Wednesday, 2010-06-23

../irclogs/#mantishelp.2010-06-23.log
--- scribe started ---00:00
dhx_mnuclear_eclipse: wrong commit?00:16
nuclear_eclipsesorry, commit 886dccd9a467c892a165e7a755a6b13d0d8ed36500:19
dhx_mah yep that's on the TODO list to fix00:29
dhx_mI'm trying to decide at the moment how to handle email notifications on bug_update.php00:29
dhx_mit's not easy :(00:30
dhx_msay someone wants to me notified on change of handler00:30
dhx_myet they don't want to know when the status changes00:30
nuclear_eclipseyeah, good luck on that one00:30
dhx_mlol I think I'll leave it how it was (mostly) for now00:30
dhx_mie. precedence of rules... handler > status > update00:31
dhx_mwhich is rather idiotic at the moment00:31
* nuclear_eclipse thinks you opened a gigantic can of worms00:32
dhx_mbut then so too are the email notification settings where half of them don't work (and can't work in many cases)00:32
dhx_mlol00:32
dhx_mI have been thinking for a long time that notifications should tap into the event system00:32
nuclear_eclipseyeah, I would completely agree -- it's just a matter of "how do you do that without pissing off > 50% of current users"00:33
dhx_mso a bundled email plugin responds to EVENT_BUG_UPDATE_POST_COMMIT to send out emails of the changes made00:33
dhx_mwell given that email doesn't really work at the moment... I think they're probably already quite annoyed :p00:33
dhx_myou might be interested in #1196700:34
nuclear_eclipsewell, it *does* work, just not the way people expect it to work out of the box; and for anyone who's taken the time to learn and config mantis to do exactly what they want, just about *anything* will be abreaking change for them00:34
dhx_myep00:35
dhx_mRE #11967 we need to add an event prior to committing a bug update to the database (as well as after)00:35
dhx_mI was thinking of having EVENT_BUG_UPDATE_PRE_COMMIT and EVENT_BUG_UPDATE_POST_COMMIT to replace the current EVENT_BUG_UPDATE event00:35
dhx_mas my new bug_update.php does all validation before making any changes to the database00:36
dhx_mtherefore database commits are "guaranteed" (race conditions unavoidable though)00:36
dhx_mwhereas before it was possible for some data to be committed (the status change for instance) yet an access check would fail later on in the script for adding a bug note... hence you only get 50% of the changes committed, the other 50% missing00:37
nuclear_eclipsenot a fan of those names, but I agree that the events around bug update should be consistent with events for bug report, where REPORT_BUG_DATA allows pre-processing, and standard REPORT_BUG allows post processing00:40
dhx_mok I'll try that approach instead (will need documentation updates too though)00:41
dhx_mI think I prefer VALIDATE to DATA but it's better to be consistent :)00:42
nuclear_eclipsewell, it's more than just validation that it allows00:44
nuclear_eclipsethe point is that REPORT_BUG_DATA allows plugins to actually modify the bug data before the database is updated00:45
dhx_mtrue, it can add/change it too00:45
dhx_myep00:45
nuclear_eclipseit the case of some proprietary plugins  I wrote, we used that to auto-assign issues based on custom field values, as well as updating custom fields based on other changes, like automatically filling in a custom field whenever changing to a certain status00:47
nuclear_eclipse(just for a bit of context)00:47
dhx_mok cool :)01:00
dhx_mcan we also change the parameters of the event where needed?01:01
dhx_mI think EVENT_BUG_UPDATE_DATA should send the existing and updated bug objects to the event handler01:01
dhx_mand EVENT_BUG_UPDATE should receive the updated bug object01:02
dhx_mdoes that sound reasonable?01:02
dhx_mthe catch is we only do a shallow copy (PHP's clone) of the BugData object01:03
dhx_mwhich at the moment is OK... but in the future if BugData becomes more complex by holding other objects will require a __clone() method on BugData01:03
dhx_mactually... bug_update only modifies primitives anyway, so I can't see it being an issue01:04
nuclear_eclipsedhx_m: just as long as you document any changes made to the events so that it's obvious in what version of mantis it changed, I'm fine with that01:05
dhx_myep01:05
nuclear_eclipseand by document changes, I mean modify the docbook event reference so that it mentions that behavior was changed in version 1.x.x01:06
dhx_mok01:09
dhx_m1.2.2?01:09
nuclear_eclipsethat's the problem, I don't like major event changes for minor versions, but I agree that it needs to change... =/01:09
nuclear_eclipseso yeah01:10
nuclear_eclipsego ahead01:10
dhx_mbtw I think I know what we need to do with the new email notification system... create a cache (in the database for now, looking towards memcache support in the future) of every change a user should be notified of01:10
dhx_mand have a cron job (or similar) check once every few minutes to collate those changes into a single email01:10
