorder-playrtp

16 years agorationalize playrtp multicast support
Richard Kettlewell [Thu, 22 Nov 2007 16:47:33 +0000 (16:47 +0000)]
rationalize playrtp multicast support

16 years agounicode search tests
Richard Kettlewell [Thu, 22 Nov 2007 13:59:33 +0000 (13:59 +0000)]
unicode search tests

16 years agofurther search testing
Richard Kettlewell [Thu, 22 Nov 2007 13:47:35 +0000 (13:47 +0000)]
further search testing

16 years agotest search
Richard Kettlewell [Thu, 22 Nov 2007 13:21:16 +0000 (13:21 +0000)]
test search

16 years agofix disorder.py search
Richard Kettlewell [Thu, 22 Nov 2007 13:20:48 +0000 (13:20 +0000)]
fix disorder.py search

16 years agosync with trunk
Richard Kettlewell [Thu, 22 Nov 2007 13:17:54 +0000 (13:17 +0000)]
sync with trunk

16 years agofixes from last change l-(
Richard Kettlewell [Thu, 22 Nov 2007 11:44:59 +0000 (11:44 +0000)]
fixes from last change l-(

16 years agoThe initial rescan now blocks any client connections. This means that
Richard Kettlewell [Thu, 22 Nov 2007 11:35:34 +0000 (11:35 +0000)]
The initial rescan now blocks any client connections.  This means that
a deadlock is unlikely to arise in the initial scan.  It also means
that the server is not usable for a short period after startup (or
quite a long period after initial startup, but that's less of a
problem).

The tests now take advantage of this by waiting until the socket
appears before assuming the daemon has started.

16 years agobuild fixes
Richard Kettlewell [Thu, 22 Nov 2007 11:08:54 +0000 (11:08 +0000)]
build fixes

16 years agoautomatically upgrade on startup if necessary
Richard Kettlewell [Thu, 22 Nov 2007 10:26:50 +0000 (10:26 +0000)]
automatically upgrade on startup if necessary

16 years agotidy up upgrade/recovery flags a bit
Richard Kettlewell [Thu, 22 Nov 2007 09:44:43 +0000 (09:44 +0000)]
tidy up upgrade/recovery flags a bit

16 years agocheck server runs after upgrade
Richard Kettlewell [Wed, 21 Nov 2007 19:20:15 +0000 (19:20 +0000)]
check server runs after upgrade

16 years agodatabase upgrade tool. needs to be run manually.
Richard Kettlewell [Wed, 21 Nov 2007 19:15:50 +0000 (19:15 +0000)]
database upgrade tool.  needs to be run manually.

16 years agosync up with disorder.dev
Richard Kettlewell [Wed, 21 Nov 2007 11:18:34 +0000 (11:18 +0000)]
sync up with disorder.dev

16 years agomore disorder.unicode test alignment
Richard Kettlewell [Wed, 21 Nov 2007 11:12:22 +0000 (11:12 +0000)]
more disorder.unicode test alignment

16 years agooverride disorder.py config
Richard Kettlewell [Wed, 21 Nov 2007 11:05:41 +0000 (11:05 +0000)]
override disorder.py config

16 years agoalign with disorder.unicode a bit
Richard Kettlewell [Wed, 21 Nov 2007 10:52:21 +0000 (10:52 +0000)]
align with disorder.unicode a bit

16 years agoborrow some test naming conventions from disorder.unicode
Richard Kettlewell [Wed, 21 Nov 2007 10:44:34 +0000 (10:44 +0000)]
borrow some test naming conventions from disorder.unicode

16 years agomore test fiddling
Richard Kettlewell [Wed, 21 Nov 2007 10:41:23 +0000 (10:41 +0000)]
more test fiddling

16 years agosync with disorder.dev
Richard Kettlewell [Wed, 21 Nov 2007 10:23:02 +0000 (10:23 +0000)]
sync with disorder.dev

16 years agomore automation of test logic
Richard Kettlewell [Wed, 21 Nov 2007 10:14:54 +0000 (10:14 +0000)]
more automation of test logic

16 years agotests figure out paths more automatically
Richard Kettlewell [Wed, 21 Nov 2007 10:03:19 +0000 (10:03 +0000)]
tests figure out paths more automatically

16 years agotest to verify current dbversion behavior
Richard Kettlewell [Wed, 21 Nov 2007 09:31:29 +0000 (09:31 +0000)]
test to verify current dbversion behavior

