Thursday, 2010-07-01

../irclogs/#mantishelp.2010-07-01.log
--- scribe started ---00:00
hellonullhey all. working on configuring my first mantis installation and need a bit of help. i'm looking to disable the "reproducibility," "severity," and "steps to reproduce" fields since they aren't applicable to the purpose for which we are using mantis. is there a way to take care of each of these with a single config option? for severity, i came across $g_severity_enum_string... i tried values of "", '', and NULL, but the field's dropd01:31
darynhellonull: are you using 1.2?01:34
hellonulldaryn: yes, 1.2.102:56
darynthere are several variables to select which fields to display on which pages but I don't think there is a way to completely disable a field in all locations with one change03:04
darynthe easiest is probably to set these g_bug_report_page_fields, g_bug_view_page_fields, g_bug_print_page_fields, g_bug_update_page_fields,g_bug_change_status_page_fields03:06
darynthe complete list of fields for each page is in config_defaults_inc.php03:07
daryncopy the variables into config_inc.php and remove any fields you don't want03:07
darynconfig_inc.php overrides values in config_defaults_inc.php03:07
hellonullsounds good, thanks daryn03:18
darynyw03:18
davidincHi08:05
dhx_mhi12:12
nuclear_eclipsehowdy dhx_m12:22
dhx_mnuclear_eclipse: howdy12:22
nuclear_eclipseseems like paulr is our only dev who doesn't like github...12:24
dhx_m"paulr> using github probably helps as git.mantisbt.org can be less stable"12:24
nuclear_eclipseyeah, that made me laugh12:24
dhx_mit seems he's for :)12:24
dhx_mbut I think his concern was that people who had checked out the repositories from git.mantisbt.org will be left in the dark if they don't know about this change12:25
nuclear_eclipseyeah, be he's also for moving *everything* to his mantisforge system, too...12:25
nuclear_eclipsethat's possible, but that's what announcements are for :P12:26
dhx_mI'd actually prefer that approach if we were a larger project with 3-4 people minimum who were interested in actively looking after load balanced infrastructure12:26
nuclear_eclipseheh12:26
dhx_mbut we don't have the manpower to setup 2-3 servers in a cluster and maintain them through faults, upgrades, etc12:27
dhx_mmy main concern is that sourceforge changes something and we don't like the change12:27
dhx_moops github12:27
dhx_mwe lose control but gain reliability12:28
nuclear_eclipsewell, they've been continually pushing their support of open source projects, so I can't really see them doing anythign detrimental, and if anything, we can always just set up our own gitweb again :P12:28
nuclear_eclipseor *shudder* set up gitorious...12:29
dhx_mno thanks :D12:30
nuclear_eclipsehehe12:31
dhx_mnote to self: don't watch jquery project on Github... SNR is low12:31
nuclear_eclipsehaha12:32
nuclear_eclipseI laughed when I saw that in my feed12:32
nuclear_eclipsethere are very few projects that I watch on github, and they're all low traffic, which is the real reason I watch them, so I know when they've actually done something :P12:33
dhx_mand don't follow daryn because I'll get 1001 forked repository messages :p12:33
dhx_mhaha12:33
nuclear_eclipseI gotta head to work, but while I'm omw, Victor also mentioned interest in using github to replace the mantisforge repos -- can you ponder on ways to try and keep track of all of them without needing to fork/host repos under the mantisbt organization?  I'd like to keep the org for officially supported repos/plugins12:36
dhx_mI guess Github doesn't support that yet so I think it'd be better to stick with mantisforge for now12:37
dhx_mwith an aim towards having a system where good quality plugins can become "first class citizens" of the MantisBT org?12:38
dhx_mie. github repository under the MantisBT name, project on our bug tracker, etc12:38
dhx_mwe'd need to write up some documentation on that idea though (what expectations are of plugin authors, etc)12:39
dhx_mthe main problem I forsee with that approach is the possibility that a plugin author has ownership over their plugin once it's considered part of the MantisBT org12:40
dhx_mmuch like how Wikipedia tries to avoid having people "own" articles12:40
dhx_mso that anyone can improve a plugin (or in Wikipedia's case, an article) if there is consensus behind the change (rather than have a benovolent dictator approach)12:41
* nuclear_eclipse nominates dhx_m to write up that process for official review :P13:11
dhx_mI'm up for that :)13:11
nuclear_eclipsewe need a new wiki13:11