dhx_mso if someone updates the handler of a bug at 03:45 and then updates the resolution at 03:46... the user will receive an email collating these changes at 03:46+5min as long as no other changes have occurred in the mean time01:11
daryndhx_m: i like the sound of that01:12
nuclear_eclipsethe problem is that it requires some sort of cronjob01:13
darynyeah...anything is better than what we get now01:13
darynwell...maybe not anything.01:14
nuclear_eclipseright, but Windows servers don't have cronjobs, and I don't like forcing people to set up a cronjob just to deal with email01:14
dhx_mI could have a look at Mediawiki's work queues too01:18
dhx_mI gather those work by running a few items from the work queue on each page request01:18
nuclear_eclipsedhx_m: I still like the idea though -- perhaps it could just work the same as right now, where it does potentially fancy stuff, but then just flushes out all the emails at the end of the page load01:18
dhx_myep01:18
dhx_mmake the timer 0 in other words01:18
dhx_m(in the event there is no cron job)01:19
nuclear_eclipseyeah, just keep using the same config option to tell it whether to rely on a cronjob or not01:19
nuclear_eclipsewhat you're talking about would also facilitate the much-requested diff-style emails, too01:19
dhx_mdaryn: btw hi :)01:20
darynhello01:20
dhx_mthat's a nice side effect01:20
dhx_mone of the main problems I had when using MantisBT last year for a University project was people receiving too many emails01:21
dhx_mso they became less relevant01:21
dhx_mhence why I suggest allowing 5-10min of no changes to a bug before sending a digest01:21
nuclear_eclipselol, and the biggest complaint when I set up mantis at my old job was not receiving *enough* email :P01:21
dhx_mlol01:21
dhx_mwell yeah that too... it's broken :p01:21
daryneveryone at our work pretty much turns them off and uses rss feeds01:21
dhx_mit's either broken and emails aren't sent... or broken and too many emails are sent01:21
nuclear_eclipsedhx_m: I'd also caution to err on the side of sending an email sooner rather than later01:22
nuclear_eclipsepeople get anxious if they don't get an email immediately after something happens01:22
dhx_mthat's true... I'd guess it'd be better left as a user customisable setting01:23
dhx_m"minimum time between notifications on the same bug" or something similar01:23
nuclear_eclipseeither way, I'd be more than happy to see the current set of notification options and configuration values get obliterated01:23
dhx_mlol01:24
dhx_mI'm not sure if we should use the event system for notifications or introduce a built-in notification system that plugins can hook into01:24
nuclear_eclipsewe just need to make sure there's *some way* of migrating old behaviors to something close enough in the new system, or else you'll hear a *lot* of complaints and bug reports01:24
dhx_myep01:25
dhx_m[master 43de33b] Refactor bug_update to fix multiple bugs01:25
dhx_m 1 files changed, 405 insertions(+), 254 deletions(-)01:25
dhx_m rewrite bug_update.php (76%)01:25
nuclear_eclipsedhx_m: I'd be more inclined to create a fresh notification API, and allow plugins to hook notification events to enable new notification methods other than email01:25
dhx_myep I was heading that direction too01:26
dhx_mso MantisBT core has some understanding of notifications01:26
daryni concur01:26
darynnot that it matters, heh01:26
dhx_mI haven't used git in months... now I have to remember it all heh01:27
daryneither of you looked at filters yet?01:27
nuclear_eclipselol, I've been a bit too busy chasing david's bugs :P01:28
daryndang it01:29
dhx_mhow was I meant to know the first element in the status enum list was the default (it's not documented at all) :p01:29
dhx_mand no clues as to that being the case hehe01:29
nuclear_eclipseall my mantis time is filled atm with a todo list of stuff I need to accomplish due to a project getting ready for a public release01:29
nuclear_eclipseit was "obvious", duh!01:30
nuclear_eclipse:P01:30
dhx_mlol01:30
daryni have a class...MantisBugFilter that basically replaces filter api...however, i want to abstract it  more and create a class that I can build other filters for that aren't necessarily querying the bug table01:32
daryni need a name...01:32
nuclear_eclipseMantisFilter? :P01:32
darynthat would be great except thats what you called the abstract field class01:33
nuclear_eclipseno, that's what I called the abstract filter class...01:33
dhx_mwith git rebase... when you squash a commit into a previous one... which commit message is used?01:33
dhx_mI guess it lets you edit it anyway01:33
nuclear_eclipseboth01:33
darynbut it's for individual fields...01:33
darynMantisBugFilter is a collection of all the fields in the filter01:34
nuclear_eclipsedaryn: I'm confused what you mean -- MantisFilter is what abstracts the ability to a) show a set of inputs to control a filter query, and b) generate a query for the given inputs01:35