16 years agofix indent
Richard Kettlewell [Wed, 21 Nov 2007 00:42:34 +0000 (00:42 +0000)]
fix indent

16 years ago(c) notice on tests
Richard Kettlewell [Wed, 21 Nov 2007 00:39:55 +0000 (00:39 +0000)]
(c) notice on tests

16 years agoneed python >= 2.4 for tests
Richard Kettlewell [Wed, 21 Nov 2007 00:36:49 +0000 (00:36 +0000)]
need python >= 2.4 for tests

16 years agounicode.c iterators can now have their notion of Word_Break tailored.
Richard Kettlewell [Tue, 20 Nov 2007 20:32:06 +0000 (20:32 +0000)]
unicode.c iterators can now have their notion of Word_Break tailored.

We use this to arrange that underscores are treated as spaces, since a
common track naming convention replaces spaces with underscores.  We
only apply this to raw filenames, not to the overrides provided by
preferences, on the assumption that if you typed an underscore there
you really meant it.

words.c/h are now gone; all references to words() and casefold() use
unicode.c functions directly.

16 years agoutf32_word_split() and utf8_word_split() splits a string into words
Richard Kettlewell [Tue, 20 Nov 2007 18:13:56 +0000 (18:13 +0000)]
utf32_word_split() and utf8_word_split() splits a string into words
using the UAX #29 word boundary algorithm.  words() is therefore now a
wrapper around this.  There is scope for improvement in the use of
this function as currently we do some needless converting back and
forth between encoding forms.

casefold() now uses the compatibility case-folding algorithm, which
seems more appropriate for searching.

dbversions are now integers not strings.  Some dbversion=2
functionality can be selectively disabled for testing purposes.

README.dbversions documents the differences between the dbversions.

16 years agonormalize recorded filenames so files.py passes
Richard Kettlewell [Tue, 20 Nov 2007 15:38:46 +0000 (15:38 +0000)]
normalize recorded filenames so files.py passes

16 years agotests/files.py; ok except for the non-ascii filenames
Richard Kettlewell [Tue, 20 Nov 2007 15:10:24 +0000 (15:10 +0000)]
tests/files.py; ok except for the non-ascii filenames

16 years agofix test config indent, oops...
Richard Kettlewell [Tue, 20 Nov 2007 12:38:40 +0000 (12:38 +0000)]
fix test config indent, oops...

16 years agoConvert track names and input lines to NFC. This is a database format
Richard Kettlewell [Tue, 20 Nov 2007 12:31:03 +0000 (12:31 +0000)]
Convert track names and input lines to NFC.  This is a database format
change so a new read-only global preference _dbversion is introduced
to record the current database version.  Upgrade will implyy
renormalizing all keys and regenerating the search/tags databases.

Fiddled with tests to put both composed and decomposed non-ASCII names
through.

Document limitations on line syntax.  This really needs improving
before release since the same thing applies to the config file!

Make 'nothing' test quicker.

16 years agosync with disorder.dev
Richard Kettlewell [Mon, 19 Nov 2007 21:28:19 +0000 (21:28 +0000)]
sync with disorder.dev

16 years agobuild fix for linux
Richard Kettlewell [Mon, 19 Nov 2007 21:27:00 +0000 (21:27 +0000)]
build fix for linux

16 years agoteach test environment about track types
Richard Kettlewell [Mon, 19 Nov 2007 21:21:50 +0000 (21:21 +0000)]
teach test environment about track types

16 years agoit's plugins not plugin
Richard Kettlewell [Mon, 19 Nov 2007 21:21:34 +0000 (21:21 +0000)]
it's plugins not plugin

16 years agoadd a nontrivial test that actually works
Richard Kettlewell [Mon, 19 Nov 2007 21:15:26 +0000 (21:15 +0000)]
add a nontrivial test that actually works

16 years agotest path fixes
Richard Kettlewell [Mon, 19 Nov 2007 21:05:25 +0000 (21:05 +0000)]
test path fixes

16 years agoremove some pointless stderr redirection
Richard Kettlewell [Mon, 19 Nov 2007 21:05:16 +0000 (21:05 +0000)]
remove some pointless stderr redirection

16 years agodon't fatal() if cannot create a root account
Richard Kettlewell [Mon, 19 Nov 2007 21:04:31 +0000 (21:04 +0000)]
don't fatal() if cannot create a root account

