Art Cancro [Mon, 6 Apr 2009 14:54:13 +0000 (14:54 +0000)]
* Set zapped_ok to 1 on nearly all calls to imap_grabroom(), because the IMAP commands in question are permitted to operate with unsubscribed (zapped) rooms.
Art Cancro [Tue, 31 Mar 2009 18:55:28 +0000 (18:55 +0000)]
* Oops ... reply/replyquoted/replyall semantics were still buggy. Restored the correct 'all replies are quoted' behavior for localuser-to-localuser messages in private mail.
Art Cancro [Mon, 30 Mar 2009 16:49:04 +0000 (16:49 +0000)]
* Backed out r7276 through r7278. This is too large a change to make during a feature freeze; it is the wrong approach; and it also produced crashes and corrupt data on my test system during evaluation (even after a make distclean). We will revisit this after the release.
Art Cancro [Fri, 27 Mar 2009 17:02:23 +0000 (17:02 +0000)]
* Fixed a multiselect bug in the mailbox view. Ctrl-click was selecting a message even if it was already selected. Amended this code to make it toggle the current state instead.
* make some pointers const...
* trace whether we were shut down regular, or we crashed/were killed/...
* send an aide message if we unconditionaly exit.
Art Cancro [Fri, 20 Mar 2009 16:20:25 +0000 (16:20 +0000)]
* Events and tasks which are missing a SUMMARY property are now displayed as 'Untitled Event' or 'Untitled Task' instead of being omitted. (I saw those two strings already in the code, so hopefully they've already been localized.)
Art Cancro [Fri, 20 Mar 2009 02:39:55 +0000 (02:39 +0000)]
* In the JS instant messenger window, encode all outgoing messages using encodeURIComponent() before handing them off to Prototype for an AJAX call. This allows non-ASCII characters to be sent without causing an error in the application. Enjoy your ũmläȕŧŝ ☺
Art Cancro [Thu, 19 Mar 2009 17:10:32 +0000 (17:10 +0000)]
* Changed ConditionalIsRoomtype() to use a full string compare instead of one character in order to improve the readability and maintainability of this code. (Also, the character being compared was the wrong one, so this also eliminates a bug.)
Art Cancro [Thu, 19 Mar 2009 04:32:15 +0000 (04:32 +0000)]
* When using the javascript instant messenger screen, transmit outgoing messages to webcit using a real webcit command (new command: ajax_send_instant_message) instead of the ajax_server_cmd framework. This gives us string unescaping for free, so we now escape all strings before transmitting them. I was hoping this would fix the problem with sending non-ASCII characters through the instant messenger, but it didn't. This will probably still be part of the solution though.
* add conditional to check room-types
* use hide 'delete' button if we're not allowed to
* check for room type BBS whether we show the one or the other button group
Art Cancro [Fri, 13 Mar 2009 18:14:44 +0000 (18:14 +0000)]
* Be more intelligent about the way we load vCard fields into the on-screen editor. Properties and parameters are now treated as separate variables. Properties are matched strictly (such as 'email') but parameters are handled with a fuzzy match. For example, we identify an 'email' property as an Internet address if the string 'internet' appears anywhere within the parameter string. This will allow it to match on 'email;internet' as well as 'email;type=internet' and 'email;type=internet,pref' or even 'email;not_the_internet'. There are so many different variations out there that we can't count on a strict match.
Art Cancro [Fri, 13 Mar 2009 02:19:05 +0000 (02:19 +0000)]
* Minor overhaul of modules/serv_vcard.c
Be less strict about properties and parameters. For example, when searching for Internet email addresses, it no longer has to be 'email;internet', it will also accept things like 'email;type=internet' or 'email;type=internet,pref' or even 'email;foo=bar,type=internet'.
Note that this is strictly server side stuff; we have a lot of work to do on the webcit side.
Art Cancro [Thu, 12 Mar 2009 02:49:59 +0000 (02:49 +0000)]
* Fixed problems related to saving vCards in config rooms. For starters, we were looking at the room name contained in the message rather than the name of the session's current room, which might actually be correct 100% of the time, but it's still the wrong place to look. We also needed to add code to distinguish between saving your own vCard and administratively saving someone else's vCard. We only reload the cs_inet_* caches when it's our own.
Art Cancro [Wed, 11 Mar 2009 16:57:36 +0000 (16:57 +0000)]
* Implemented the first of several steps required to fix the bug that causes mailing list bounces to be misdirected. This commit adds a new field in the SMTP delivery instructions to force a particular envelope sender.
Art Cancro [Mon, 9 Mar 2009 16:45:21 +0000 (16:45 +0000)]
* Removed references to Trebuchet and Verdana fonts; go with 'sans-serif' instead. This will cause most of WebCit to be rendered in the user's preferred Sans Serif font, which will hopefully fix any remaining weird font issues.
Art Cancro [Mon, 9 Mar 2009 04:00:43 +0000 (04:00 +0000)]
* Removed the defunct METADATA parameters to our unfinished LIST-EXTENDED implementation.
* Rewrote any comments in the IMAP module sources that may have been copied directly out of various RFC's, in order to satisfy the license police.
Art Cancro [Thu, 5 Mar 2009 21:28:24 +0000 (21:28 +0000)]
* Corrected the 'walking timestamp' problem in the Tasks view by converting to UTC while loading a task into WebCit. This is not a problem for Tasks because there are no calculations or recurrences being performed on task dates. Thanks to feivel for reporting this.
Art Cancro [Tue, 3 Mar 2009 03:42:29 +0000 (03:42 +0000)]
* When converting cid: URL's in to /webcit/mimepart... URL's, prefix the inserted 'src=' with an additional space. Sometimes it's needed when other parameters such as alt text appear prior to the src= declaration, and in cases where it's not needed, the extra space won't hurt anything.
Art Cancro [Mon, 2 Mar 2009 04:11:04 +0000 (04:11 +0000)]
* Added some comments next to portions of the code which are obsoleted by the final form of RFC 5464 (IMAP METADATA extension). The dependency on LIST-EXTENDED has been removed, and the Bynari Connector never needed it in the first place. After the code freeze is over we will remove the portions of our unfinished LIST-EXTENDED implementation which were to provide metadata, *or* we may simply remove the unfinished LIST-EXTENDED implementation altogether.