dhx_mGRRR I squashed the wrong things01:35
nuclear_eclipsedhx_m: git rebase --abort01:36
dhx_malready done01:36
nuclear_eclipseor check `git reflog` for the old tip's commit hash, and do `git reset --hard <ref>`01:36
dhx_mah nice, thanks :)01:37
darynnuclear_eclipse: help me understand... a) show a set of inputs to control a filter query...01:38
darynthis is for one field per instance, correct?01:38
nuclear_eclipsewell, one input element, if that's what you mean01:39
darynso...if it were applied to standard fields, status would be one instance, priority a different instance, etc.01:39
darynyes01:39
nuclear_eclipseright01:39
dhx_mdaryn: I'll commit these changes and backport them... then I'll have a look at filters :)01:39
darynok...so i'm talking about a collection of all the input elements on the 'filter'01:40
dhx_mprepare for the worst... :p01:40
darynto me the filter is the whole thing01:40
dhx_m...01:40
CIA-22Mantisbt: hickseydr * r4c7915236c46 /core/bug_api.php: Issue #12096: Add reporter and handler to monitor list on duplicate01:40
CIA-22Mantisbt: hickseydr * recc102ed24be /core/constant_inc.php: Fix #12094: ERROR_BUG_READ_ONLY_ACTION_DENIED is not obsoleted01:40
CIA-22Mantisbt: hickseydr * r69c55f958680 /core/bug_api.php: Fix #12095: bug_monitor_copy() should check users exist01:40
CIA-22Mantisbt: hickseydr * r035a1302792a /bug_update.php: Refactor bug_update to fix multiple bugs01:40
CIA-22Mantisbt: hickseydr * rfdaef4f8a6cc / (core/constant_inc.php lang/strings_english.txt): Add new error ERROR_BUG_RESOLVE_DEPENDANTS_BLOCKING01:40
dhx_mthere we go01:40
darynbut that is semantics01:40
dhx_mI'd use the term "rule" perhaps?01:41
dhx_ma filter consists of multiple rules01:42
daryndhx_m: i hope that 'prepare for the worst' isn't directed towards filters!01:42
darynperhaps01:42
dhx_mit wasn't sorry (my bad timing)... it was aimed at those commits01:42
daryn:)01:42
daryndhx_m i don't suppose you've looked at filters either...01:43
daryn:P01:43
dhx_mnot yet sorry, been busy with those changes for 1.2.201:43
daryni want to abstract more but i'd like some feedback01:43
darynya, cool01:44
dhx_mI really hate maintaining 2 branches01:44
nuclear_eclipsedhx_m: then don't worry about it,01:44
dhx_mI can see why the Linux Kernel went with their approach :)01:44
nuclear_eclipsewe can just push 1.3.0 in a couple weeks and call it a 6 mo release cycle :P01:45
dhx_mlol really? I'm for that haha01:45
darynlet's push filters and break it good!01:45
dhx_mquick, get the beta out today!01:45
darynbtw...i took the MantisTemplate branch and made a new plugin with it to replace phptal plugin01:45
nuclear_eclipseso dhx_m, wanna learn how to build a release? :P01:46
darynnot pushed yet01:46
dhx_mnuclear_eclipse: sure :)01:46
dhx_mactually I don't have a non-expired GPG key at the moment01:46
dhx_mstill TODO pending my new email address01:46
nuclear_eclipsewell, I can just do the tag myself01:47
dhx_mthanks for the config from before ;)01:47
dhx_mI don't mind posting/etc01:47
nuclear_eclipsestep 1) figure out everything noteworthy that's changed between 1.2.x and 1.3.0-dev01:48
nuclear_eclipsestep 2) write up release notes in doc/RELEASE with a high level overview of those changes, and remove the old 1.2.x release notes01:48
dhx_mRELEASE NOTES: blame daryn :p01:49
nuclear_eclipselol01:49
darynwhaaaa?01:49
dhx_mactually blame me for the new bug_update() :p01:49
dhx_mfilters ;)01:49
dhx_mdaryn: do I need remotes/daryn/filter-constants01:51
darynwe should probably get the Ajax call on bug report removed first...it breaks collapse js01:51
darynno01:51
darynfilters01:51
nuclear_eclipsestep 3) commit the new release notes, and update the codebase to reflect the next designated release version01:51
nuclear_eclipsein this case, 1.3.0-beta1 ?01:52
daryngoing straight to beta, eh?01:52
nuclear_eclipsemeh, it's not like mantis doesn't have a giant base of stable code...01:53
nuclear_eclipseIMO, alpha is generally reserved for completely new codebases, or times when a huge portion of the release has been untested01:53
dhx_mfilters? :p01:54
darynyeah...01:54
darynuh huh01:54
* nuclear_eclipse pulls up github to look at filters01:54
dhx_mbtw 1.2.0 was only 3 months ago01:54
nuclear_eclipse?01:55
nuclear_eclipsereally?01:55
dhx_mso perhaps we should make it a quarterly release01:55
dhx_mMon, 22 Feb 2010 20:21:46 +0000 (15:21 -0500)01:55
nuclear_eclipsejreese@mach ~/workspace/mantisbt ±master-1.2.x » git show release-1.2.001:55
nuclear_eclipsetag release-1.2.001:55
nuclear_eclipseTagger: John Reese <jreese@leetcode.net>01:55
nuclear_eclipseDate:   Mon Feb 22 15:29:36 2010 -050001:55
darynthis will blow everyone's mind after the last release01:55
nuclear_eclipsethat's more than 3 months ago..01:55
dhx_mlol01:55
dhx_m2 years from 1.1.x to 1.2.x and 3 months to 1.3.x01:55
dhx_mindeed it is01:55
darynbarely01:56