16 years agofix syslog test
Richard Kettlewell [Mon, 19 Nov 2007 21:04:08 +0000 (21:04 +0000)]
fix syslog test

16 years ago--[no-]syslog for disorder-normalize too
Richard Kettlewell [Mon, 19 Nov 2007 20:54:28 +0000 (20:54 +0000)]
--[no-]syslog for disorder-normalize too

16 years agoserver subprocesses should log to the same place the server does
Richard Kettlewell [Mon, 19 Nov 2007 20:47:24 +0000 (20:47 +0000)]
server subprocesses should log to the same place the server does

16 years agoDisobedience now honors config->short_display. This is a bit grotty
Richard Kettlewell [Mon, 19 Nov 2007 17:01:19 +0000 (17:01 +0000)]
Disobedience now honors config->short_display.  This is a bit grotty
as we have access to enough information to do a better job of
truncation, but it'll do for now.

16 years agotruncate_for_display() now uses unicode.c grapheme boundary deetection
Richard Kettlewell [Mon, 19 Nov 2007 16:56:21 +0000 (16:56 +0000)]
truncate_for_display() now uses unicode.c grapheme boundary deetection

16 years agoUTF8 versions of NFC/NFKC
Richard Kettlewell [Mon, 19 Nov 2007 16:15:09 +0000 (16:15 +0000)]
UTF8 versions of NFC/NFKC

16 years agosync up with disorder.dev
Richard Kettlewell [Mon, 19 Nov 2007 16:06:23 +0000 (16:06 +0000)]
sync up with disorder.dev

16 years agodoxygen fixes
Richard Kettlewell [Mon, 19 Nov 2007 16:05:06 +0000 (16:05 +0000)]
doxygen fixes

16 years agoNFC and NFKC support
Richard Kettlewell [Mon, 19 Nov 2007 15:58:27 +0000 (15:58 +0000)]
NFC and NFKC support

16 years agoGenerate, but do not use, composition mappings.
Richard Kettlewell [Mon, 19 Nov 2007 10:17:04 +0000 (10:17 +0000)]
Generate, but do not use, composition mappings.

UTF-32 sequences for decomposition etc are now de-duped across the
full set of sequences, rather than privately by each generator.  Total
saving is a couple of thousand tables currently.

16 years agodistinguish the different decompositions via a flag bit
Richard Kettlewell [Sun, 18 Nov 2007 23:49:51 +0000 (23:49 +0000)]
distinguish the different decompositions via a flag bit

16 years agotest and fix utf32_iterator_set()
Richard Kettlewell [Sun, 18 Nov 2007 23:07:22 +0000 (23:07 +0000)]
test and fix utf32_iterator_set()

16 years agomore efficient utf32_iterator_set()
Richard Kettlewell [Sun, 18 Nov 2007 22:53:32 +0000 (22:53 +0000)]
more efficient utf32_iterator_set()

16 years agoreplace main utf-8 parser with table-driven one
Richard Kettlewell [Sun, 18 Nov 2007 22:34:06 +0000 (22:34 +0000)]
replace main utf-8 parser with table-driven one

16 years agotransition various bits of code to unicode.h interfaces
Richard Kettlewell [Sun, 18 Nov 2007 22:23:35 +0000 (22:23 +0000)]
transition various bits of code to unicode.h interfaces

16 years agotable-drive UTF-8 validity checker
Richard Kettlewell [Sun, 18 Nov 2007 22:03:51 +0000 (22:03 +0000)]
table-drive UTF-8 validity checker

16 years agonew unicode string iterator give cleaner boundary testing
Richard Kettlewell [Sun, 18 Nov 2007 21:36:39 +0000 (21:36 +0000)]
new unicode string iterator give cleaner boundary testing

16 years agoneed gnu sed these days
Richard Kettlewell [Sun, 18 Nov 2007 16:59:44 +0000 (16:59 +0000)]
need gnu sed these days

16 years agomore space efficient unicode tables
Richard Kettlewell [Sun, 18 Nov 2007 15:44:50 +0000 (15:44 +0000)]
more space efficient unicode tables

16 years agomore idiomatic grapheme breaking
Richard Kettlewell [Sun, 18 Nov 2007 15:44:12 +0000 (15:44 +0000)]
more idiomatic grapheme breaking

