Monday, 2009-04-13

../irclogs/#mantishelp.2009-04-13.log
--- scribe started ---00:00
CIA-11Mantisbt: daryn * r4e629fc69443 / (3 files in 3 dirs): =Implement bug 4723, display issues I've added notes to in a my view box01:55
mantisbotBug 4723 - kgrubbs - fixed - resolved01:55
mantisbotAdd new capability to "My View".... "Issues I've added Notes To" - http://www.mantisbt.org/bugs/view.php?id=472301:55
CIA-11Mantisbt: vboctor * ra6773fcd3fe5 / (7 files in 3 dirs): Fixed #9102: webservice api call causes SYSTEM NOTICEs \nAdded unit tests for SOAP API.01:56
dhx_mohhh changes :)02:27
dhx_mAHHH sourceintegration (websvn) just got stuck in an infinite loop of importing the same changeset.... 20,000+ times!03:37
dhx_mthis could be bad for the server03:38
nuclear_eclipsedhx_m: I'm going to bed now, I'll investigate that tomorrow at work, probably a side effect of a recent "feature" (issue #14)- if you revert to commit fc0c7b, you should be fine03:41
dhx_mok thanks :) I'll check it out03:41
dhx_mcrisis averted as well... by renaming the svn trunk directory while the import was in progress ;)03:41
nuclear_eclipsecheers03:42
CIA-11Mantisbt: vboctor * rffff385dacfc / (api/soap/mc_issue_api.php tests/soap/IssueUpdateTest.php): Fixed #9102: webservice api call causes SYSTEM NOTICEs --- fixed for mc_issue_update().03:48
mantisbotNew bug: Bug 10322 - vboctor - open - new03:59
mantisbotNew bug: Html markup is not handled correctly by the soap APIs - http://www.mantisbt.org/bugs/view.php?id=1032203:59
mantisbotNew bug: Bug 10323 - vboctor - open - new04:04
mantisbotNew bug: mc_issue_update() will set optional fields to default values rather than leave them with current values - http://www.mantisbt.org/bugs/view.php?id=1032304:04
CIA-11Mantisbt: vboctor * r974883d711bc / (api/soap/mc_issue_api.php tests/soap/IssueUpdateTest.php): Fixed #9472: Issues updating duplicates notes.04:35
CIA-11Mantisbt: vboctor * ra5f7dbcd540c / (bug_report.php core/bug_api.php tests/soap/IssueAddTest.php): Fixes #9132: mc_issue_add does not set values for all submitted fields.05:30
vboctornuclear_eclipse, are you there?05:38
vboctorbug update advanced is broken by the following checkin:05:38
vboctorhttp://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=c0d04c65c53f004f5f37364b80559a83243e54cf05:38
vboctorYou have defined $t_show_profiles and checked for $t_show_platform.05:39
dhx_mvboctor: he is sleeping I think :)05:45
dhx_moddly enough, I can't reproduce it...05:46
dhx_mthe profile fields are missing, as expected05:46
vboctordhx_m, it depends on your error reporting settings.05:58
vboctorI think he mis-named the configuration option and the variable.05:59
dhx_mI have PHP errors on... or is there another level of verbosity?06:17
dhx_mand I have profiles disabled... and those fields don't show on the advanced update06:17
dhx_munless the if() turns into more of an "ifdef", and it is ignored06:18
CIA-11Mantisbt: vboctor * reeb012b41654 / (3 files in 2 dirs): Better handling for attempts to add blank notes, issue notes unit tests, test for mc_issue_exists().07:24
mantisbotNew bug: Bug 10324 - rolfkleef - open - new11:16
mantisbotNew bug: Add filter options "Next target version" and "Last released version" - http://www.mantisbt.org/bugs/view.php?id=1032411:16
mantisbotNew bug: Bug 10325 - thraxisp - open - assigned11:16
mantisbotNew bug: Move should have the option of changing categories - http://www.mantisbt.org/bugs/view.php?id=1032511:17
mantisbotNew bug: Bug 10326 - prabhurangan - open - new11:52
mantisbotNew bug: Facing issues with changelog feature in LINUX(FC10)? - http://www.mantisbt.org/bugs/view.php?id=1032611:52
nuclear_eclipsemorning all12:25
dhx_mhey12:34
nuclear_eclipsedhx_m: was it import or checkin that you were having issues with last night?12:39
nuclear_eclipsenevermind, it only loops on import :P12:42
* nuclear_eclipse hasn't had is caffeine yet...12:42
nuclear_eclipsedhx_m: hmm, I'm not replicating the infinite loop on my system12:47
nuclear_eclipseit completes correctly when a) there are changesets to import, and b) when there are no changesets to import...12:48
dhx_mnuclear_eclipse: import all (initial import)12:54
dhx_mhmm I was up again an old version of subversion...12:54
dhx_mas I see it, it keeps looping until the number of changesets returned is < 112:54
nuclear_eclipseright, and the websvn plugin should return 0 objects when it gets to the end of the revision history, unless older versions of `svn log` don't error out on unknown revs like 1.4.4?12:56
nuclear_eclipsein this case, "older" to me being older than the 1.4.4 binaries on my workstation =\12:57
dhx_mbut this wasn't specifying a revision range12:57