dhx_mhttp://git.mantisbt.org/?p=mantisbt.git;a=summary must be wrong01:56
nuclear_eclipsewell, in all truthfulness, I'll giwe it some slack since it's exactly four months, but if it still says three tomorrow...01:57
dhx_mah I see :)01:57
nuclear_eclipseeither way, two months from beta to rc to final seems reasonable...01:57
dhx_mheadline features would be... filters?01:57
darynthen 1.4 templates in 3 more months?01:58
daryn:D01:58
dhx_mhaha get busy :p01:58
nuclear_eclipsedaryn: sounds good to me :P01:58
dhx_mI'm up for that, thanks for offering to do all the work too :)01:58
nuclear_eclipsehttp://www.mantisbt.org/bugs/changelog_page.php?version_id=10101:58
dhx_mah so I did make a few minor user noticeable changes01:59
nuclear_eclipsedhx_m: seems like the highlight is "lots of small, but progressive changes"01:59
dhx_mfancy new page footer too01:59
darynif we're going to put filters in i need to add a  bunch of filter bugs that will be fixed01:59
dhx_mlol02:00
nuclear_eclipsedhx_m: I can name a few important fixes for 1.3 though02:00
dhx_mby how many megabytes does nuclear_eclipse need to expand the mantisbt.org database server by to accommodate these bugs?02:00
nuclear_eclipsewell, before an official 1.302:00
dhx_myep02:01
darynno, i mean target exisitng bugs for the release and mark them fixed02:01
dhx_m*cough* that preselection of next status feature I blogged about02:01
darynof course the other might be true also...02:01
dhx_mah yep02:01
nuclear_eclipsea) version inheritence is incomplete and has a few bad side effects, like roadmap/filter handling, etc02:01
nuclear_eclipseb) new random/crypto stuff is broken by default02:02
dhx_mit's fixed now02:02
nuclear_eclipseah02:02
dhx_mie. Windows is made to be insecure as there is no other working option02:02
dhx_mit uses the weak pseudo-random number generator02:03
dhx_moh it's not fixed02:05
nuclear_eclipseI didn't think so :P02:05
dhx_m(for Windows)02:05
nuclear_eclipsec) we need to document changes that plugin authors need to be aware of when making plugins compatible with 1.302:06
nuclear_eclipsed) rewrite admin guide :P02:06
darynfilters will affect that too02:06
dhx_mnuclear_eclipse: db_get_table hah02:07
nuclear_eclipse;)02:07
darynyeah...02:07
nuclear_eclipsedaryn: filters is branched from master, right?02:09
darynyes02:09
darynalthough...i may not have your changes from today in it. don't remember02:09
nuclear_eclipsek, gonna pop a diff onto my phone so I can review it later02:09
daryncool. thx02:10
nuclear_eclipsetsk tsk daryn, git is highlighting all your bad whitespacing... :P02:11
darynyeah...i try to remove it all but you know i missed some02:11
nuclear_eclipse`git diff master...daryn/filters` highlights bad spacing in red for me02:12
darynit does for me too but sometimes i get in a hurry02:12
* nuclear_eclipse wags his finger at daryn02:12
nuclear_eclipseanywho, bedtime for me02:12
daryni usually check all that and clean it up before pushing but was in a hurry for you guys02:13
darynof course it didn't help any02:13
daryn:P02:13
nuclear_eclipsedhx_m: you going to be around in ~10hours?02:13
dhx_myep02:13
nuclear_eclipsek, I'll ping you when I get to work02:13
dhx_mwill try and have lots of bugs fixed by then02:13
dhx_mok, cya then :)02:13
nuclear_eclipseawesome02:14
nuclear_eclipsecheers02:14
dhx_m'night02:14
daryngnite02:15
daryndhx_m you in australia?02:15
dhx_mdaryn: yep02:15
darynmy wife tells me we're going there next year.  her friend is getting married there02:16
dhx_mjust be careful of the sharks, world's most poisonous snakes, dropbears, poisonous spiders, lethal jellyfish, etc :p02:17
daryn:)02:17
daryndon't say snakes...she might change her mind02:17
dhx_mhah02:17
dhx_mI even forgot crocodiles from that list duh :p02:18
darynhehe02:18
dhx_mare you using Windows at the moment?02:22
daryn_no...linux only02:27
CIA-22Mantisbt: hickseydr * r3436d148684f /core/crypto_api.php: Use insecure built-in PRNG on Windows (no other options available)02:28
dhx_myep same here, I'll ask paul later to check the latest version of PHP 5.3 on Windows with openssl_random_pseudo_bytes()02:28
daryndhx_m any chance when reworking notifications you can make it so a user can completely remove notification regarding a specific bug even if they have written  bugnotes?02:53
dhx_mthat sounds to me like the monitor list should act as both a whitelist and blacklist02:54
darynmaybe change it so when adding a note they are added to the monitor list and only use that for sending emails02:54
dhx_malso while notifications are high priority... so too is theming :)02:54
darynyeah...02:54
dhx_mpersonally I'd like to focus on templating/theming for the next MantisBT release02:55
dhx_mas I think it's something that may interest developers in helping out the MantisBT project02:55
dhx_mit's easier work writing HTML templates than it is rewriting core parts of Mantis :)02:55