16 years agographeme boundary check can now use tables
Richard Kettlewell [Sun, 18 Nov 2007 15:27:36 +0000 (15:27 +0000)]
grapheme boundary check can now use tables

16 years agoword break now comes from the table
Richard Kettlewell [Sun, 18 Nov 2007 15:22:01 +0000 (15:22 +0000)]
word break now comes from the table

16 years agoall characters now in the table, which is broken to save space
Richard Kettlewell [Sun, 18 Nov 2007 14:57:56 +0000 (14:57 +0000)]
all characters now in the table, which is broken to save space

16 years agostart on ut32__unidata() which will provide a uniform interface
Richard Kettlewell [Sun, 18 Nov 2007 14:26:59 +0000 (14:26 +0000)]
start on ut32__unidata() which will provide a uniform interface

16 years agodon't be fooled by First/Last entries in UnicodeData.txt
Richard Kettlewell [Sun, 18 Nov 2007 12:41:24 +0000 (12:41 +0000)]
don't be fooled by First/Last entries in UnicodeData.txt

16 years agounicode_gc_cat -> unicode_General_Category
Richard Kettlewell [Sun, 18 Nov 2007 12:14:24 +0000 (12:14 +0000)]
unicode_gc_cat -> unicode_General_Category

16 years agocompute Sentence_Break as well
Richard Kettlewell [Sun, 18 Nov 2007 12:10:05 +0000 (12:10 +0000)]
compute Sentence_Break as well

16 years agotest and corrections for utf32_is_word_boundary()
Richard Kettlewell [Sun, 18 Nov 2007 11:31:31 +0000 (11:31 +0000)]
test and corrections for utf32_is_word_boundary()

16 years agountested utf32_is_word_boundary() and associated table changes
Richard Kettlewell [Sun, 18 Nov 2007 02:47:51 +0000 (02:47 +0000)]
untested utf32_is_word_boundary() and associated table changes

16 years agotests and a correction for grapheme cluster boundary detection
Richard Kettlewell [Sun, 18 Nov 2007 01:16:16 +0000 (01:16 +0000)]
tests and a correction for grapheme cluster boundary detection

16 years agountested grapheme cluster boundary detection
Richard Kettlewell [Sat, 17 Nov 2007 22:01:22 +0000 (22:01 +0000)]
untested grapheme cluster boundary detection

16 years agoImplement the s3.13 compatibility casefolding transformation as
Richard Kettlewell [Sat, 17 Nov 2007 20:34:40 +0000 (20:34 +0000)]
Implement the s3.13 compatibility casefolding transformation as
utf32_casefold_compat().

Fix checking the surrogates range!

16 years agoStart of Unicode support rewrite
Richard Kettlewell [Sat, 17 Nov 2007 20:07:23 +0000 (20:07 +0000)]
Start of Unicode support rewrite

unicode.c/.h provide a new (not yet complete) Unicode API which implements
algorithms as described in the Unicode 5.0.0 standard.  lib/test and words.c
are somewhat modified to use them.

By default the make-unidata script is NEVER run.  It must be run manually.
This policy might be changed in the future; the script is capable of
downloading any files it needs to operate.

16 years agotone down logging some more
Richard Kettlewell [Fri, 16 Nov 2007 19:17:42 +0000 (19:17 +0000)]
tone down logging some more

16 years agotone down logging
Richard Kettlewell [Fri, 16 Nov 2007 19:14:46 +0000 (19:14 +0000)]
tone down logging

16 years agobuild fix
Richard Kettlewell [Fri, 16 Nov 2007 19:14:39 +0000 (19:14 +0000)]
build fix

16 years agomore systematic elimination of feedback from the gui
Richard Kettlewell [Fri, 16 Nov 2007 13:24:49 +0000 (13:24 +0000)]
more systematic elimination of feedback from the gui

16 years agodisobedience survives losing TCP connections better
Richard Kettlewell [Fri, 16 Nov 2007 13:08:32 +0000 (13:08 +0000)]
disobedience survives losing TCP connections better

16 years agodiscard data sent to logging client; more close fixing
Richard Kettlewell [Tue, 13 Nov 2007 22:23:41 +0000 (22:23 +0000)]
discard data sent to logging client; more close fixing

16 years agofix server crash if log connection fails
Richard Kettlewell [Tue, 13 Nov 2007 20:44:37 +0000 (20:44 +0000)]
fix server crash if log connection fails