6905626e24f39feca86cd35948dbb0d43a3b7a62: fix the pending async count logging. also, use the same connection as sync
Andrew Sutherland <asutherland@asutherland.org> - Thu, 30 Oct 2008 21:20:44 -0700 - rev 1000
fix the pending async count logging. also, use the same connection as sync and async. We don't actually use the syncConnection as such anymore, but if we did, we would want to be using the same connection anyways.
b4d077faa9994fd5708fa1c14f42a3189de32678: * add content processing idiom, with a generic quoting processing one.
Andrew Sutherland <asutherland@asutherland.org> - Thu, 30 Oct 2008 15:58:09 -0700 - rev 999
* add content processing idiom, with a generic quoting processing one. * add ability to expose attributes flaged as domExpose via helper method.
035f9d589e162fa6b13d5a5cfbeeb756388599dc: comment out a lot of queryFromQuery-related debug that hopefully isn't needed.
Andrew Sutherland <asutherland@asutherland.org> - Mon, 27 Oct 2008 03:56:46 -0700 - rev 998
comment out a lot of queryFromQuery-related debug that hopefully isn't needed. I'll purge them when I'm confident I don't need to debug this junk again.
1a1437838c7c932dd9c1dde84e0c49d458d8dabf: * autocomplete on message tags
Andrew Sutherland <asutherland@asutherland.org> - Mon, 27 Oct 2008 03:48:52 -0700 - rev 997
* autocomplete on message tags * populate set of known freetags based on bound parameters; avoids ugly need for a contact to have already been loaded for us to know about the tag.
942330424dd7626562381d1a63060e7a07e2e3bb: Changes for our JS message representation made in parallel with my changes for
Andrew Sutherland <asutherland@asutherland.org> - Mon, 27 Oct 2008 01:29:40 -0700 - rev 996
Changes for our JS message representation made in parallel with my changes for the patch on bug 447842. This allows us to just index the text body parts. We should probably have a fallback to transforming the HTML to text if we have no text parts, but baby steps...
4f73d25ae3c1b3bf1f27cfdec534580139263e57: notifications for the identity -> contact -> identity loading were not working
Andrew Sutherland <asutherland@asutherland.org> - Sun, 26 Oct 2008 20:21:11 -0700 - rev 995
notifications for the identity -> contact -> identity loading were not working right; the latter identity was short-circuiting the notification back to its originator (message) rather than going through contact, which resulted in contact never having its _deps resolved. This resulted in invariant violation and unhappiness downstream when that item was later loaded from cache and a resolve-deps attempted to happen, but without the data required.
cd97741b526eabc6472b24271bc76205af69d3ff: * trigger re-indexing of messages on keyword changes
Andrew Sutherland <asutherland@asutherland.org> - Sun, 26 Oct 2008 09:11:21 -0700 - rev 994
* trigger re-indexing of messages on keyword changes * wrap collection notifications in try/catch blocks to protect other code.
0a0218f1f119f0021071cabe7749d800562e4801: ugh, had failed to back out some db attribute def stuff that broke bound
Andrew Sutherland <asutherland@asutherland.org> - Sun, 26 Oct 2008 07:57:36 -0700 - rev 993
ugh, had failed to back out some db attribute def stuff that broke bound parameters.
da619e0bce23fc1cac3ce56aa9dc183ae18a5df5: various fixes involving usesParameter. No one claimed they supported it,
Andrew Sutherland <asutherland@asutherland.org> - Sun, 26 Oct 2008 06:41:12 -0700 - rev 992
various fixes involving usesParameter. No one claimed they supported it, which was just as well, because no one checked it/used it properly either. Tagged contacts now works correctly. (Before it was lumping everyone under the null-parameter id value.)
11ad0ebad23c57a9dd2b33bb9510d57d60ade2be: fix-up autocomplete logic, also resolving some bugs in grokNounItem, plus
Andrew Sutherland <asutherland@asutherland.org> - Sun, 26 Oct 2008 05:44:48 -0700 - rev 991
fix-up autocomplete logic, also resolving some bugs in grokNounItem, plus other minor fixes. Autocompletion of tagged contacts works again now (it was indexing of them that was broken, though.)
16c6a95db57ed5db0a990778ce6a93d64e7c11d9: make fulltext work again
Andrew Sutherland <asutherland@asutherland.org> - Fri, 24 Oct 2008 09:24:15 -0700 - rev 990
make fulltext work again
dde2741d963e2e3bb72e9c74b394d3661f22a8e4: things are working with exptoolbar
Andrew Sutherland <asutherland@asutherland.org> - Fri, 24 Oct 2008 08:30:34 -0700 - rev 989
things are working with exptoolbar
1be012dd374f93311ca7e932d922cffd55f0e961: we can use Cu instead of Component.utils when we define Cu...
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 16:38:36 -0700 - rev 988
we can use Cu instead of Component.utils when we define Cu...
8dd37fc89a60f3b179f9a4638fc0882cee4e6dc7: refactoring status commit; address book indexing glitch oversight is the most
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 11:23:55 -0700 - rev 987
refactoring status commit; address book indexing glitch oversight is the most significant change.
9409c30e08c4d57218442358f4261dac65fe7259: refactoring in a very good state now; expmess seems to work fairly nicely.
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 09:40:05 -0700 - rev 986
refactoring in a very good state now; expmess seems to work fairly nicely.
d01c90782c0c49d13006ff863400654ad8a8d1d5: whoa, bad confusion about the read/write timeout. Do not set it!
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 07:30:10 -0700 - rev 985
whoa, bad confusion about the read/write timeout. Do not set it!
d01202682cfe273a0159412b0512992fd5aa843b: er, and of course, global escaping is important. (to my credit, I remembered
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 04:16:07 -0700 - rev 984
er, and of course, global escaping is important. (to my credit, I remembered to do this without getting bit by it. in fact, I never escaped anything! (the want of testing may not be as much to my credit...))
19f6ee39f3beea5c675574d50759d8eb69e1684f: implement Log4J XMLLayout style logging support in log4moz, hook it up so that
Andrew Sutherland <asutherland@asutherland.org> - Mon, 20 Oct 2008 03:31:38 -0700 - rev 983
implement Log4J XMLLayout style logging support in log4moz, hook it up so that if you have a /tmp/chainsaw.ptr file that contains "hostname:port", it can target the output to said destination. When using Chainsaw, you would create an XMLSocketReceiver (default port 4448), so you could put "localhost:4448" in /tmp/chainsaw.ptr and then party down. It's advisable to delete chainsaw.ptr when you're not using it, but the SocketAppender defaults to only trying to open the connection every 5 seconds, so the performance impact won't be too horrid if you leave it around.
93af40a644fa28e6ced6d14ad4068045780b56f7: Added tag unstable for changeset 8e9fe32109b8
Andrew Sutherland <asutherland@asutherland.org> - Sun, 19 Oct 2008 04:21:49 -0700 - rev 982
Added tag unstable for changeset 8e9fe32109b8
8e9fe32109b8a7232a5844320ca273bf5c3a4528: fix invariant glitch unstable
Andrew Sutherland <asutherland@asutherland.org> - Sun, 19 Oct 2008 04:19:14 -0700 - rev 981
fix invariant glitch
(0) -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 tip