CIA-22Mantisbt: hickseydr * r373d998a1aab /bug_update_advanced_page.php: Issue #12097: Remove usage of update_mode parameter03:25
CIA-22Mantisbt: hickseydr * r981621614d60 / (43 files in 4 dirs): Fix #12085: Remove $g_allow_close_immediately03:25
socorropchello good evening. I've recently upgraded my mantis installation from 1.2rc3 up to 1.2.1. I upgraded the application, and everything is working great except from the "View Issue Detail" page which is not showing up the issue detail (numbre of issue, assigned to, etc.) Only shows custom fields I created... Any ideas? I guess it might be a configuration parameter which I might be using incorrectly...03:26
socorropc"View Issue Detail" page which is not showing up the issue detail (numbre of issue, assigned to, etc.) Only shows custom fields I created... Any ideas?03:44
CIA-22Mantisbt: hickseydr * r0b6540cb1e76 / (3 files in 3 dirs): Fix #11967: Fix bug update events (both pre and post commit)04:10
dhx_msocorropc: screenshot please?04:10
socorropchow can I give them to you?04:11
dhx_mjust upload them to a free image hosting site perhaps?04:16
socorropcgive me 1 sec...04:16
socorropcalright, dhx_m, please have a look @ http://www.4shared.com/photo/DoIJjwKD/newmantis.html & http://www.4shared.com/photo/PVu1XYON/oldmantis.html04:29
CIA-22Mantisbt: hickseydr * r11fb95de538e /bug_reminder_page.php: Fix #11981: Factor in view status when showing user monitor list04:29
dhx_mhmmm04:31
dhx_mah04:31
socorropc?04:31
dhx_m$g_bug_view_page_fields04:32
dhx_mcheck config_inc.php for anything relating to which fields to show on view_bug_page04:33
dhx_mit changed from 1.2.0RC2 => 1.2.004:33
socorropcwill have a look at it and let you know... give me a couple of minutes04:37
socorropcI searched view_bug in config_inc.php. only have $g_my_view_bug_count=10... nothing about view_bug_page neither in config_defaults...04:46
dhx_mare you sure you upgraded correctly?05:08
dhx_mie. backup old directory, extract new version into a clean directory05:08
dhx_mthen copy over your old config_inc.php and custom_ pages?05:08
socorropcpretty, I received no error messages, no warning, etc...05:09
socorropcfollowed the upgrade instructions in the website05:09
siebranddhx_m: in the commits where you removed that message: there is no need to remove messages from languages other than English if you don't want to. This will be done automagically each time I export translations.06:21
dhx_msiebrand: ok thanks for the heads up, I might use that approach from now on06:21
dhx_mI was previously using sed so it wasn't any time out of my day :)06:22
dhx_msiebrand: I was talking to nuclear_eclipse before and I think we may be pushing for a 1.3.0 release in the next month or two06:22
siebranddhx_m: oh, nice!06:23
siebranddhx_m: guess I'll have to change the supported version from master-1.2.x to master-1.3.x then...06:23
dhx_msiebrand: aiming for a ~6 month release cycle so we're not maintaining two vastly different branches of MantisBT for a long time06:24
siebrandsounds like a plan.06:24
dhx_mthe problem I see at the moment (and I'm sure you notice it too with translations) is that it is hard/complex/time consuming to deal with two branches which are quite different to each other06:24
siebranddhx_m: I don't support multiple branches at the moment.06:25
siebranddhx_m: that's because of technical limitations. If I could, it wouldn't be a big deal.06:25
dhx_mis that related to the language changes Paul keeps talking about? or is it limitations on the translatewiki end?06:26
siebranddhx_m: limitations in our backend.06:26
dhx_mto be honest I don't see us changing strings in minor releases in the future if we speed up our major release cycle06:28
siebranddhx_m: on an abstract level we have File Format Support classes (FFS). Those classes have features. The features we use for branches are only supported in one particular FFS, the one for MediaWiki core.06:28
siebrandI'd like that to be more generic, and that's on a wish list, but there are higher priorities all the time.06:29
siebrandFOSS projects can always use more developers ;)06:29
dhx_mdo you have a link to the source code... it's something I could be able to help with06:29
dhx_mif it isn't 10,000 LOC complexity :)06:30
dhx_mthe reason I'm interested is that there are plans to add templating/a new language subsystem/etc to MantisBT (probably using PHPTAL at this stage)06:31
dhx_mPHPTAL has an awareness of i18n built in and all we'd have to do is plug it into our lang/strings_* formatted files06:33
siebranddhx_m: our code is here: http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/Translate/06:35
siebranddhx_m: specific code we use for MantisBT Is http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/Translate/ffs/PhpVariables.php?view=annotate and http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/Translate/groups/Mantis.php?view=annotate06:36