dhx_mI'll place it at position 150 on the agenda :p13:11
nuclear_eclipseone that doesn't suck ass13:11
nuclear_eclipselol13:11
dhx_midd13:11
dhx_mI really like this website: http://www.mono-project.com/Main_Page13:12
nuclear_eclipsehmm, we could use the Github wiki :P13:12
dhx_mit's MediaWiki... but it doesn't look like it :)13:12
nuclear_eclipsehmm, that oes look like a pretty nice template13:12
dhx_msteal it please :p13:13
nuclear_eclipselol13:13
dhx_mI couldn't find it anywhere last time I looked13:13
dhx_mperhaps I should ask :)13:13
nuclear_eclipseit's probably a one-off template done just for mono13:13
dhx_mit'd be nice to use as a reference if we were to do a similar thing13:14
dhx_mMediawiki templates are publicly accessible via Special: pages13:14
dhx_mbut that doesn't cover the changes needed to Mediawiki's PHP files (used for templating each style)13:14
CIA-22Mantisbt: s.mazeland master-1.2.x * ra16cd8abcbae / (6 files in 2 dirs): Localisation updates from http://translatewiki.net17:20
paulrmoo17:44
* paulr pokes dhx_m17:44
paulrnuclear_eclipse: I've had an idea regarding github17:58
nuclear_eclipseuh oh18:00
nuclear_eclipse(hi paulr)18:00
* daryn thinks this could be good...18:01
paulrdiner brb18:03
nuclear_eclipse /sigh18:03
* daryn ...again?18:03
paulrbk18:23
paulrthinking of setting up github to mirror whats on mantisforge18:23
paulrthat way18:23
paulrwe get benefits both ways18:23
nuclear_eclipsepaulr: the point is so that github is the primary access point18:24
paulr(we still need to work out how people release plugins18:24
paulrwell18:24
paulrcan't really do that for plugins18:24
nuclear_eclipseif we just have github as a mirror, we might as well not use it at all18:24
nuclear_eclipsewe're not planning to use it for anything except the core mantis repos...18:24
paulryea18:24
paulrhence why I can mirror mantisforge plugins to github18:25
paulras a nice 'backup'18:25
nuclear_eclipseI'm not sure why that would be useful -- they are really discoverable at all on github18:25
nuclear_eclipsearent*18:25
paulryea, i've not played to see how github works18:25
paulrbut I was more thinking of it could be useful to have a backup somewhere18:26
darynwe're just waiting for the proper organization api's from github then write an interface for users to add themselves, teams, and new plugins to a mantis plugins organization18:26
darynthen they can be discoverable and manageable through github18:26
nuclear_eclipsetrying to automate mirroring/syncing between two separate public endpoints just sounds like it will cause norhing but pain and frustration18:28
nuclear_eclipsetheres multiple different ways where it can go wrong without any warning18:29
paulrwell, I've actually found I tend to prefer gitweb's interface for browsing source then github's (for annotating/diffing etc)18:32
paulrbut where one pushes to is irrelevant18:32
paulrall I'd say is if we push'd to github, if we could keep mantis bug tracker pointing at either git.mantisbt or git.mantisforge.org so when we click on a diff it goes to gitweb18:32
paulrthat would be fine with me18:32
paulrbut right now, need to fix some stuff18:33
paulrand try and catch dhx I guess on sunday now18:33
nuclear_eclipsewell, I was going to update the mantisbt.org tracker so that everything points to the github repos18:34
nuclear_eclipseitd be silly to be inconsistent IMO18:34
paulr:((18:34
nuclear_eclipsewhy use github if you're not going to have everything point to the same place?18:34
paulrwell, lets not bother then :)18:35
paulrI dont really see point of changing18:35
nuclear_eclipsewe'd bother because of the benefits I listed in theemail18:35
paulrgiven that mantisbt.org doesn't play up18:35
paulrgit's distributed18:35
nuclear_eclipseso then why do you care if we change wher the canonical repo is hosted at?18:36
paulrI prefer the speed of http://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=a16cd8abcbaeca39fcc6ec57a6fd8d159f8f017918:36
paulrit's cleaner, and loads faster18:37
paulrdespite the server being in the US18:37
nuclear_eclipsethe point is to offload the repo management and collaboration to a site that simplifies all of that18:37
paulrwell, I'd still carry on trying to improve repo management for plugins on mantisforge18:37
paulrand mirroring core repo18:37
paulrso all that github would be is a push location18:38
paulrthat we can't then control as easily18:38
nuclear_eclipsebut we can18:38
paulrfor example18:38
nuclear_eclipseit's easier to control a github repo than to control a gitweb repo18:38
paulri'd like to build a release tarball of each commit in trunk so that if users grab a tarball18:38