nuclear_eclipsewell, I'm talking about how the WebSVN plugin actually handles imports12:58
nuclear_eclipseit makes a call to `svn log -v -r XXX:HEAD --limit 200` and then parses the output12:58
dhx_mimport_full: $t_svnlog = explode( "\n", `$svn log -v $t_url` );12:58
dhx_mno revision range specified for the initial import12:58
dhx_mimport_latest seems to be ok12:59
nuclear_eclipseoh, I see12:59
dhx_mand thanks for making me look at that again12:59
nuclear_eclipseI missed the fact that import_full didn't spec a rev range12:59
* nuclear_eclipse kicks himself12:59
dhx_mI realise how it checks the revision range now :)12:59
nuclear_eclipsethis is the problems found when primary development is done against Git plugins, and then ported to the SVN plugins :P13:00
nuclear_eclipseI forgot that I had implemented import types differently in SVN, while the Git plugins use the same logic for both full and latest imports13:01
dhx_mI'm not using SVN by choice :p13:01
nuclear_eclipsesame here, and I'm also not using SCCS by choice....13:01
dhx_malso, won't you hit the script execution time limit with very large repos?13:01
nuclear_eclipseright, that's the purpose of helper_begin_long_process(), it extends the execution time (assuming safe_mode is off)13:02
dhx_mnice :)13:03
nuclear_eclipsedhx_m: I just pushed a fix for the SVN problem13:05
dhx_mthanks, I'll try it out13:05
nuclear_eclipsethis just proves to me that I was right for holding off an "official" release with all the new features I added over the past week :P13:06
dhx_mwith ticket #17, it looks like it'd be nice to have the ability to setup regex matches per repo... rather than globally13:11
dhx_mso you don't get incorrect linking of change sets13:11
dhx_mwait that's odd13:12
dhx_myou must have just used the wrong bug number I guess13:12
nuclear_eclipse?13:12
nuclear_eclipseoh, you mean the "3d games" changesets?13:13
nuclear_eclipsethose are an artifact of having imported those repositories into a different tracker than they were originally developed against =\13:14
dhx_mah I see :)13:15
dhx_measily fixed by running some manual SQL queries to remove those links I guess13:15
nuclear_eclipseprobably, or I could just be lazy and leave them there :P13:16
dhx_m:)13:16
nuclear_eclipsethey run out around #20ish IIRC13:16
dhx_mwell it isn't really a problem then :)13:16
dhx_mI take it the source repository is checked for matching entries for newly reported bugs?13:17
nuclear_eclipseno, it just creates the links in the database without making sure the bug actually exists or not :P13:17
nuclear_eclipsethat should probably be considered yet another "bug"; I just haven't found it enough of a problem to take time to fix it, even though it's probably a dirt easy fix13:18
dhx_myep13:20
dhx_mwell I think I prefer it the way it is now13:20
dhx_mso you can commit something, then file a bug 5 minutes later13:20
mantisbotBug 5 - prescience - fixed - closed13:20
mantisbotNeed show Next/Prev XYZ bugs - http://www.mantisbt.org/bugs/view.php?id=513:20
nuclear_eclipsehehe13:20
* nuclear_eclipse kicks mantisbot 13:20
dhx_mthat was really strange13:21
nuclear_eclipseexactly13:21
nuclear_eclipsethe bug hasn't been touched in 9 years...13:21
dhx_mit predates Mantis tracking bug history13:21
nuclear_eclipsehehe13:21
* nuclear_eclipse is working on your timecard contribution13:23
nuclear_eclipsedhx_m: do you use *nix or Windows?13:24
nuclear_eclipseand also, do you git command line, or a gui?13:24
dhx_mnuclear_eclipse: Gentoo Linux13:26
dhx_mnuclear_eclipse: git command line13:27
nuclear_eclipseok13:27
nuclear_eclipseone random thing I get absurdly picky about for no good reason is trailing whitespace; it bugs my inner OCD :P13:27
dhx_moh, did I have some? :o13:28
nuclear_eclipsebut if you configure git to use colors when diffing, it makes it glaringly obvious (probably another reason I OCD on it)...13:28
nuclear_eclipseI'll take care of it via `git commit --amend` which I'd do to signoff anyways, but I just thought I'd let you know13:29
nuclear_eclipseok, I pushed your changes; thanks again :)13:32
dhx_mno probs, thanks13:32
dhx_mI think the white space problem13:32
dhx_mis because I haven't setup options for that repo13:33
dhx_mI'll see if I can make them global13:33
nuclear_eclipseif you do `git config --global`, you can set your color prefs there, and then it always shows up13:33
dhx_myep13:34
nuclear_eclipsehow would you see #18 being implemented?13:34
dhx_mthat's a good question13:34
nuclear_eclipse:)13:34
nuclear_eclipse:)13:34
dhx_mI'm not sure if it makes sense to let all users configure it on their own13:35
dhx_mmaybe just add a threshold level for it13:35
nuclear_eclipseyeah, that's one of the reasons I haven't yet done anything like that... =\13:35
