Fri, 10 Apr 2026 07:03:40 +0000 https://bugs.horde.org/ticket/9773 Date column not updated after midnight This is really just a cosmetical issue. But after midnight, yesterday's messages are not updated to only show the date. This would actually be fine, but if you navigate to a mailbox after midnight that hat receiced new messages before midnight, you now have mixed date and time for yesterday's messages. Fri, 01 Apr 2011 23:13:38 +0000 https://bugs.horde.org/ticket/9773#t62991 Changes have been made in Git for this ticket: [mms] Purge browser cache daily in dynamic view; updates time stamps to proper format (Request #9773). 6 files changed, 54 insertions(+), 25 deletions(-) http://git.horde.org/horde-git/-/commit/b56b7893d8fbba53ea6498709809edb65a06f0ba Wed, 23 Nov 2011 01:23:53 +0000 https://bugs.horde.org/ticket/9773#t69003 Changes have been made in Git for this ticket: Fix regressions in search mailboxes due to Ticket #9773 Move UIDVALIDITY check before empty mailbox check, since a UIDVALIDITY failure should cause all viewport information to be reset, regardless of whether the mailbox contains messages. UIDVALIDITY will also fail when the day changes, so metadata creation needs to appear below this line since metadata will be reset in this instance. 2 files changed, 36 insertions(+), 36 deletions(-) http://git.horde.org/horde-git/-/commit/6f435e07cd2a1419a35c1de8f73fa81cc3d22b9c Wed, 30 Nov 2011 22:29:46 +0000 https://bugs.horde.org/ticket/9773#t69194 > [...] > UIDVALIDITY will also fail when the day changes, so metadata creation > needs to appear below this line since metadata will be reset in this > instance. doesn't work for me (see #12940) with IMP 6.1.6, Horde_Imap_Client 2.17.1. IMAP server is Cyrus 2.4.17. Do I read it correctly, that I should trace the first IMAP call after a (server) date change to see what is going wrong? Thu, 30 Jan 2014 09:22:58 +0000 https://bugs.horde.org/ticket/9773#t82189 >> [...] >> UIDVALIDITY will also fail when the day changes, so metadata creation >> needs to appear below this line since metadata will be reset in this >> instance. > > doesn't work for me (see #12940) with IMP 6.1.6, Horde_Imap_Client > 2.17.1. IMAP server is Cyrus 2.4.17. Verified working. The browser cache ID for a mailbox is appended with a string of the form "D###" where ### indicates the current day of the year. This cache id is guaranteed to change after midnight, at which point the browser empties its message cache and fetches a new mailbox slice from the server (with the updated dates). Thu, 30 Jan 2014 19:44:30 +0000 https://bugs.horde.org/ticket/9773#t82200 >> doesn't work for me (see #12940) with IMP 6.1.6, Horde_Imap_Client >> 2.17.1. IMAP server is Cyrus 2.4.17. > > Verified working. The browser cache ID for a mailbox is appended > with a string of the form "D###" where ### indicates the current day > of the year. This cache id is guaranteed to change after midnight, > at which point the browser empties its message cache and fetches a > new mailbox slice from the server (with the updated dates). attached a "proof-of-notworking4me" image from today - you can clearly see yesterday's messages with time-only timestamps. It's from a session run over-night. According to "pear list-ugrades", the server is running the latest modules as of today. I had run a "horde-clear-cache" after the recent upgrade (which, iirc, included an update to some caching module). But these symptoms have been here since forever, not just after the recent upgrade. And in case it might matter, we're running Kolab backends here. How