dhx_msiebrand: you've been busy writing code, I don't remember it being so well organised/neat a year ago :)06:36
siebranddhx_m: people learn, and as no one else did it, I had to learn it...06:37
dhx_msiebrand: I very much like the idea of FFS where you can easily integrate different file formats by writing a simple class06:37
siebranddhx_m: Nikerabbit from #mediawiki-i18n has written most of our extension's code.06:38
siebranddhx_m: for the PHPVariables FFS there are two changes to be made.06:39
siebranddhx_m: The current FFS needs to be migrated to a new cached serialised structure that reduces parsing.06:39
siebranddhx_m: all base classes are there already, it's just the FFS specifics that needs to be put into a new class.06:40
siebranddhx_m: config of the message groups is then done using YAML.06:40
siebranddhx_m: the second is adding Mangler support, and I think with a specific extensions, so that messages can be inherited from another group, based on another template. Nikerabbit can explain if you really want to work on it :)06:41
dhx_mwell I like the simplicity of the YAML formatted files for MantisBT plugins06:41
dhx_mas I had plans to rip some of the core functionality out of MantisBT into core plugins (as was done with graphs)06:42
dhx_mso it should be pretty easy for me to send you some .yml's if that eventuates06:42
dhx_mnot that those files are the main problem... I imagine it's more work at your end setting up projects/etc on translatewiki?06:43
dhx_mI'm looking at http://svn.wikimedia.org/viewvc/mediawiki/trunk/extensions/Translate/groups/MediaWikiExtensions.php?view=markup06:43
siebranddhx_m: the technical side is no longer the challenge if FFS support is there.06:43
dhx_mto see how extensions are handled in MediaWiki... quite different I see06:44
siebranddhx_m: MediaWiki extensions is a different beast. More features (like magic files and special page aliases), and volume.06:44
siebranddhx_m: A better example is to look at the YAML for Shapado.06:45
siebranddhx_m: that's what the MantisBT config should look like in the future.06:45
dhx_mnice, thanks :)06:46
XimeraДобрый день. Вчера я произвел обновление своего MantisBT до версии 1.3.0, но столкнулся с проблемой. Из имеющейся документации я понял, что записать в параметр g_crypto_master_salt. Может кто знает, что нужно записать в этот параметр?   Microsoft IIS 7.0 MantisBT 1.3.007:01
XimeraGood afternoon.  Yesterday I made a MantisBT update to version 1.3.0, but encountered  problem. From the available documentation, I realized that to write a parameter  g_crypto_master_salt. Maybe someone knows what to write in this setting?   Microsoft IIS 7.0  MantisBT 1.3.007:02
dhx_mXimera: hi07:09
dhx_mXimera: you need to create a randomly generated value for $g_crypto_master_salt in config_inc.php07:09
dhx_mXimera: this value is kind of like a private key for MantisBT07:10
dhx_mXimera: it is used when generating passwords, CSRF tokens, etc07:10
Ximerasorry, how generate master_salt?07:15
dhx_mXimera: a random number generator... wait a sec and I'll find out how you do it on Windows07:17
XimeraSorry, David, you have not forgotten about me? :)07:32
dhx_mI've been trying to find an inbuilt way in Windows to pull random data from CryptoAPI07:32
dhx_mbut it seems like it's only usable if you write your own C++/.NET application to use this library07:33
dhx_mXimera: you can just keyboard mash out a "random" value yourself but it isn't the ideal approach to use07:36
Ximerait's don't work :( i get ERROR_CRYPTO_CAN_NOT_GENERATE_STRONG_RANDOMNESS07:37
Ximera"Unable to find a source of strong randomness for cryptographic purposes."07:38
dhx_mah yes that has been fixed as of a few hours ago07:39
dhx_mhttp://git.mantisbt.org/?p=mantisbt.git;a=commit;h=3436d148684f0216a565b7d744eef98fc3eac7e907:40
dhx_myou can backport that patch to 1.2.1 if you like07:40
Ximerathanks :)07:40
dhx_m"fixed" is used loosely as it actually disables strong crypto PRNG on Windows07:42
dhx_mnothing that a normal MantisBT user would really need to worry about though07:42
dhx_malmost killed Projax and replaced it with jQuery Autocomplete :)08:33
dhx_mthis new system will be 1000000000000 times better as all you have to do is add the "autocomplete" class to the input type="text" field you want to add autocomplete support for08:35
dhx_mthen you just need to add a function to xmlhttprequest_api.php to send the JSON data08:35
kirillkahi all12:43
kirillkanuclear_eclipse12:44
kirillkanuclear_eclipse: git can import cvs repository?12:44
nuclear_eclipsekirillka: I don't know about CVS, but it can import from an SVN repo12:53
kirillkanuclear_eclipse: I see drupal work with cvs, and don't migrate to svn or dvcs. I don't want install cvs or svn client12:55
nuclear_eclipseafaik, if they use CVS, you're gonna need to install CVS12:57
nuclear_eclipseor find a project that uses SVN or Git12:57
kirillkaO! I found on github mirror for git.mantisfr.org13:07