dhx_mso developers/admins can do that13:35
dhx_mbut then you'd also need a configuration page where admins can override on a per-user basis13:35
nuclear_eclipseI wonder...13:36
dhx_mmanage_user_edit_page.php13:36
dhx_mand13:36
dhx_maccount_page.php13:36
nuclear_eclipseI recently added plugin events for the user's account prefs page13:36
dhx_mthose 2 places probably13:36
nuclear_eclipseI wonder if I set it so that the user's commit name could be set only if a) the user exceeded a certain threshold, or b) an admin was setting someone *else's* ...13:38
dhx_mI just added a note to that effect :)13:39
dhx_mall we need is one threshold13:39
dhx_mand then we only show it on the user page if the user is above that threshold13:39
nuclear_eclipseI'd of course then also need to create space on the source-integration config page to set the user threshold...13:39
dhx_myep13:39
nuclear_eclipsehmm...13:39
dhx_malso the "Update Threshold"... does that let people manually import the latest revisions?13:40
nuclear_eclipsethe good news is that with my recent extraction of mapping commits to users, it's far more prudent to implement now than it was a couple weeks ago...13:40
dhx_mor just attach changesets to tickets?13:40
nuclear_eclipsegood question :P13:40
dhx_mI couldn't find a button to that effect for users who have "update" permissions13:41
* nuclear_eclipse looks into the code13:41
dhx_mas the button only appears when you click "Manage" next to a repo13:41
nuclear_eclipseoh, update_threshold is used for a) attaching and detaching commits to issues, and b) for the optional porting status feature13:44
dhx_mok :)13:45
dhx_mI did have some wording suggestions that I'll make later13:45
dhx_mfor this + Timecard13:45
dhx_mjust to make things easier to understand13:45
nuclear_eclipseporting status being a feature my employer uses to help keep track of fixes that have/haven't been ported to older branches, etc13:45
nuclear_eclipseesp considering there are times where fixes need to get ported back to three or more legacy versions that are still in active use around the globe...13:46
dhx_mso it tracks merges?13:48
dhx_mor is it more general... and just based on a commit message saying "ported"13:48
nuclear_eclipsewell, not directly; it requires developers to explicitly update commits to say what branch they ported the fix to13:48
dhx_mcool, I should try that out :)13:49
dhx_mdo you then define what branches you have?13:49
nuclear_eclipsewe still use a set of Perl scripts on top of SCCS for our primary project... (ugh)13:49
dhx_moh :o13:50
dhx_mI really like git now that I've started using it13:50
nuclear_eclipseso anyways, the porting status feature just lets you specify whether a commit a) doesn't need to be ported, or b) what branch it was ported to by selecting from a list of all branches available13:50
dhx_mnever liked SVN, and haven't really used much else13:51
dhx_mcool13:51
nuclear_eclipseI used to use SVN for everything until I found out about Git, now I use Git for everything :)13:51
nuclear_eclipseand with SCCS, "branches" are only a concept manifested in our miriad of perl scripts....13:52
nuclear_eclipseotherwise, SCCS barely handles revision control for a single file, let alone an entire source repository13:52
dhx_msounds like someone set your system up with their CVS knowledge13:52
dhx_mthose sort of scripts were apparently very common in the CVS days13:53
nuclear_eclipsebut we can't switch to anything newer, because "then we can't grep through the revision history files".....13:53
dhx_mor else you'd need to type 15 commands to commit a patch :p13:53
dhx_mor some other hideously stupid thing like that13:53
dhx_mcommit messages are stored in separate files, not in your SVN repo?13:54
nuclear_eclipseno, I don't think you're understanding; we use SCCS, from the 70's, not SVN13:54
dhx_moh I see13:55
dhx_mthat makes sense now13:55
nuclear_eclipseeg, it was *the* first source control system ever created13:55
dhx_mI suppose you could port all your history to git13:55
dhx_meven in an ugly way... walk through the history and commit each change one by one to the git repo13:56
dhx_mand then update your system clock in between (if git doesn't have a timestamp override... I'm sure it does though) :)13:56
nuclear_eclipsewell, it's a matter of our senior development staff refuses to learn anything new, because SCCS works "perfectly well"13:56
dhx_mmust be a headache for others to use13:56
nuclear_eclipseindeed13:56
dhx_mI suppose there wouldn't be too many branches that get merged13:57
dhx_mbecause it'd be so hard to do so?13:57
nuclear_eclipseha, you wish13:57
nuclear_eclipsefor generaly bug fixes, there's generally about 3-7 "branches" that the fix has to be ported to...13:57
