don [Fri, 19 Aug 2005 12:44:28 +0000 (04:44 -0800)]
[project @ 2005-08-19 05:44:28 by don]
* Fix spamming on closing of merged bugs (closes: 323846) by actually
understanding what the for $ref (@process) loop was doing.
don [Thu, 18 Aug 2005 04:46:17 +0000 (20:46 -0800)]
[project @ 2005-08-17 21:46:13 by don]
* Add Debbugs::Mail module that has two important functions:
send_mail_message and encode_headers. All mail handling in service.in
and process.in now uses send_mail_message to send mail messages which
tries as hard as possible to send a message; if it fails, only
warnings are returned. This fixes bad addresses causing sendmail to
exit and destroying the bug log. (closes: #191306)
* Add rudimentary Test::More modules for testing Debbugs::Mime and
Debbugs::Mail. These are currently not called as part of the build
process.
ajt [Thu, 11 Aug 2005 15:57:00 +0000 (07:57 -0800)]
[project @ 2005-08-11 08:57:00 by ajt]
Improvements to the CGIs, in particular for moving towards some sane URLs
that people needn't be ashamed to tell other people. Detailed changes from
the darcs history:
* allow running cgis from command line (./bugreport.cgi bug=12345)
* support for getting params from cookies
* add cookie setter (cookie.cgi, it's pretty lame)
* initial smart url parser (smarturl.cgi)
* make smarturl actually display bug info
* move per-msg mbox handling into bugurl()
* deprecate mboxurl in favour of bugurl(..., "mbox")
* deprecate dlurl in favour of bugurl too
* replace & in urls with ; so it doesn't get munged by urlsanit :-/
* merge all the pkgreport *url() functions
* initial support for internal links in leet urls
* make bug urls leet too
* cope with attachment filenames, single messages as mbox
"Leet" URLs look like "/x/123456" and "/x/package/dpkg" etc.
ajt [Thu, 11 Aug 2005 15:48:09 +0000 (07:48 -0800)]
[project @ 2005-08-11 08:48:09 by ajt]
* remove "use IO::Lines" line from bugreport.cgi
* add "report spam" link to bugreport.cgi
* improve by-*.idx handling to support archive indexes and be slightly faster
* add link to archived bug pages to pkgreport.cgi
cjwatson [Wed, 10 Aug 2005 08:39:45 +0000 (00:39 -0800)]
[project @ 2005-08-10 01:39:45 by cjwatson]
Add an allstates() method, which returns the versionwise state of the bug
at every known version at once. The answer to whether this is more
efficient than just calling buggy() for every version you're interested in
is roughly "it depends".
don [Thu, 4 Aug 2005 11:22:29 +0000 (03:22 -0800)]
[project @ 2005-08-04 04:22:29 by don]
* Wrap the read_log_records call in eval so we output a useful error
message using quitcgi on munged bugs
* Really output boring records if we're only outputing one record
* Add the appropriate meta http-equiv tag to indicate that we're
sending utf-8 so things that save the pages display in an appropriate
charset
don [Mon, 1 Aug 2005 13:47:54 +0000 (05:47 -0800)]
[project @ 2005-08-01 06:47:54 by don]
* Rewrite encode_rfc1522 from scratch with a much saner
implementation which now appropriately deals with non-ASCII trailing
characters.
don [Sun, 31 Jul 2005 11:43:01 +0000 (03:43 -0800)]
[project @ 2005-07-31 04:43:01 by don]
* Add links to cloned bugs (closes: #217960) and forwarded records
that look like urls in the html records output by bugreport.cgi.
* Things that look like urls in message bodies are now linked
(closes: #168962)
don [Sat, 30 Jul 2005 10:22:36 +0000 (02:22 -0800)]
[project @ 2005-07-30 03:22:36 by don]
The "FUNNY FERRIGN THINGIES DON'T WORK" Memorial Commit
* All messages that originate from the BTS and either go to .log
files or out to users are now properly RFC1522 encoded. (closes:
#306068)
* Use Encode.pm to convert from other charsets to UTF8; this
introduces an (as yet unexpressed) dependency on a newish version of
perl. [And decodes more charsets to boot.]
* Stop depending on libunicode-maputf8-perl because of the above.
* Really make encode_rfc1522() do what it is supposed to do without
stripping out spaces and doing other funny things; A patch for
MIME::Words was ripped out and modified to make this work:
http://rt.cpan.org/NoAuth/Bug.html?id=13027
* Decode every single header, not just some of them
don [Fri, 29 Jul 2005 14:05:37 +0000 (06:05 -0800)]
[project @ 2005-07-29 07:05:37 by don]
* Add temporary hack to work around #320406 (which ramified itself in
#204514) [Failures with BIG5 charsets... we now punt]
don [Fri, 29 Jul 2005 11:34:11 +0000 (03:34 -0800)]
[project @ 2005-07-29 04:34:11 by don]
* Add an appropriate X-$gProject-PR-Message: list inactivebug for
archived or non-existant bugs for list related messages.
* Display old severity when changing severity (closes: #196947)
don [Thu, 28 Jul 2005 05:51:25 +0000 (21:51 -0800)]
[project @ 2005-07-27 22:51:25 by don]
* Return to the previous mailbox download default of showing only the
non-boring messages. Boring messages can now be shown by adding
&boring=yes to the url.
don [Tue, 26 Jul 2005 11:30:53 +0000 (03:30 -0800)]
[project @ 2005-07-26 04:30:53 by don]
* Add X-$gProject-PR-{Message,Package,Title} headers for messages
that are sent through the subscription mechanism; these will be used
in crafting the subscription confirmation message.
don [Tue, 26 Jul 2005 06:30:22 +0000 (22:30 -0800)]
[project @ 2005-07-25 23:30:22 by don]
* Change bugreport.cgi to use Debbugs::Log and greately simplify the
process of outputing the bug log.