dhx_mnuclear_eclipse: any good with jQuery autocomplete? :)13:13
kirillkadhx_m: nice to see you13:14
dhx_mthe documentation on autocomplete sucks :p13:14
dhx_mkirillka: hi :)13:14
rolfkleefcvs -> git: a friend of mine does that (works on drupal stuff) but seems not to work well with branches in CVS (but if you only use HEAD it seems to work ok bidirectionally) -- can ask what he uses13:16
rolfkleefdrupal will move to git when they start working on version 813:16
kirillkarolfkleef: Thanks. I want give only HEAD revision, but I found mirror with git of cvs repository.13:18
nuclear_eclipsedhx_m: unfortunately not13:21
dhx_mok thanks :)13:22
dhx_mI am returning the following JSON:13:22
dhx_m713:22
dhx_m["abc"]13:22
dhx_m013:22
dhx_mI don't know why the numbers are there (from json_encode() in PHP)13:22
nuclear_eclipsedhx_m: try doing a file_put_contents("/tmp/json", print_r($data,true)) on your data before you put it through json, and make sure you have an appropriate data structure?13:23
dhx_moh it's OK... it's the HTML chunked transfer encoding at work13:27
dhx_mit works :)13:41
dhx_mjust need to work out a way of getting $(this) to work inside $().autocomplete13:42
kirillkadaryn: mo13:42
darynmorning13:42
daryn:)13:42
darynor evening as may be the case13:42
dhx_mdaryn: hi13:42
darynhallo13:43
darynwhat did i miss?13:43
darynyou converting projax autocomplete to jquery?13:44
dhx_myep, almost done13:44
darynyay!13:44
darynnow i don't have to13:44
dhx_m$('input.autocomplete').autocomplete({13:44
dhx_msource: function(request, callback) {13:44
dhx_mso any input field with the "autocomplete" class will work for now13:45
daryngreat13:45
dhx_mI'm not sure if the attribute selector is allowed there13:45
darynwhere?13:45
dhx_mbut... the problem is I need the name attribute of the input field which the autocomplete event fired on13:45
dhx_mie input[type=text].autocomplete13:46
darynyou can get that...13:46
dhx_malert($(this).attr('name'));13:46
dhx_mundefined13:46
darynyeah... it's your setup i think13:46
dhx_mI assume $(this) refers to the callback for the source: thingy13:46
darynis the code where i can look at it?13:46
dhx_mit can be :)13:47
darynthis changes constantly in js13:47
darynlet me take a look at it13:47
darynyou can do.... $(input.autocomplete').each( function() { alert($(this).attr('name')); $(this).autocomplete( ...13:49
dhx_m$('input.autocomplete').autocomplete({13:51
dhx_msource: function(request, callback) {13:51
dhx_m$.getJSON('xmlhttprequest.php', {'entrypoint' : 'os_get_with_prefix', 'os' : request.term}, function(data) {13:51
dhx_mvar results = [];13:51
dhx_m$.each(data, function(i, value) {13:51
dhx_mresults.push(value);13:51
dhx_m});13:51
dhx_mcallback(results);13:51
dhx_m});13:51
dhx_m}13:51
dhx_m});13:51
dhx_mI need ($this).attr('name') instead of the 'os' string13:51
darynright13:51
dhx_mhowever it isn't defined at that point for some reason13:52
darynyeah because the this no longer references the input at that point13:52
darynit's inside the function and the input is not passed into the function13:52
darynyou should be able to use the each above, set a variable for it and then the variable should be in scope for the function...i think13:53
darynor...13:53
dhx_mhmm13:53
dhx_mhttp://github.com/jquery/jquery-ui/blob/master/ui/jquery.ui.autocomplete.js13:54
darynthis is the plugin you're using?13:56
dhx_maha13:57
dhx_mhttp://blog.seanja.com/2010/03/using-the-jquery-ui-autocomplete-widget/13:57
dhx_m$(this).attr('element').attr('id');13:57
dhx_mlet's see13:57
darynthere you go13:57
dhx_mbingo!13:58
darynsweet.  that's cleaner for sure13:58
dhx_mwell almost...13:59
dhx_mvar field_name = $(this).attr('element').attr('id');13:59
dhx_m$.getJSON('xmlhttprequest.php', {'entrypoint' : field_name + '_get_with_prefix', field_name : request.term}, function(data) {13:59
dhx_mmy concatenating of strings must be wrong13:59
kirillkadhx_m: is it autocomplete implements in mantisbt?14:00
kirillkawill?14:00
dhx_mkirillka: I'm replacing Projax with jQuery Autocomplete14:00
dhx_mso for fields where it makes sense, yes14:00
kirillkagood idea14:00
dhx_mand it'll be trivial to work with in the future14:01
dhx_munlike the kludge that is Projax14:01
daryndhx_m aren't you supposed to be using $j rather than $?14:02
dhx_mdaryn: I chopped out the compatibility... do we want it back?14:02
dhx_mone of the comments indicated it was temporary14:02
darynoh...is projax totally removed?14:02
dhx_myes14:02
dhx_mdead14:02
darynawesome14:02
dhx_mhaha indeed14:02
dhx_mjust this last bug to go14:02
darynnp. just means i have to update all my plugin scripts14:03
daryn:P14:03
dhx_mI think people expect to work with $() when using jQuery anyway14:03
darynoh, definitely14:03
dhx_moh14:06
dhx_mfield_name : request.term14:06
dhx_mfield_name is treated as a string14:06
dhx_mrather than a variable14:06
kirillkafootball begin14:06
dhx_mthe entrypoint is OK though14:06