dhx_mouch13:58
nuclear_eclipseother departments use SVN though, and one is looking to switch to Mercurial, which was the major reason I created the generalized source-integration framework, rather than hardcoding everything like the previous system that Mantis will be replacing13:59
dhx_mthe new framework you've created is excellent13:59
dhx_myou've done a great job making it modular13:59
dhx_mthat is exactly how good software is written :)13:59
nuclear_eclipseI try :)14:00
dhx_mwill there be plans to rip out the old cruft from Mantis once this plugin 'takes off'?14:00
dhx_mand is this the sort of thing that may become officially "managed" by the project?14:00
nuclear_eclipseprobably not for at least another full release cycle at the shortest14:00
dhx_mie. bugs reported to the official Mantis tracker, the plugin is bundled, etc14:01
nuclear_eclipseeg, need to maintain some sort of legacy compatibility for the time being14:01
dhx_myep14:02
nuclear_eclipsewe have given consideration to maintaining a second set of releases containg popular/useful plugins out of the box14:02
dhx_mand I suppose it'd need an upgrade path14:02
nuclear_eclipsewell, I don't think there's any real way to provide an upgrade path for any existing notes-based source integration14:02
dhx_mah, probably not then14:03
dhx_mI haven't used it before, so didn't know it was based on notes (UGH!)14:03
nuclear_eclipseor at least, nothing more than "install and set up this plugin"14:03
dhx_myep14:03
nuclear_eclipseyeah, that's the biggest reason I didn't like the old style integration14:03
dhx_mwell I was thinking more in terms of if good plugins become managed by the project14:03
dhx_mit means they get to use the existing infrastructure (bug tracker, release cycle, website, etc)14:04
nuclear_eclipsebasically, the old "integration" was nothing more than piping a post-commit text payload into checkin.php which just added a note to any linked issues14:04
dhx_m... more users... more development activity (?)... and if someone gets hit by a bus, it isn't the end of the plugin14:04
dhx_mugly14:04
nuclear_eclipsewell, hopefully by using Git plugins on mantisforge, the bus situation should be "solvable" just by forking the repo, or paulr modifying the maintainer of the primary repo14:05
nuclear_eclipseGit repos*14:05
nuclear_eclipsealthough I hope the bus situation never happens, because 90% of the repos on mantisforge are mine :P14:06
dhx_mlol14:11
dhx_mit'd be nice to have a Mantis bugtracker for plugins/stuff on mantisforge14:11
dhx_mie... 1 tracker for everything... official & non-official?14:11
nuclear_eclipseagreed, but paulr basically stopped before getting anything set up...14:11
dhx_mif it was me, I'd struggle to find the time to do that :)14:14
nuclear_eclipseyeah, between work, university, and world of warcraft, I struggle as well :P14:15
slestaki just started playing wow this weekend.  i can see how it kills productivity :)14:21
dhx_mhaha14:21
nuclear_eclipseI only started playing because a friend of mine runs a private server, so I don't have to pay that horrid monthly fee14:22
slestakmy wife and kids do, so its the only way to get "family time" ;)14:22
nuclear_eclipsehehe14:22
slestakpersonally, id rather be outside on my bike14:23
slestakim usign the soap connector to integrate with deki.  the bug links are pointing to the soap uri instead of linking to the mt login page to allow viewing the ticket.  is that typical?14:24
nuclear_eclipsenot to my knowledge :P14:24
slestaki'll ping over there and see if anyone else is using mc in this fashion.14:25
dhx_mwell I'm off for now14:35
dhx_mthanks for the fixes nuclear_eclipse :)14:35
dhx_mcya later14:35
nuclear_eclipseyou're welcome14:35
nuclear_eclipsecheers14:36
CIA-11Mantisbt: s.mazeland * rfd43db318727 /lang/strings_english.txt: I've -> I Have in message14:56
paulr_moo16:33
darynhello paulr_16:33
paulr_lo16:33
paulr_did you see mail ?16:33
darynon dev?16:34
paulr_to you16:34
darynhm...don't think so...just a sec16:34
paulr_aboutabout dates16:34
paulr_i've sent 3 emails in last 48 hours16:34
paulr_2 today16:34
darynoh yes16:34
daryni tried to respond on the test db but my mail relay wasn't working at home16:35
darynanyway.  for that one show queries and count were both off16:35
paulr_really?16:35
darynyes16:35
paulr_can you try running the latest version against your test db16:36
darynsure.16:36
darynwill take me a bit16:36
paulr_what about the comments on commits today/16:36
darynfor exceptions?16:37
paulr_I was more thinking for ETA/status16:38
paulr_i've not tested yet, but as far as I can tell16:38
paulr_we've just modified bug_Report.php so I can submit a resolved bug as a normal user16:38