may I trace this "browser cache ID" thing - I ran a network trace last night, but unfortunately forgot to tune down the encryption methods... wireshark wouldn't decode the SSL traffic despite having the private server key :[ Is that the right traffic to look at (so I'd repeat the test next night) or is this something that acts either client-side only or server-side only? This still is low prio to me, but since I've already started looking into this, I offer to take it further... Fri, 31 Jan 2014 09:41:22 +0000 https://bugs.horde.org/ticket/9773#t82208 > How may I trace this "browser cache ID" thing Firebug is probably best. The cacheid is sent in the imp:viewport task object contained in the JSON responses to various mailbox actions. It looks something like this: "VTIxMTkyLFYxMjU1Njg1MzM4LEg1MDEwNQ==|9|0|D34" Tue, 04 Feb 2014 07:08:15 +0000 https://bugs.horde.org/ticket/9773#t82250 >> How may I trace this "browser cache ID" thing > > Firebug is probably best. The cacheid is sent in the imp:viewport > task object contained in the JSON responses to various mailbox > actions. > > It looks something like this: "VTIxMTkyLFYxMjU1Njg1MzM4LEg1MDEwNQ==|9|0|D34" Finally I got around to testing this with Firebug - I'm sorry for the delay. After a fresh page reload (I had logged in previously, selected my INBOX for display and selected your latest Horde bug reply - then I hit the browser's "reload" icon), I do see the mentioned imp:viewport with it's cache id as part of the imp/dynamicInit POST request: cacheid "VTM5MDAyLFYxMjU2NTc2OTUwLEgzMjY5Ng==|1|1|D42" From then on, imp/poll POST requests (besides the toolbar updates) are the only requests I see, either automatically or by pressing IMP's "update" link. Their response JSON only has imp:poll and imp:quota objects in "tasks" - neither contain a cacheid. The first imp/poll after midnight then actually has a viewport object with a cacheid element: cacheid "VTM5MDAyLFYxMjU2NTc2OTUwLEgzMjY5Ng==|1|1|D43" This is indeed different from the cacheid reported initially, but as observed previously, the time stamps in the message list remain as "time-only" even after the imp/poll requests after midnight. Further requests again have no viewport object in their responses. Seems like the cache id change goes by unnoticed? Regards, Jens The currently installed version is imp 6.1.6 from PEAR. Wed, 12 Feb 2014 23:13:48 +0000 https://bugs.horde.org/ticket/9773#t82370 Here's the JSON from the first poll after mid-night. I see no other requests but the imp/poll and imp/toolbarUpdate at regular intervals (unless, of course, I'd manually interact with IMP - which I didn't do during this test). {"response":true,"tasks":{"imp:viewport":{"label":"Posteingang","totalrows":38,"rowlist_reset":true,"rowlist":{"39001":1,"38989":2,"38965":3,"38963":4,"38904":5,"38903":6,"38785":7,"38722":8,"38632":9,"38571":10,"38569":11,"38560":12,"38535":13,"38297":14,"38187":15,"37993":16,"37893":17,"37878":18,"37871":19,"37826":20,"37684":21,"37412":22,"37137":23,"35362":24,"35255":25,"35109":26,"34242":27,"33979":28,"33926":29,"31722":30,"30608":31,"28412":32,"23231":33,"18485":34,"11078":35,"38537":36,"9889":37,"6262":38},"data":[],"cacheid":"VTM5MDAyLFYxMjU2NTc2OTUwLEgzMjY5Ng==|1|1|D43","view":"SU5CT1g"},"imp:poll":{"SU5CT1g":0,"SU5CT1gvQWJsYWdlL21haWxpbmctbGlzdHM":3,"SU5CT1gvQWJsYWdlL05ERSBBRy9LdW5kZW4vT0ZEIE5pZWRlcnNhY2hzZW4vTlotUHJvYmxlbWU":0,"SU5CT1gvQWJsYWdlL1ByaXZhdC9GaW5hbnplbi9EZXV0c2NoZSBCYW5r":0,"SU5CT1gvanVuaw":0,"dXNlci9hbW96ZHplbg":1},"imp:quota":{"m":"Keine Begrenzung","p":0,"l":""}}} Further imp/poll