paulrthey always get docs etc18:38
nuclear_eclipsethat's what we already do with the nightlies paul18:39
paulryea18:39
nuclear_eclipseyou're a bit late to that party18:39
paulrbut we don't keep them for ever18:39
paulrnor link them to specific commits18:39
nuclear_eclipsewhy would we need to keep them forever?18:39
paulrwell, a while at least18:39
nuclear_eclipsewho wants to download a tarball of an old commit, unless it's a release?18:39
paulralso18:41
paulrmore to point is github reliable18:42
paulrI just put into github search 'mantis'18:42
paulrclicked first link which goes to : http://github.com/jreese/mantisbt/tree/9f8516e83a9871077ea98ce2df2db8659b896667/core/classes/MantisCorePlugin.class.php18:42
paulrand got:18:42
paulr"That page doesn't exist!"18:42
nuclear_eclipsethat's because I deleted that repo18:43
nuclear_eclipseI'd imagine their search results get cached for a while, so they haven't flushed old data yet18:43
paulrwhen did you delete it?18:44
nuclear_eclipsewhen I search for mantis, I get victor's mantis-plugins as the top result18:44
nuclear_eclipseyesterday18:44
paulrso 24 hours?18:44
nuclear_eclipseactually, when I search for mantis, I get mantisbt/mantisbt (the new repo) in the result set at #418:45
nuclear_eclipseyes, < 24hours18:45
paulrtbh, i'd rather we just pointed bugtracker to the git.mantisforge.org and used git.mantisbt.org as central point or whatever18:46
paulrthen we can build git/testing/plugins into one place18:46
nuclear_eclipsethat doesn't even make any sense18:47
paulrrather then having one place for viewing git and one place for test results18:47
paulri've got 3 weeks off in a month when I plan to start pushing mantisforge and fixing automation issues18:49
paulror well when I say that18:50
paulrsetting up mantisforge to run your release tarball scripts etc18:50
nuclear_eclipseI'm still not sure what you think you're getting at that we don't already have with the nightly builds...18:50
nuclear_eclipsethe nightly builds already compile docbook and everythign18:51
paulryes, but if we are doing testing/continous integration18:51
paulrshould build tarballs as part of that also18:51
nuclear_eclipsewhatever, I'm still not sure what any of that has to do with why you don't like the idea of using github...18:52
paulrwe could take some load off of mantisbt.org box as that's apparently struggling :P18:52
nuclear_eclipseit's not struggling at all18:52
paulras I say, I prefer working with the simplicity of gitweb's user interface18:53
nuclear_eclipseso you can mirror the github repo to mantisforge and use the mantisforge interface; just because we switch to github doesn't mean you can't keep a mirror18:54
paulrhence why when I normally resolve bugs I normally include a link to git.mantisbt.org in the resolution18:54
paulrfor the patch18:54
paulras I said, don't mind core going to git hub18:54
paulrbut it would be good if bug tracker carried on pointing at a gitweb interface18:55
paulrso I can just click on links to patches18:55
nuclear_eclipsethat wouldn't make ansy sense at all paul18:55
paulrwhy?18:55
nuclear_eclipseif we have github as our primary repository, we're going to have the tracker point at github urls18:55
paulryou wanted github to ease admin18:55
nuclear_eclipsebecause it makes no sense to have the tracker point at some repo host that's not where our primary repo is hosted at18:55
paulrthen in which case, i'd rather not use github18:56
nuclear_eclipsenevermind paul18:56
paulras I tend to waste more time trying to find stuff on it18:56
paulrI'd rather invest my time to write a bot to post a bugnote to a bug linking to a gitweb interface when a commit references a bug number then try to compare diffs using github18:57
nuclear_eclipseyou're totally insane18:57
paulrincidently, since you typed:18:57
paulr<nuclear_eclipse> I'm still not sure what you think you're getting at that we don't already have with the nightly builds...18:57
paulri'm still waiting for a page to load looking at 'Loading commit data...'18:57
paulrso i've made 10 webrequests so far to github in last 48 hours18:58
nuclear_eclipsewhat?18:58
paulrand 3 have failed18:58
paulr;)18:58
paulrwhen I go to view a repo18:58
nuclear_eclipseyou're the only one I know of who can find so many "errors" with github18:58
paulrthere's a column that says 'message'18:58
paulrwhich shows in it 'Loading commit data... ' and an hour glass18:58
nuclear_eclipseI've never seen an error message from them in the two years I've been using them18:58
paulrmaybe i've just picked a bad day or something :P19:00
