Tuesday, 2010-06-29

../irclogs/#mantishelp.2010-06-29.log
--- scribe started ---00:00
dhx_m#xoscommit is really annoying on our ManTweet page ;)00:00
nuclear_eclipseI know there's a bug for the roadmap/changelog issue, but I'm not sure if there's a bug posted for the filter issue00:00
dhx_mI'll try to remember to look into it with daryn's new filter code00:01
dhx_motherwise remind me and I'll do a thorough test00:01
dhx_m(of versions... everywhere :))00:01
nuclear_eclipseyeah, I know daryn's filter changes would make it difficult, so I was just wondering if there might be a quick/dirty fix that we could do just for 1.2.x...00:01
dhx_mwe're still working on 1.2.x?00:02
nuclear_eclipsewell, bug fixes for 1.2.x would be great :)00:03
nuclear_eclipsewe do still need/want to support it00:03
dhx_mbtw I finally upgraded my local install from lighttpd+fastcgi-wrapper+php-cgi (5.2)+xcache to nginx+php-fpm (5.3.3)+pecl-apc... so much faster :)00:04
dhx_mit's time consuming :(00:04
nuclear_eclipseI'd like to at least get a few of these fixes in and push a 1.2.2 release00:04
dhx_mthe problem is that 1.2.x and 1.3.x are already quite different to each other so not much backports cleanly... it requires 2x the work to fix something00:05
nuclear_eclipseeven though we're working on 1.3, people are still using 1.2 (or even 1.1 =\) for everything00:05
dhx_mhow do I reproduce this bug?00:05
dhx_madd an inherited version00:06
dhx_mand see if it shows up in the "fixed in version" (etc) filter dropdowns?00:06
nuclear_eclipseselect the child project, and go to view issues, and the inherited versions aren't shown in the filter dropdowns00:06
dhx_maha00:06
dhx_mare they meant to inherit?00:07
dhx_mthere is no option to toggle inheritance (as there is with categories)00:07
nuclear_eclipsewell, there is an option, but it's a global option00:07
nuclear_eclipseand yes, it's enabled, because we use the inherited versions everywhere00:08
dhx_m$g_subprojects_inherit_categories00:08
nuclear_eclipseyep00:08
dhx_m$g_subprojects_inherit_versions00:08
dhx_mboth default to ON00:08
nuclear_eclipsebasically, the filter dropdowns aren't looking for versions from parent projects00:09
dhx_mif ( $p_project_id == ALL_PROJECTS ) {00:10
dhx_m$t_inherit = false;00:10
dhx_m} else {00:10
dhx_mif ( $p_inherit === null ) {00:10
dhx_m$t_inherit = ( ON == config_get( 'subprojects_inherit_versions' ) );00:10
dhx_mso if you've got ALL_PROJECTS selected on the bugs view page00:11
dhx_myou won't see inheritied versions (there aren't any)00:11
dhx_mfor me, nothing shows up for "Fixed in version" when ALL_PROJECTS is selected00:12
dhx_mseems wrong... I'd expect to see the versions of all projects and child projects, etc00:12
nuclear_eclipseyeah, I think the logic for that might have just been busted altogether or something =\00:13
dhx_mdo categories work ok for you?00:16
nuclear_eclipseyeah00:17
dhx_mah00:19
dhx_mversion_get_all_rows_with_subs has no concept of inherited versions00:19
dhx_mhmmm that didn't work00:26
nuclear_eclipseyeah, I couldn't find the root cause during my initial investigation00:27
dhx_mI suspect it may be the same issue that is affecting the roadmap/changelog00:39
dhx_mversion_get_project_where_clause creates the list of projects from which versions are gathered from00:40
dhx_mand this function calls project_hierarchy_inheritance00:40
nuclear_eclipseyay fun00:40
dhx_mthe where clause is returning projects 0 and 100:40
dhx_mwhen it should be returning 1, 3, 4 in my case00:41
dhx_moh oops00:41
dhx_mI am looking at this the wrong way00:41
dhx_mit works now :)00:41
dhx_m... sort of00:42
dhx_mprojects nested 3 deep don't inherit anything00:42
dhx_mand subprojects aren't included in the version lists00:42
dhx_msigh, project_hierarchy_inheritance isn't even a recursive function00:43
dhx_mhow useless00:44
nuclear_eclipsewelcome to the hell I went through :P00:44
dhx_mlol00:44
dhx_mwhat's your take on disabled projects... should they stop the recursive lookup or should they just be hidden from the total search results?00:49
dhx_mie. ProjectA => ProjectB (disabled) => ProjectC00:50
dhx_mwould the inherited versions list of Project C return Project A?00:50
dhx_mI would have though disabled projects should work recursively too00:50
dhx_m*thought00:50
dhx_mbut that is broken too :p00:51
