*** Quits: dhx1 (~anonymous@60-242-247-232.static.tpgi.com.au) (Remote host closed the connection) | 02:22 | |
*** Joins: dhx1 (~anonymous@60-242-247-232.static.tpgi.com.au) | 02:27 | |
*** Joins: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) | 06:10 | |
dhx1 | Paul_46: ping | 06:36 |
---|---|---|
Paul_46 | . | 06:37 |
dhx1 | the vulnerability report sent through today is a waste of time from what I can see | 06:39 |
dhx1 | bogus non-issues | 06:39 |
Paul_46 | where? | 06:39 |
dhx1 | paul@mantisforge.org | 06:39 |
Paul_46 | waste of time :) | 06:42 |
dhx1 | yea | 06:43 |
Paul_46 | I stripped out null bytes from gpc stuff in mantisbt2 with https://github.com/grangeway/mantisbt/commit/f3c032f8e46abce74dd3ebf0a50e0cbbd677653b | 06:44 |
Paul_46 | back in 10 | 06:44 |
Paul_46 | oh well | 06:45 |
Paul_46 | the nusoap stuff needs dumping | 06:45 |
Paul_46 | we've already removed prototype i think | 06:45 |
Paul_46 | database api got rewritten | 06:46 |
Paul_46 | adodb is dead | 06:46 |
Paul_46 | so that basically leaves that import [a plugin] can upload files | 06:47 |
Paul_46 | and back | 07:00 |
dhx1 | yeah | 07:07 |
Paul_46 | I replied | 07:12 |
Paul_46 | but dhx | 07:12 |
Paul_46 | this is where i'm hoping I can use objects to tidy things up | 07:13 |
Paul_46 | and exploring that angle | 07:13 |
Paul_46 | as atm | 07:13 |
Paul_46 | we have 200000000 functions that do random things | 07:13 |
Paul_46 | some expect valid input | 07:13 |
Paul_46 | some check | 07:13 |
Paul_46 | and it's not consistent :) | 07:13 |
Paul_46 | dhx1: I looked at implementing/using zeta's mailer component instead of phpmailer - so we can do html mails etc easily and have fewer dependencies- https://github.com/grangeway/mantisbt/commit/9e13b3b0035f6fe481fb32824cb7f449e5af634d | 07:14 |
dhx1 | will check it out | 07:15 |
Paul_46 | whilst doing that I looked at our use of the variable to limit email domains | 07:15 |
Paul_46 | and did two things | 07:15 |
Paul_46 | allow it to take a array of domains not just a single string | 07:15 |
Paul_46 | but also, whilst doing that released that you can signup, update a user or a manager can update a user | 07:16 |
Paul_46 | IIRC, it appeared that the limit of domain's only applied to updating for both user+admin | 07:16 |
Paul_46 | but you could sign up with any domain :) | 07:16 |
Paul_46 | it was either that or the disposable email stuff [which is enabled by default with no config to disable???] has that functionality | 07:17 |
Paul_46 | anyway, this is why i dont mind ignoring 1.2/1.3 atm | 07:20 |
Paul_46 | there's no one deving big features | 07:20 |
Paul_46 | and bugfixes are easy to port if needed later on | 07:20 |
dhx1 | email notifications are only one aspect of many | 07:20 |
Paul_46 | but we need a simpler core we can build on | 07:21 |
Paul_46 | and to me, that means project/bug/user object | 07:21 |
dhx1 | objects are inflexible though | 07:22 |
dhx1 | they're OK for simple purposes such as editing/viewing a project, user, etc | 07:23 |
dhx1 | but useless otherwise | 07:23 |
Paul_46 | well, yea | 07:24 |
Paul_46 | however in any case, atm, the code base is a mess | 07:24 |
Paul_46 | soap api has different security checks to core | 07:24 |
Paul_46 | it's just bugs waiting to happen | 07:24 |
dhx1 | yes | 07:24 |
Paul_46 | so if making users,projects,bugs an object - and leaving custom fields etc 'as is' a bit more | 07:25 |
Paul_46 | and then making a mobile/api/web interface use those objects for the simple purposes of viewing a bug | 07:26 |