again have only the "short version" answer: {"response":true,"tasks":{"imp:poll":{"SU5CT1g":0,"SU5CT1gvQWJsYWdlL21haWxpbmctbGlzdHM":3,"SU5CT1gvQWJsYWdlL05ERSBBRy9LdW5kZW4vT0ZEIE5pZWRlcnNhY2hzZW4vTlotUHJvYmxlbWU":0,"SU5CT1gvQWJsYWdlL1ByaXZhdC9GaW5hbnplbi9EZXV0c2NoZSBCYW5r":0,"SU5CT1gvanVuaw":0,"dXNlci9hbW96ZHplbg":1},"imp:quota":{"m":"Keine Begrenzung","p":0,"l":""}}} Wed, 12 Feb 2014 23:20:23 +0000 https://bugs.horde.org/ticket/9773#t82371 The key to this might be the following observation: The data element of the viewport object is empty, in the first poll response after midnight. A future poll, which reports a new message, does have the viewport object, but with a data element containing the new entry's field values: {"response":true,"tasks":{"imp:viewport":{"label":"Posteingang","totalrows":39,"rowlist_reset":true,"rowlist":{"39002":1,"39001":2,"38989":3,"38965":4,"38963":5,"38904":6,"38903":7,"38785":8,"38722":9,"38632":10,"38571":11,"38569":12,"38560":13,"38535":14,"38297":15,"38187":16,"37993":17,"37893":18,"37878":19,"37871":20,"37826":21,"37684":22,"37412":23,"37137":24,"35362":25,"35255":26,"35109":27,"34242":28,"33979":29,"33926":30,"31722":31,"30608":32,"28412":33,"23231":34,"18485":35,"11078":36,"38537":37,"9889":38,"6262":39},"data":{"39002":{"flag":["personal","unseen"],"size":"5 KB","date":"00:13:48","from":"noreply@bugs.horde.org","subject":"[Tickets #9773] Re: Date column not updated after midnight"}},"cacheid":"VTM5MDAzLFYxMjU2NTc2OTUwLEgzMjY5Nw==|1|1|D43","view":"SU5CT1g"},"imp:poll":{"SU5CT1g":1,"SU5CT1gvQWJsYWdlL21haWxpbmctbGlzdHM":3,"SU5CT1gvQWJsYWdlL05ERSBBRy9LdW5kZW4vT0ZEIE5pZWRlcnNhY2hzZW4vTlotUHJvYmxlbWU":0,"SU5CT1gvQWJsYWdlL1ByaXZhdC9GaW5hbnplbi9EZXV0c2NoZSBCYW5r":0,"SU5CT1gvanVuaw":0,"dXNlci9hbW96ZHplbg":1},"imp:quota":{"m":"Keine Begrenzung","p":0,"l":""}}} I tend to assume that the browser's message list is not replaced (with a version with updated time stamps) because no new entries' data was sent with that "after midnight" poll response. Wed, 12 Feb 2014 23:34:29 +0000 https://bugs.horde.org/ticket/9773#t82372 An issue might be that you are running 6.1 and I am running 6.2. Comparing the two, I wonder if this is causing problems (i.e. make this change on your file): diff --git a/imp/lib/Ajax/Application/ListMessages.php b/imp/lib/Ajax/Application/ListMessages.php index 00cd4aa..a084baa 100644 --- a/imp/lib/Ajax/Application/ListMessages.php +++ b/imp/lib/Ajax/Application/ListMessages.php @@ -141,7 +141,7 @@ class IMP_Ajax_Application_ListMessages $parsed = $imp_imap->parseCacheId($args['cacheid']); $uid_expire = true; - if (!$parsed['date'] != date('z')) { + if ($parsed['date'] == date('z')) { try { $imp_imap->sync($mbox, $parsed['token'], array( 'criteria' => Horde_Imap_Client::SYNC_UIDVALIDITY Fri, 14 Feb 2014 03:42:14 +0000 https://bugs.horde.org/ticket/9773#t82373 > An issue might be that you are running 6.1 and I am running 6.2. > > Comparing the two, I wonder if this is causing problems (i.e. make > this change on your file): > [...] thank you for your suggestion - I have updated the file and will report back the results! Regards, Jens Fri, 14 Feb 2014 12:27:56 +0000 https://bugs.horde.org/ticket/9773#t82377 > An issue might be that you are running 6.1 and I am running 6.2. > > Comparing the two, I wonder if this is causing problems (i.e. make > this change on your file): Confirmed :) After changing that single line, the time/date display switches correctly with the first poll after midnight. Regards, Jens Sat, 15 Feb 2014 14:23:52 +0000 https://bugs.horde.org/ticket/9773#t82382