nuclear_eclipsethey should stop the recursion IMO00:51
dhx_myep00:51
nuclear_eclipseoh, never really thought about how to deal with disabled parent projects...00:51
nuclear_eclipsebluh00:52
nuclear_eclipseyet more hell00:52
nuclear_eclipseIMO that shouldn't be possible, or the disabled-ness should extend to the child projects too, but then a project can have multiple parents, and it all gets insane very quickly00:52
dhx_mindeed00:54
dhx_mREWRITE! :)00:54
nuclear_eclipselol00:54
dhx_mI'll add that to my ever growing list of things to rewrite at some distant stage in the future00:55
dhx_mcustom fields/fields being #1 still00:55
dhx_malthough I think Daryn is interested in that too :)00:55
dhx_mthis code sucks... it should be a simple tree search but isn't...00:56
nuclear_eclipsewell, what makes it worse is knowing that the tree can potentially have loops, or children can have multiple parents... =/00:57
nuclear_eclipseanywho, it's my bedtime00:57
nuclear_eclipsecheers mate00:57
dhx_mI'll have it fixed soon :)00:57
dhx_mcya later00:57
dhx_moh no, this is really broken... the cache isn't recursive either01:00
dhx_mperhaps I will rewrite it now01:00
dhx_mdaryn: hi :)03:36
dhx_mhowdy13:33
nuclear_eclipsehi dhx_m13:36
dhx_mnuclear_eclipse: I've had a look at project hierarchy and I'm working on rewriting it13:39
darynoh goody...13:40
daryn:)13:40
dhx_mmy current thoughts are to implement MantisTree.class.php that implements a tree (computer science terms) and reads/writes to a database using the nested set model13:40
nuclear_eclipseoh dear13:40
dhx_mthen MantisProjectHierarchy extends MantisTree13:40
dhx_mand we can reuse it for MantisVersion, etc13:40
dhx_mif that is needed13:41
dhx_mdaryn: hey :)13:41
daryngdday13:41
nuclear_eclipsedhx_m: KISS13:41
dhx_mwe can't apply a simple fix to the problem you identified before nuclear_eclipse as project_heirarchy_api.php has no concept of 'trees'13:41
dhx_mit assumes there are two levels of projects ONLY13:42
dhx_mit is pretty simple :)13:42
nuclear_eclipsewell, I just don't like having a bugfix turn into a complicated rewrite of a core API... =\13:42
nuclear_eclipsethat's a lot of changed code to push out for a maintenance release...13:43
dhx_mthere is no other way to get nested projects to work for more than 2 layers13:43
daryni'd say leave it broke in 1.2 then13:43
dhx_mwell I might be able to get it to work with 2 layers of projects13:44
nuclear_eclipsewell, IMO if it never worked with big chains before, is it possible to just fix the other bug in 1.2.x, and then do a rewrite for 1.3?13:44
dhx_mwith minimal changes13:44
darynnuclear_eclipse: you ever converted a git repo to svn?13:47
nuclear_eclipsegood haevens no :P13:47
daryni know...13:47
dhx_mgarrgh my ears!13:47
nuclear_eclipseI know two potential methods though13:47
darynwe're using hudson to build mantis with plugins13:47
darynhudson doesn't properly handle multiple git repos13:48
darynbut does svn13:48
nuclear_eclipsea) git can host a cvs server, and you could use SVN to convert the CVS repo...13:48
nuclear_eclipseb) there's a project available that can translate lots of various repo types to one another13:48
nuclear_eclipseI just can't remember the name of b)13:49
daryntailor?13:49
nuclear_eclipseI think it starts with a t13:49
nuclear_eclipseyep13:49
darynnuclear_eclipse: so much for tailor...:(14:54
docgnomeis there a way to make it so reporters can't change the status of a ticket? I'm particular about how tickets are filed and we have a reporter who doesn't bother to read the docs and keeps changing the status15:38
daryndocgnome see Manage/Workflow Thresholds15:40
darynUpdate issue status15:40
darynand Limit reporter's access to their own issues15:41
docgnomeah. thanks. :-\ should have seen that. my bad15:41
darynnp15:41
docgnomehuh. there isn't a checkbox for that option.15:42
docgnomelike i see it in the table, but there are no controls to configure it15:42
docgnomehow about a way to stop it from trying to assign tickets to me when i change the status? I keep forgetting that it puts me in that drop down automatically and keep having to unassign tickets15:43
darynare you administrator?15:45
docgnomeyeaha15:45
darynhm..maybe it's new to dev15:45
docgnomerunning 1.2.115:45
darynfor Limit reporter, right?15:45
docgnomeyeah15:46