Paul_46 | means we get less code and less bugs | 07:26 |
Paul_46 | the one thing I want to look at coding is an open source mobile interface for mantisbt2 | 07:27 |
dhx1 | hmmm | 07:27 |
Paul_46 | charging people for mantis touch i disagree with | 07:27 |
dhx1 | sounds like REST (or other object get/set architecture) | 07:27 |
Paul_46 | I did look at rest other week ;p | 07:28 |
Paul_46 | I was actually looking at what redmine do with http://www.redmine.org/projects/redmine/wiki/Rest_api | 07:28 |
dhx1 | I don't like REST | 07:28 |
dhx1 | it's trying to force something complex into something simple | 07:29 |
dhx1 | and causing a major mess in the process | 07:29 |
Paul_46 | well, i'd like a web interface where basically | 07:29 |
Paul_46 | we have a user object | 07:29 |
Paul_46 | we use say zend's webservices stuff | 07:29 |
Paul_46 | that can implemenet soap/xmlrpc/whatever | 07:29 |
Paul_46 | but as we control the 'object', the soap/xmlrpc thing is looser | 07:30 |
dhx1 | hmmm | 07:30 |
Paul_46 | then basically generate an auth key for the soap/xmlrpc layer [seperate to user/pass] that says "this key can update/edit this type of object(s)" | 07:30 |
dhx1 | I don't see how OO or REST are required to achieve this result | 07:30 |
dhx1 | the OO/REST approach sounds like Reflection mess | 07:31 |
Paul_46 | sidenote: i'd like a json api so i can display mantis data in a portal using jscript | 07:31 |
Paul_46 | yea sure | 07:31 |
Paul_46 | what I do know is we need to scrap the current webservice api and think about what we want to do and rewrite it | 07:32 |
Paul_46 | and I don't overly care what we need rest/soap/json/xmlrpc etc | 07:32 |
dhx1 | right | 07:32 |
Paul_46 | but just atm, it's buggy | 07:33 |
Paul_46 | yes, it's getting 'better' | 07:33 |
Paul_46 | as people are implementing features to work around bugs | 07:33 |
Paul_46 | but you still have the fundamental issue when i last checked: | 07:33 |
Paul_46 | most of mantis is designed to support multiple projects and customising per project | 07:33 |
Paul_46 | mantisconnect was designed when mantis was single project and didn't allow such customising | 07:34 |
Paul_46 | then extended | 07:34 |
Paul_46 | for example | 07:36 |
Paul_46 | function mc_enum_priorities( $p_username, $p_password ) { | 07:36 |
Paul_46 | priority I believe can be custom to a project | 07:36 |
Paul_46 | and yet the soap api doesn't allow the project to be specified | 07:37 |
Paul_46 | therefore bug | 07:37 |
Paul_46 | therefore the soap interface is wrong | 07:37 |
dhx1 | yeah | 07:41 |
Paul_46 | so for now, i'm just experimenting with core | 07:41 |
dhx1 | fair enough | 07:48 |
Paul_46 | well, my hope would be to come up with something people like | 07:51 |
Paul_46 | :) | 07:51 |
Paul_46 | have you followed what i've been doing? | 07:53 |
dhx1 | as of 2-3 weeks ago, yeah | 07:53 |
Paul_46 | but you dont overly like so far? | 07:54 |
dhx1 | I don't see it being merged with master any time soon | 07:55 |
dhx1 | that is my main concern | 07:55 |
dhx1 | lots of random disruptive changes | 07:55 |
dhx1 | I think we'd need to look at both next and your mantisbt2 branch | 07:56 |
dhx1 | and merge stuff manually one thing at a time | 07:56 |
dhx1 | picking and choosing | 07:56 |
Paul_46 | well, equally, i'm against some of the commits that have gone into 1.2 and 1.3 | 07:57 |
Paul_46 | so really that's a mess :) | 07:58 |
dhx1 | any thoughts on next? | 07:58 |
Paul_46 | I dont like the file system layout in it | 07:58 |
dhx1 | neither do I actually | 07:59 |