* All RFC1522 subject lines are decoded, both in the html information
and the message headers. All messages are converted to UTF-8 whereever
possible; all bugreport.cgi pages are now completely in UTF-8 to the
degree possible. (closes: #46848, #238984)
* Add a convert_to_utf8 function to Debbugs::Mime to make the above
possible; abstracts functionality that was already present in the
decode_rfc1522 fucntionality.
* Individual messages can now be downloaded from each bug report
(closes: #95373)
* Uninteresting headers are now hidden by default, can be renabled
with &trim=no (closes: #188561)
[project @ 2005-07-25 09:18:02 by cjwatson]
Add an incremental read_record interface to allow processing records one at
a time rather than having to slurp them all into memory.
don [Mon, 25 Jul 2005 01:42:41 +0000 (17:42 -0800)]
[project @ 2005-07-24 18:42:41 by don]
* Rest of subscription patch that wasn't applied; adds support for
sending messages to affected bugs in service.in (the control bug)
and actually receiving the subscription messages (receive.in)
[project @ 2005-07-24 18:01:52 by cjwatson]
Fix 'found' to clear the done field if addfoundversions() cleared the most
recent version from fixed_versions, rather than trying to guess itself.
[project @ 2005-07-24 17:31:18 by cjwatson]
Improve version canonicalisation when adding found/fixed versions and
attempting to remove any corresponding fixed/found versions respectively.
Try harder to store fully-qualified versions (i.e. with source package
name).
[project @ 2005-07-24 16:11:22 by cjwatson]
Move $gPackageSource fallback from binarytosource() to
makesourceversions(), so that it's possible for applications to detect the
"nothing known about this version" condition.
[project @ 2005-07-24 15:06:54 by cjwatson]
Move binary/source package handling from common.pl to a new
Debbugs::Packages module. This will allow service to do some
binary-to-source mapping for version tracking.
don [Sat, 23 Jul 2005 04:37:31 +0000 (20:37 -0800)]
[project @ 2005-07-22 21:37:31 by don]
* When the spam has already been detected, we should output
$mail->get_pristine as the $munged_mail is not defined yet.
DAMN YOU MISSING 'use strict;'!
don [Fri, 22 Jul 2005 10:49:44 +0000 (02:49 -0800)]
[project @ 2005-07-22 03:49:44 by don]
* Added per bug subscription support to debbugs, which relies on an
external MLM to actually deal with the requests; currently works with eoc
and sends messages to bugnum\@$gListDomain. (closes: #34071) This feature
still needs to be modified slightly so that it can be enabled or
disabled, as not all installs of debbugs will want to run eoc.
don [Fri, 22 Jul 2005 04:24:09 +0000 (20:24 -0800)]
[project @ 2005-07-21 21:24:09 by don]
* Change spamscan.in to work with SA 3.0; This is an incompatible
change with SA 2.6, and as such, we now Suggests: spamassassin
(>=3.0) (closes: #290501)
[project @ 2005-07-18 04:13:21 by cjwatson]
Prefer source architecture for src= queries, but allow other architectures
if source isn't present for a given package. Fixes handling of multi-binary
packages.
don [Sun, 17 Jul 2005 04:13:17 +0000 (20:13 -0800)]
[project @ 2005-07-16 21:13:16 by don]
* Add Makefile rule to actually install apache.conf in examples
* Don't remove the maintainer address if the message was sent by the
maintainer and we're submitting to maintonly (closes: #140061)
* Update changelog for the changes that I've made.
don [Sun, 17 Jul 2005 01:49:31 +0000 (17:49 -0800)]
[project @ 2005-07-16 18:49:31 by don]
* Use uri_escape to escape URI's before sending them back out so the name
of the file doesn't munge the query string. (URI handling really needs to
be improved, but this will work for starters.) [#301606 part 1]
* call decode_rfc1522 on the filename returned so that encoded filenames
get decoded to something "reasonable." [#301606 part 2]
don [Thu, 14 Jul 2005 23:21:50 +0000 (15:21 -0800)]
[project @ 2005-07-14 16:21:50 by don]
* Added apache.conf, which is a sample apache configuration file which
implements the rewrite rules that are currently used on bugs.debian.org.
[When commited this will enable #222264 to be closed]
[project @ 2005-04-20 04:40:18 by cjwatson]
Add note to nonumnosub error that "Your message has been filed under junk
but otherwise ignored", following the form of other messages.
[project @ 2005-04-09 16:21:02 by cjwatson]
Decode RFC1522 mail headers for display in the web interface. bugreport.cgi
and pkgreport.cgi now output UTF-8.
This introduces .summary format version 3, in which all metadata is stored
in UTF-8. Due to a lack of testing and no migration script, this is not yet
the default. When reading/writing earlier format versions, RFC1522 metadata
is automatically decoded/encoded.
cjwatson [Thu, 17 Mar 2005 04:23:21 +0000 (20:23 -0800)]
[project @ 2005-03-16 20:23:20 by cjwatson]
Add Debbugs::MIME::de_rfc1522() function, modified slightly from a patch by
Don Armstrong <don@debian.org>. Decodes RFC1522-style mail header encoding
to UTF-8.
cjwatson [Sat, 28 Aug 2004 06:36:44 +0000 (22:36 -0800)]
[project @ 2004-08-27 23:36:44 by cjwatson]
Update history; note that nCipher allowed me to distribute modifications I
made while I was an employee there.