darynmine shows at bottom of list15:46
docgnomefound a different way to do it though15:46
docgnomeright, i see Limit Reporter blah blah blah there15:46
docgnomebut no controls15:46
docgnomeif i go to Workflow Transitions though i can set who has permissions to move tickets to what statuses15:46
docgnomeso i just made it so they had to be higher than reporter to move ticket status15:47
darynoh...limit reporter's just makes only see their own tickets15:47
darynthat works15:47
docgnomeright. the second thing is bugging me more though heh15:47
daryni don't know about that one15:47
docgnomedurn.15:49
nuclear_eclipsedaryn: I'm already exploiting github's new "organizations" feature... :P21:51
darynhaven't even looked21:52
darynic...i've been added to mantisbt organization21:52
darynnifty21:52
nuclear_eclipseit gives a "canonical" endpoint for projects to host repos without requiring a shared account21:53
nuclear_eclipseI'd personally like to see Github become the official repo location for Mantis21:54
nuclear_eclipsejust because it has so many nice features, and it makes collaboration so easy21:54
daryni think i'm ok with that21:55
daryni'm using it anyhow21:55
nuclear_eclipseyeah, my biggest21:55
nuclear_eclipseoops21:55
darynwe wouldn't have to worry about mantisforge crashing either21:55
nuclear_eclipsemy concern is that we'd be losing post-commit emails and CIA.vc notifications, or at least not without a bit of custom scripting attached to Github's hooks21:57
nuclear_eclipseoh, hmm, Github supports CIA now21:57
nuclear_eclipsenifty, they've completely revamped their service hooks since I last looked21:58
nuclear_eclipseapparently I could write an "official" service hook that notified my source-integration plugins21:58
nuclear_eclipseI'd just have to learn how to write it in Ruby :P21:58
daryncool21:58
nuclear_eclipsenevermind then :P21:58
nuclear_eclipseit seems the only hurdle is a) getting people to agree, and b) having people update their git remotes21:59
nuclear_eclipsedhx_m: you on board?22:00
nuclear_eclipsesiebrand: ?22:00
darynso...i have a mantisbt fork where i do dev before pushing to the official repo...should that go away and dev on branches in official? or should still use a fork?22:00
nuclear_eclipseI'd still use a fork for anything that's not an officially-supported branch22:00
darynk22:00
nuclear_eclipsethis would certainly make it easier to manage developers who have access to push to the repository22:01
nuclear_eclipseand it would also give an official pull-request mechanism that could send email to the mantisbt-dev list22:02
daryni'm in22:04
nuclear_eclipseok, the IRC notification from Github is pretty nifty22:07
nuclear_eclipse18:06 - - Github [~Github@sh1-ext.rs.github.com] has joined #leetcode22:07
nuclear_eclipse18:06 < Github> leetcode: master John Reese * 5ff0525 (1 files in 1 dirs): Added Piwik stats tracking code22:07
nuclear_eclipse18:06 < Github> leetcode: master John Reese * 550b683 (1 files in 1 dirs): Updated URL to repo on Github22:07
nuclear_eclipse18:06 < Github> leetcode: master commits 950e1d5...550b683 - http://bit.ly/9yjOCd22:07
nuclear_eclipse18:06 - - Github [~Github@sh1-ext.rs.github.com] has left #leetcode []22:07
nuclear_eclipsethe email notification, not so much22:08
nuclear_eclipseemail doesn't contain a diff, but I guess the real question is how much do people care about that or actually look at it?22:08
darynweird...why does it show the official is forked from my fork?22:09
nuclear_eclipsebecause I forked it from mine, and then deleted mine, so yours in the oldest in that forking group22:09
darynah...so i should delete mine too22:09
nuclear_eclipsemeh22:10
nuclear_eclipseyou can if you want, but I wouldn't worry too much about it22:10
darynjust makes me look more important than i am, heh22:10
nuclear_eclipseI like the auto-generated bitly link in the IRC hook, just wish the email hook contained the same thing...22:11
nuclear_eclipseanywho, dinner time22:11
nuclear_eclipsecheers22:11
daryncheers22:11
darynnuclear_eclipse: this notice just popped up on my github page - "Know anyone that prefers Mercurial? Tell them they can use GitHub with our hg-git plugin." Now maybe paulr will be happy...22:33
rb__hello all22:45
rb__any warm bodies here22:45
rb__anybody?23:02

Generated by irclog2html.py