dhx1 | it's the Zend approach though | 07:59 |
dhx1 | and that is what I thought you were keen on? | 07:59 |
Paul_46 | I like the exceptions stuff though | 07:59 |
Paul_46 | no, daryn/giallu iirc | 07:59 |
dhx1 | me... I don't like frameworks or being forced into using someone else's standards | 07:59 |
Paul_46 | [albeit, we'd both agreed on the exceptions and namespaces - i'd just gone "no point implementing exceptions until we decide what to throw" | 07:59 |
dhx1 | they never work because "one size fits all" doesn't work either | 08:00 |
Paul_46 | the gettext stuff i'm not sure about | 08:00 |
Paul_46 | albeit, I appreciate you added a php implementation | 08:00 |
Paul_46 | siebrand can support all 3 formats of language files we have so it's probably more a case of deciding what people prefer and use in php | 08:01 |
dhx1 | I had a long chat to Siebrand a month or two ago about gettext | 08:01 |
dhx1 | basically, they don't like it much because people that use it tend not to provide translation hints | 08:01 |
dhx1 | and don't have a proper i18n toolchain configured | 08:02 |
Paul_46 | define 'translation hints' | 08:02 |
dhx1 | I think we can overcome that | 08:02 |
Paul_46 | the only reason I like | 08:02 |
dhx1 | in the current state we have translation lookup keys such as 'bug_view_title' | 08:02 |
dhx1 | with gettext, we use the English translation as the lookup key | 08:02 |
Paul_46 | the only reason I like "lookup keys" | 08:02 |
dhx1 | but that doesn't help the translator know what the message is meant to do | 08:03 |
dhx1 | 'OK' is a vague message to translate | 08:03 |
Paul_46 | is the english translation becomes a translation too | 08:03 |
Paul_46 | i.e. | 08:04 |
dhx1 | Paul_46: there is nothing preventing the creation of a en_US locale | 08:04 |
Paul_46 | 'private_project_msg' => 'This project is private. Only administrators and manually added users have access.', | 08:04 |
Paul_46 | what i mean though is | 08:04 |
Paul_46 | in the example above | 08:04 |
Paul_46 | when someone wants to reword that to | 08:04 |
Paul_46 | Private Project: You have not been granted access to this project." | 08:04 |
Paul_46 | it doesn't become a code change | 08:05 |
Paul_46 | and the lookup for other languages doesn't change whilst people debate the best way to word it :) | 08:05 |
dhx1 | /* L10N: Notification to users when they view the details page of a project which is private and they don't have access to */ | 08:06 |
Paul_46 | as a side note: This project is private. You have not been granted access to this project." | 08:06 |
dhx1 | _('This project is private. Only administrators and manually added users have access.'); | 08:06 |
Paul_46 | is probably a better way to phrase it | 08:06 |
dhx1 | in that example, the string starting L10N: is a translation hint to help translators | 08:06 |
Paul_46 | then 'manually added users' which isn't a grant term :) | 08:06 |
Paul_46 | in your example there, i'd almost rather have those translation hints in the strings file :) | 08:07 |
dhx1 | the thing is though... gettext is a standard toolchain | 08:07 |
dhx1 | translators can use whatever tools they like | 08:07 |
dhx1 | including TranslateWiki | 08:07 |
dhx1 | we don't need to maintain any of our own stuff on top of it... much simpler, more robust and accessible to users and translators | 08:08 |
Paul_46 | apart from not overly like _ as a name of a function | 08:08 |
dhx1 | we can change those if needed | 08:08 |
Paul_46 | and aside from whether i'd overly like /* L10N stuff scattered | 08:08 |