daryni saw the email but haven't had a chance to look into it at all16:38
paulr_http://git.mantisforge.org/w/mantisbt.git?a=commitdiff;h=a5f7dbcd540c533546fc80533866d808d4c64f9d16:42
* paulr_ doesn't get it16:42
paulr_that's a soap bug if anything16:43
paulr_not a ui bug16:43
mornethi16:45
paulr_hi16:45
mornetis there anyone that perhaps have made a custom function to email a seperate person not in the user table16:47
mornetusing a custom field that is required to filled in and then to use that email when an issue has been resolved16:48
morneti have made the changes but I'm not very happy to hard code stuff16:48
mornetthis is only required for one project though, that is why have hardcode the project id for now, but it would be nice to have this more dynamic16:49
paulr_it's not something that core project does16:49
* paulr_ is thinking a) custom functions as you've probably done16:49
paulr_b) in 1.2.x plugin16:49
mornetnope, not even a custom function, I just added a little piece of code to email_api.php16:51
morneti thought, actually hope that there would be a plugin in the next release16:51
mornetfor functionality like this16:52
darynpaulr_ installation successful16:54
daryni haven't reduced the memory though so that was at 256M16:55
paulr_daryn: I *think* you must be calling one of the debug stuff17:01
paulr_that populates the array of database calls17:01
paulr_as from what i can tell17:01
paulr_without that, it's static17:01
nuclear_eclipsehi Victor17:02
darynwell...it worked this time. i'll reduce the memory back down and see if it still works for me. If it doesn't I'll look into that a bit more. If it does...no worries.17:02
vb123hi nuclear_eclipse17:03
vb123did you see my message relating to update advanced page broken?17:03
nuclear_eclipseyeah, I was just about to look into that...17:04
nuclear_eclipseI haven't noticed any troubles on my end...17:04
vb123you defined a config / var for profiles, but attempt to hiden platform with undefined variable.17:04
paulr_nuclear_eclipse: did you have a look at my note about user preferences?17:04
paulr_vb123: if you've got any time atm, can you have an initial glance at dates/bugobjects branches17:05
nuclear_eclipsevb123: just looked at the diff you linked, and it's obvious now :P17:05
vb123paulr_: sorry can't look now.17:05
paulr_ok17:05
vb123nuclear_eclipse: yep it is!17:05
* paulr_ would like to get dates in sooner rather then later17:06
paulr_and both need testing/polishing by me17:06
paulr_so just want initial opinions17:06
vb123with the dates one, every use has his/her own timezone.  And database uses a single time zone?  e.g. UTC?17:07
paulr_database stores unixtimestamp17:07
paulr_then we localise17:07
paulr_that's the plan17:07
vb123great.17:07
vb123the unixtimestamp is dependent on the server time zone.17:07
vb123so we need to know the server timezone and user time zone, to calculate correctly, right?17:08
paulr_time() states: Unix Epoch (January 1 1970 00:00:00 GMT).17:08
nuclear_eclipsebiggest problem we have is that we'll need to convert the db schema for everything from datetime to integer, which is a hefty conversion task17:08
paulr_nuclear_eclipse: well, seems to work for me ;p17:08
paulr_I tried it on old mantisbt.org/bugs db17:09
nuclear_eclipseright, I'm just saying that's a large change to the schema, and it's certainly going to tick off aptituz_ and/or debian with their separate system of handling schema changes....17:09
paulr_heh17:09
nuclear_eclipsehe was already frustrated enough with the category migration tasks17:10
paulr_well, I don't see how you can avoid that ;/17:10
vb123have we talked lately about when 1.2.x will see the light as an rc?17:10
nuclear_eclipsevb123: I think it'd be nice to drop an rc as soon as dates get solidified?17:10
paulr_vb123: with me off work for the next week and in a coding mood, next weekend would be a good time to see where we are ;p17:10
nuclear_eclipseI was tempted to release 1.1.7 this week as well17:11
nuclear_eclipsein fact, I think I'll make a note on the dev list asking for any final targets for 1.1.717:11
vb123It would be great to identify the issues to focus on.  The two branches are really new features.17:11
vb123I'm focusing on soap, since it was totally broken.17:12
paulr_they more 'attempt to tidy stuff up'17:12
paulr_victor: out of interest, do you like the current soap implementation we have atm?17:12
vb123I like the fact that it is the only one we have now and it needs to be working.  However, I have some concerns about it.17:13
vb1231. authentication.17:13
vb1232. scalability17:13
vb1233. incompleteness.17:13
paulr_i've been hesitant to work on it too much given17:14
vb1234. versioning17:14
paulr_a) not having a proper test setup17:14