dhx_mhttp://stackoverflow.com/questions/739151/javascript-variable-evaluation-in-a-map14:08
dhx_mstackoverflow to the rescue14:08
daryndhx_m: filters? :D14:26
dhx_mdaryn: not yet :(14:37
dhx_mdaryn: almost done now... just need to escape HTML entities with jQuery14:37
daryn:) i figured. sounds like you've been working on good stuff14:37
daryni'm working on filters now.  writing a base class and changing some of property names14:38
darynso...it will hopefully be a bit cleaner shortly14:38
daryndhxm: nuclear_eclipse: what if i want plugins to be able to build filters other than adding to the view issues page?  the token_api only allows one filter token per user and it uses a constant to define the type...I need more than one and I need to be able to dynamically add them...thoughts?14:56
dhx_mdaryn: I guess a new table of filters vs users?14:58
dhx_mdaryn: columns include user_id, filter_id, filter_type?14:58
dhx_mfilter_type could be "last known filter for view_all_..." or "last known filter for blah_..." or "saved filter for ..."14:59
dhx_m(of course, change the schema to be more correct) :)14:59
darynwell...we have a filter table we could just add the type to but there are two parts to this...14:59
darynproblem is that persistent filters use the php session cookie but temporary filters use the token table...i think i just answered my own question15:00
darynthey are temporary so overwriting because i'm on a different page won't matter.  I can use the same token type15:01
darynbut...for the filters table a type...15:01
dhx_mlol 'persistent' using the session cookie?15:04
darynwell...it stores it in the database and passes the id around in the session cookie15:04
darynthe filter table stores both 'current' filters and 'named filters'15:05
dhx_mwhat JavaScript compressor am I meant to use?15:12
daryni don't know... i was just using http://javascriptcompressor.com/15:13
darynbecause it was at the top of google query15:13
dhx_mhttp://refresh-sf.com/yui/15:18
dhx_mthat one seems to work OK too15:18
dhx_mthis sucks though... it should be handled by our build script :)15:18
daryni agree15:19
CIA-22Mantisbt: hickseydr * rb77ea9cd2333 / (34 files in 9 dirs): Fix #9117: Replace Projax with jQuery UI Autocomplete15:21
nuclear_eclipsedhx_m: feel free to update the build scripts :P19:16
nuclear_eclipseI can always publicly humiliate you later by reverting your changes ;)19:16
ToffeePopsHi, is it possible to stop users from editing an issues description or additional info after an issue has been raised?22:02
ToffeePopsThe problem is when a user edits (eg) a large description, all we see in the history is 'description updated', so we have to review the description and try to figure out what has changed (i.e without knowing explicitly what has changed, we don't know if it is important or not).22:02
ToffeePopsWe are trying to stop our users from editing description and additional info fields and get them to always add notes, but we'd like to formalise it a bit...22:02
nuclear_eclipseToffeePops: in 1.2.x, Mantis keeps track of old revisions to all the long text fields so that you can audit changes to them after the fact22:09
nuclear_eclipsedoh22:09
nuclear_eclipseToffeePops: in 1.2.x, Mantis keeps track of old revisions to all the long text fields so that you can audit changes to them after the fact22:10
ToffeePopsnuclear_eclipse: Awesome! Is there anything special that needs to be done to audit changes?22:11
ToffeePopsCan anyone do it or only the administrator?22:11
nuclear_eclipseanyone can see the old revisions22:12
ToffeePopsfantastic22:12
nuclear_eclipsebut iirc it's a configurable threshold22:12
ToffeePopsBut where possible we're trying to encourage people to add notes rather than edit existing ;-)22:13
ToffeePopsWe're still on 1.1.8 so I guess it's time we updated22:14
ToffeePopsHmmm, I clicked on a link to 1.3.0 changelog to see what else was coming up... and got an app error 1601 :(22:16
ToffeePopsURL http://www.mantisbt.org/bugs/changelog_page.php?project=mantisbt&version=1.3.0-dev22:16
nuclear_eclipsehttp://www.mantisbt.org/bugs/changelog_page.php?project=mantisbt&version=1.3.x22:16
nuclear_eclipseI'd definitely not recommend running 1.3.0 yet though22:16
ToffeePopsYeah no worries, was just curious ;-) just a shame to have a dodgy link on http://www.mantisbt.org/download.php22:24
ToffeePops(Change log link under development section)22:24
ToffeePopsBTW, our mantis admin said the reason he hasn't updated was that when he tried 1.2.0 the upgrade spat our sql errors updating the database. He's going to try again shortly with 1.2.1 :-)22:26
ToffeePopss/spat our/spat out/22:27
ToffeePopsnuclear_eclipse: Cool, we're upgraded to 1.2.1... I guess the migration improved :-)22:50
ToffeePopsAt least we think it's 1.2.1... the version number down the bottom LHS of the screen has been removed o_O ?22:50

Generated by irclog2html.py