paulrbut so far i'm being really unlucky :P19:00
nuclear_eclipseare you *trying* to break it?19:00
paulrnot really19:01
paulrI went to github.com19:01
paulrput 'mantisbt' in search and clicked the top link for first test19:01
paulrI then went back and looked down list for more sensible link19:01
paulrand it's not loading the ajax commit messages in19:01
paulrDon't really think that's trying to break it :)19:02
paulrI could understand if I was trying to open a filename with a ' in or something :P19:03
nuclear_eclipseit loads everything for me almost instantly...19:03
paulr;/19:03
darynoh I see that on the .gitignore file19:04
paulrdaryn: yes, that's where I see it19:04
paulrand then it shows nothing in messages for rest of page19:05
paulras I think it loads them in order?19:05
* nuclear_eclipse is not seeing it at all19:05
darynbut only for my forked repo and only for that file. all the rest of mine load19:05
* paulr is glad he's not only one19:05
paulrhmm19:05
nuclear_eclipseI know they do eavy caching, so unttil the data has been pulled into cache, sometimes it can take a bit longer, but it still isn't forever19:06
paulranyway, If you change where we push to19:06
paulrjust add my ssh key19:06
paulrfrom existing box19:06
nuclear_eclipsepaulr: that's the point, using github, I don't have to be the one to deal with everyone's ssh keys19:06
nuclear_eclipseI just add somebody's github account to the team19:06
nuclear_eclipseand then the user can manage their own ssh keys19:07
paulrmaybe in 3 weeks we can try and plan something sensible out19:10
paulrwould be nice to link test results into mantis bugtracker19:11
paulrwhen a commit is added19:11
paulrI proably need to explain my point more there at some point :P19:11
paulrnuclear_eclipse: will you be around when dhx is on sunday?19:13
nuclear_eclipseno clue19:13
nuclear_eclipseyou can always ping me though, and see if I respond19:14
paulrever used app-v btw?19:14
nuclear_eclipsenever heard of it19:14
paulrMS windows tech19:14
nuclear_eclipsethat would be why19:15
nuclear_eclipseI don't run Windows19:15
paulrnuclear_eclipse: still there?21:37
nuclear_eclipseI am now22:28
nuclear_eclipsepaulr:22:28
paulrI dont get the github's branch stuff22:30
paulrif someone has a fork22:30
paulrwhen it shows branches22:30
Githubmantisbt: master-1.2.x Siebrand Mazeland * a16cd8a (6 files in 2 dirs): Localisation updates from http://translatewiki.net - http://bit.ly/amn4e122:30
paulris that forward/behind the fork or the non-fork?22:30
nuclear_eclipsehuh?22:30
paulrhttp://github.com/redhog/pad/tree/postgres22:31
paulrthat's 300 behind but 19 ahead22:32
paulris that based on where the fork is22:32
paulror where the original is?22:32
paulras it's a branch of a fork22:32
paulrbtw22:32
paulr* Github (~Github@sh1-ext.rs.github.com) has joined #mantishelp22:32
paulr<Github> mantisbt: master-1.2.x Siebrand Mazeland * a16cd8a (6 files in 2 dirs): Localisation updates from http://translatewiki.net - http://bit.ly/amn4e122:32
paulr* Github (~Github@sh1-ext.rs.github.com) has left #mantishelp22:32
paulrthat needs to sit on channel22:32
paulrto not be annoying22:32
nuclear_eclipseprobably based on where that branch is compared to the original fork's version of that branch22:32
nuclear_eclipseI can set it up to use our existing CIA project22:33
paulri'm still not sure tbh what we're trying to fix :P22:34
paulrnot like people change ssh keys every week after all22:35
nuclear_eclipsewe're not *fixing* things paul; we're taking advantage of some things that github automates for us22:35
nuclear_eclipseand other various benefits that we don't have from using our own home-rolled services22:36
nuclear_eclipsebtw paulr, if you find join/part info to be annoying, you should just have your IRC client filter those out22:41
nuclear_eclipse90% of the time this channel is just a series of join/part messages anyways...22:41
paulryea22:41
paulrthat's annoying too22:41
paulrpeople should run shells/bnc's ;)22:42
nuclear_eclipseso filter them out if they annoy you -- any client worth its salt will do that for you22:42
paulrbut join/parts are useful for seeing when daryn or whatever times out22:42
nuclear_eclipseso stop complaining! :P22:43
paulrthings that join, 'advertise' a url and part tend to hit my ircop spam filters at some point22:43
nuclear_eclipseso add an exception to ircop...22:43
paulri.e. at some point it'll probably trigger an auto ban22:44
paulrIt didn't then though22:44
* paulr has spent too long diffing java app makefile tonight22:47

Generated by irclog2html.py