paulr_b) not really knowing what you should/shouldn't do in soap17:14
CIA-11Mantisbt: jreese * r807d7826148b /bug_update_advanced_page.php: Fix PHP notices with undefined variables due to commit c0d04c65.17:14
paulr_c) backwards compatibility17:14
paulr_for example17:14
paulr_In mantis, you can disable display of mantis version e.g. 1.2.0 - that could be a 'security' feature17:15
paulr_however, the soap api if turned on exposes it in readonly mode17:15
paulr_however, we can't turn that off easily as we use that for versioning the scope api17:15
paulr_;)17:15
vb123yes, I'm aware of that.17:16
vb123I would also change the APIs to throw exceptions that soap APIs can catch and hence remove the duplicate validation in soap.17:16
paulr_and as we know there's a bunch of similar stuff where we've added config per project/user stuff since the soap api started17:16
vb123I think that per project config is a bug farm17:17
vb123This needs a release on its own to fully test / fix across all scenarios.17:17
paulr_we have 'per-project' config atm17:18
paulr_just no nice ui17:18
vb123yep, my point is whether in all the appropriate scenarios we call config_get() with the right parameters.  Mainly in cases where we are executing an API for a project / user different than the current one.17:18
paulr_anyway, soap api is a bit of can of worms ;/17:19
vb123the short story is that there is a lot of work in soap, but we need to take it one step at a time.17:19
paulr_especially if we want to try and maintain some level of compatibility17:19
vb123automation is critical for it.  I used to depend on .NET unit tests that run on top of MC client, but I wanted to move to something which runs directly on top of the PHP web service.17:20
paulr_anyway, both the date + bugobject branches are reworks of stuff to try to fix existing bugs17:20
vb123for soap, I avoid breaking compatibility.17:20
paulr_so in theory once complete, should be drop-in replacements17:20
vb123however, big re-works / diffs are likely to introduce new bugs.17:21
CIA-11Mantisbt: jreese * rffbfd85cfbef / (3 files in 2 dirs): Fix #10299: Fixed incorrect or missing HTML colspan attributes.17:23
CIA-11Mantisbt: jreese master-1.1.x * r185c60dfef52 / (3 files in 2 dirs): Fix #10299: Fixed incorrect or missing HTML colspan attributes.17:23
paulr_hmm17:24
paulr_need to setup a instance of mantis as part of build to test soap17:24
vb123nuclear_eclipse: did you add the new configuration option to the manual?17:25
nuclear_eclipseI didn't realize we were documenting config options in the manual :P17:25
nuclear_eclipsethe code is it's own documentation!17:25
vb123yep, it should be part of the same commit as adding the configuration option :)17:25
nuclear_eclipse;)17:25
paulr_vb123: regarding reworks+bugs, I think it depends on how likely we think they are to cause major breakage. atm, we can always do a short beta release to stabalize anything new17:27
paulr_nuclear_eclipse: any thoughts on where to put the timezone stuff? ;p17:28
paulr_(Setting17:28
nuclear_eclipseMy Account -> Preferences17:29
nuclear_eclipseI'd suggest right under Language17:29
paulr_i meant, when to hook it in ;p17:30
nuclear_eclipse?17:30
paulr_actually first question17:30
paulr_when I click on 'user prefs reset' here17:30
paulr_it errors17:30
paulr_doyou get the same or does it work for you?17:31
nuclear_eclipseit "worked" on my test install, but it puts user prefs in a bad state, ie not any sort of "sane" reset17:32
paulr_I seem to get:17:32
paulr_atabase query failed. Error received from database was #1048: Column 'advanced_report' cannot be null for the query: UPDATE mantis_user_pref_table17:32
paulr_SET default_profile = 0, default_project = 0, advanced_report = NULL, advanced_view = NULL17:32
paulr_with mysql ;/17:32
nuclear_eclipseor rather, 'reset prefs' reset them to nothing at all like what the default user prefs are17:32
paulr_in fact, shouldn't reset pref's just delete the pref_table entry?17:33
paulr_or is it more17:33
nuclear_eclipseI dunno17:33
paulr_'reset prefs for this project'17:33
paulr_but from what i can tell17:33
nuclear_eclipsewell, considering account prefs (AFAIK) is supposed to be project-agnostic, it shouldn't matter what project is selected17:33
paulr_we use ->Get() to load17:33
paulr_or well17:34
paulr_->Get('default_profile') means17:34
paulr_load default_profile, if not lookup and return default17:34
paulr_but then we tend to use ->default_profile17:34
paulr_to do stuff17:34
* paulr_ wants to do same to user_preferences as he did to bugs last thursday ;p17:34
* paulr_ wants feedback on what he did to bugs last thursday ;/17:35
* nuclear_eclipse has not enough time for reviewing things atm =\17:35