dhx1 | it also adds complex features such as pluralisation support | 08:09 |
Paul_46 | I can see both benefits and downside to having the english strings | 08:09 |
* Paul_46 nods | 08:09 | |
dhx1 | english strings are easy to grep the source code for | 08:09 |
Paul_46 | I just wondering if having the english strings in source | 08:09 |
dhx1 | which is the predominant language we use | 08:09 |
Paul_46 | mean people are more likely to fiddle with them | 08:09 |
Paul_46 | [which could be both a good or bad thing] | 08:09 |
dhx1 | agreed | 08:09 |
dhx1 | but at least those changes become more visible (they're not hidden away on TranslateWiki) | 08:10 |
Paul_46 | 'reauthenticate_message' => 'You are visiting a secure page, and your secure session has expired. Please authenticate yourself to continue.', | 08:10 |
Paul_46 | i'd probably be inclined to add 're' to please authenticate | 08:10 |
Paul_46 | :) | 08:10 |
dhx1 | + we can quickly see where strings are reused multiple times | 08:10 |
* Paul_46 nods | 08:10 | |
dhx1 | anyway, more work is needed | 08:11 |
Paul_46 | the main thing we dont have atm is a plan | 08:12 |
Paul_46 | and I dont see that getting fixed any time soon | 08:12 |
dhx1 | meh, who would follow it? :D | 08:12 |
Paul_46 | given I want to code a ipod interface to mantis and commit it to '2.0' | 08:12 |
dhx1 | yeah | 08:13 |
Paul_46 | but then equally, I dislike people commiting changes to mantis to support their private project to make $$$'s and not doing other stuff | 08:14 |
* Paul_46 shrugs | 08:15 | |
dhx1 | I'd rather focus on code for now and work that out later | 08:16 |
Paul_46 | btw, how do we revert a commit from 1.3? | 08:16 |
dhx1 | which one(s)? | 08:17 |
Paul_46 | you'll love it :P | 08:17 |
Paul_46 | [and i'm actually being serious for a change] | 08:17 |
Paul_46 | https://github.com/mantisbt/mantisbt/commit/839f1d68bc771e579a9f48324624973e0c996ebc | 08:17 |
Paul_46 | that one | 08:17 |
dhx1 | no need to revert | 08:18 |
dhx1 | I want to merge fields and custom fields support together | 08:19 |
dhx1 | they shouldn't be different | 08:19 |
Paul_46 | well only reason I say revert is | 08:19 |
Paul_46 | in the 2.0 branch, I allow the long text in the main field | 08:19 |
Paul_46 | so no need to have seperate | 08:19 |
dhx1 | then revert it in your branch, but not master | 08:20 |
Paul_46 | well it also means we can't do a 1.3 release | 08:20 |
dhx1 | ? | 08:20 |
dhx1 | I know what you're saying | 08:21 |
Paul_46 | https://github.com/grangeway/mantisbt/blob/master-2.0.x/admin/schema.php | 08:21 |
Paul_46 | i collapsed the _Text tables into the bug/bugnote tables | 08:21 |
Paul_46 | so if we do do that | 08:22 |
dhx1 | I know... but I think we'll make a lot more database changes soon | 08:23 |
dhx1 | at least, I want to do that... | 08:23 |
dhx1 | CHECK constraints, referential integrity, transaction-based approach | 08:23 |
GitHub161 | [mantisbt] siebrand pushed 1 new commit to master-1.2.x: http://git.io/y6tN7w | 17:17 |
GitHub161 | [mantisbt/master-1.2.x] Localisation updates from http://translatewiki.net. - Siebrand Mazeland | 17:17 |
*** Quits: giallu (~giallu@fedora/giallu) (Ping timeout: 240 seconds) | 19:37 | |
*** Quits: sdfjkljkdfsljkl (~sdfjkljkd@static.96.23.63.178.clients.your-server.de) (Remote host closed the connection) | 20:00 | |
*** Joins: sdfjkljkdfsljkl (~sdfjkljkd@static.96.23.63.178.clients.your-server.de) | 20:00 | |
*** Quits: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) (Quit: Say What?) | 20:12 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!