paulr_it's more initial feedback ;p17:36
* paulr_ takes dusts off sledgehammer and looks at user_pref_api.php17:38
* daryn_away is away: Gone away for now17:40
vb123nuclear_eclipse: for 1.1.7, we should make sure there is a way to turn of validation tokens for users that have trouble with it.  This should allow them to continue working while we are troubleshooting the issues.18:00
vb123that is my 2c worth.18:00
nuclear_eclipseah, yes, I attached a fix to bug 9999 to implement that, but skay is saying that patch causes troubles with something seemingly unrelated18:03
mantisbotBug 9999 - arul_k_kumar - open - assigned18:03
mantisbotAPPLICATION ERROR #2800 - While submit a new bug - http://www.mantisbt.org/bugs/view.php?id=999918:03
nuclear_eclipseI'll try to look into that more to see if it's a side effect of what I did change...18:03
vb123ok, as long as this is the plan.18:03
* nuclear_eclipse apparently forgot to target that for 1.1.718:05
CIA-11Mantisbt: paul * r047e1535e743 /print_all_bug_page_word.php: Perf: reduce number of lang_get and user_pref_get_pref calls whilst looping around in word reports.18:05
* paulr_ pushes more changes18:05
CIA-11Mantisbt: paul * ra1e20db97fe5 /print_all_bug_page_word.php: this should be $j not $t_row_count (thanks victor)18:05
CIA-11Mantisbt: paul * rf6d71ff36607 /bug_report.php: Don't allow gpc to set status in bug_report - this is only for SOAP api.18:05
CIA-11Mantisbt: paul * rcf155ebc47f1 /core/classes/MantisEnum.class.php: Remove unnecessary loop in MantisEnum (thanks victor)18:05
CIA-11Mantisbt: paul * r3b921fa98c0a /core/print_api.php: Html Validation: dont print empty span tags18:05
CIA-11Mantisbt: paul * ra982f75fedef / (changelog_page.php roadmap_page.php): encode version in url.18:05
paulr_mysql> select * from mantis_user_profile_table where platform = '';18:13
paulr_+-----+---------+----------+----+----------+-------------+18:13
paulr_| id  | user_id | platform | os | os_build | description |18:13
paulr_+-----+---------+----------+----+----------+-------------+18:13
paulr_|  14 |       0 |          |    |          |             |18:13
paulr_|  16 |       0 |          |    |          |             |18:13
paulr_|  17 |       0 |          |    |          |             |18:13
paulr_can't we do an upgrade to delete blank profiles? ;/18:13
CIA-11Mantisbt: paul * r4ca787696ee6 /billing_inc.php: HTML Validation18:14
CIA-11Mantisbt: paul * r8b21ab5870a8 / (6 files): fix invalid html (extra ") and suppress html validation errors for missing alt tags.18:14
* daryn is away: Gone away for now18:28
* daryn is back.18:34
paulr_daryn: did you try with lower limit?18:35
darynnot yet18:35
darynjust back from lunch18:35
CIA-11Mantisbt: jreese master-1.1.x * r72235214532b / (config_defaults_inc.php core/form_api.php): Fix #9999: allow form security to be disabled for sites that use 'bad' proxy servers.18:45
paulr_bad proxy servers?18:47
paulr_what's this? :)18:47
CIA-11Mantisbt: jreese * r10040dee7386 / (config_defaults_inc.php core/form_api.php): Fix #9999: allow form security to be disabled for sites that use 'bad' proxy servers.18:47
nuclear_eclipsepaulr_: some proxy servers ignore the caching/modified headers that we send, which breaks CSRF protection because the proxy serves stale pages, so this patch allows people to disable the protection to work around terrible proxy servers....18:48
paulr_are you sure we are sending the right headers? :)18:48
nuclear_eclipseyes, we've been through this all by now...18:49
nuclear_eclipseproxy servers like Squid obey the headrs appropriately, but some do not18:49
* nuclear_eclipse uses a Squid proxy every day18:50
paulr_;)18:50
* paulr_ enjoys headers :P18:50
paulr_do we know what (broken) proxies users are hitting ?18:53
nuclear_eclipsevb123: do I really need to document a configuration option that shouldn't be touched? :P18:54
paulr_no18:54
nuclear_eclipsepaulr_: I don't know off hand, but everyone that has complained about #2800 errors in 1.1.6+ has said that they are behind a proxy...18:54
paulr_1.1.6 issue ?18:55
paulr_only18:55
paulr_not 1.218:55
nuclear_eclipsewell, 1.1.6+ includes 1.2.x ;)18:55
nuclear_eclipseie, ever since we fixed caching headers for 1.1.6 and 1.2.0a3, the only people still having #2800 errors are behind a proxy, everyone else has had the problem fixed18:56
* nuclear_eclipse is off to class19:06
nuclear_eclipsecheers all19:06
* daryn_away is away: Gone away for now19:33
* daryn_away is away: Gone away for now19:33
* daryn_ is back.20:11
* paulr_ strokes daryn_20:12
daryn_uhm...20:12
paulr_did you get a chance retest with low memory?20:13
* paulr_ goes back to game of dota - cya in 45 minutes :)20:13
daryn_not yet.  i reloaded database and reset memory but had a meeting...now to rerun20:13
paulr_right20:47
daryn_paulr_ the upgrade completed successfully with 64M memory.  Took awhile though.  I have NOT checked that the dates are accurate.21:07
paulr_cool it works :)21:09
paulr_at some point could you check whether one bug's dates are21:09
paulr_and try changing date21:10
paulr_:)21:10
paulr_hmmm21:41
paulr_nuclear_eclipse: i dont get 999821:41
darynpaulr_ the dates seem to be ok.  are they currently ignoring time zones?21:49
paulr_yes/no21:50
paulr_if you go to user preferences21:50
paulr_my account -> preferneces and change timezone21:50
darynmy time zone was set weird...so i changed it21:50
darynbut it didn't change what is displayed in the bug21:50
paulr_set weird?21:51
darynwrong21:51
daryni don't remember setting it at all21:51
paulr_it's "set" to what server is21:51
daryni think it just had the first item selected21:51
darynno...it wsa something in africa21:51
paulr_ahh21:51
darynwhich the server is definitely not set to21:51
paulr_if you change what timezone you have and save21:51
paulr_for me it was updating21:51
paulr_(but at the same time, not quite worked out where21:52
paulr_also21:52
paulr_did you update to my ltest copy of branch right?21:52
darynyes21:52
darynwhen we were discussing earlier21:52
paulr_@@ -150,6 +150,7 @@ function user_pref_cache_row( $p_user_id, $p_project_id = ALL_PROJECTS, $p_trigg21:52
paulr_        $g_cache_user_pref[$p_user_id][$p_project_id] = $row;21:52
paulr_+       date_default_timezone_set( $row['timezone'] );21:52
paulr_        return $row;21:52
paulr_ }21:52
paulr_atm, i just randomnly set timezone when it laodsd userpref's21:53
darynrandomly?21:53
paulr_in a random place21:53
paulr_as opposed to a random value21:53
darynwhat about date custom fields?21:53
paulr_but when i was changing what i had under preferences21:53
paulr_top of screen that shows date21:54
paulr_was immediately showing different timezone21:54
* paulr_ hadnot thought about date custom fields21:54
darynso...top of screen shows correctly now...i don't recall whether it was correct timezone or not before.  The time/date was correct21:55
darynactually...i can figure that out. just a sec21:55
paulr_so whatever top of screen says21:55
paulr_should be timezone you are now in21:55
paulr_I dont like preferences api atm21:55
* paulr_ wants to bugobjects it ;p21:55
paulr_as from what I can tell21:55
paulr_its broken ;/21:55
nuclear_eclipsepaulr_: what don't you "get" about bug 9998?21:56
mantisbotBug 9998 - jburt - open - feedback21:56
mantisbotExport functions generate bad URL - http://www.mantisbt.org/bugs/view.php?id=999821:56
paulr_it only breaks with projects with spaces in?21:57
Weplhi, is it possible to change the config so that reporters can close their own issues?21:58
paulr_nuclear_eclipse: would you agree preferences are 'broken'?21:58
nuclear_eclipsepaulr_: it's not the spaces that break 9998, it's the change in caching headers, and IE being retarded/braindead21:59
paulr_why doesn't it happen to me then21:59
paulr_or is it only ie6?21:59
nuclear_eclipseit's only IE622:00
nuclear_eclipseeg, IE6 is too dumb, and when you choose "Open With", it will delete the file before the app gets a chance to read it22:01
paulr_right22:02
paulr_anyway22:02
paulr_userpreferences22:02
paulr_reset is broken yes/22:02
paulr_did you get a chance to look at where I was going with bug api branch ?22:02
paulr_can I do similar to user prefs :)22:02
nuclear_eclipseI haven't looked, no22:03
paulr_could you have a quick scan of bug api? ;p22:03
paulr_before I do the same to preferences22:03
nuclear_eclipseI'm still in the "I have part of a term paper due" mode...22:04
paulr_ahh22:04
darynpaulr_ it takes a very long time to run this upgrade22:09
Weplby22:09
paulr_daryn: sure22:10
paulr_that's not really that surprising22:10
paulr_although could probably speed it up22:11
paulr_by changing it to do >1 column at a time22:11
darynso i don't have time zone set in my preferences but it is showing correctly at the top of the page22:11
paulr_e..g it'll do bug_table twice22:11
paulr_once for bug_submitted / row22:11
paulr_+ once for last_update / row22:11
paulr_all of those could get merged into 122:11
paulr_daryn: it should be pulling server timezone22:12
paulr_so now if you change to say london22:12
daryngood.  i just wasn't sure and wanted to recheck22:12
paulr_do we want to try to speed up the upgrade22:12
darynprobably.  i tend to have a much longer script run time allowed than most will.22:13
daryni have to run.  i may be on after a bit22:13
paulr_well, install.php calls @set_time_limit( 0 );22:13
paulr_which basically means we should check if users are running safemode ;p22:14

Generated by irclog2html.py