| ../irclogs/#mantishelp.2010-01-02.log | ||
| --- scribe started --- | 00:00 | |
| nuclear_eclipse | hi paul__ | 00:32 |
|---|---|---|
| paul__ | i'm really starting to hate ldap | 00:32 |
| paul__ | so you'll like this | 00:32 |
| paul__ | ldap_auth atm | 00:33 |
| nuclear_eclipse | lol, I still cant even wrap my head around it | 00:33 |
| paul__ | does a search for ldap entries matching the ldap username | 00:33 |
| paul__ | if there's more then one | 00:33 |
| paul__ | it tries to authenticate against each one | 00:33 |
| paul__ | let's assume for a momenth that the ldap username isn't unique | 00:33 |
| paul__ | and consider how | 00:33 |
| paul__ | we get other ldap data | 00:34 |
| paul__ | e.g. realname | 00:34 |
| paul__ | I'm going to make the assumption that username is actually unique | 00:34 |
| paul__ | as otherwise | 00:35 |
| paul__ | what we really should be doing | 00:35 |
| paul__ | is creating an plugin_ldap table | 00:35 |
| paul__ | with fields of like... | 00:35 |
| paul__ | mantisuserid, username, DN, email... | 00:35 |
| paul__ | where mantisuserid,DN are unique | 00:35 |
| paul__ | even if username isn't | 00:36 |
| paul__ | and we really need that table ;/ | 00:40 |
| paul__ | as we shouldn't really be updating our user database with ldap stuff | 00:41 |
| paul__ | as doing that stops us from allowing people to have flexibility to override the ldap data | 00:41 |
| * nuclear_eclipse is cl;ueless | 01:04 | |
| paul__ | it's just set up in a way to be a pain ;/ | 01:04 |
| paul__ | nuclear_eclipse: btw, have you run check.php in trunk recently? | 01:09 |
| paul__ | give it a run :P | 01:09 |
| paul__ | and tell me what happens | 01:09 |
| paul__ | then i'm off to bed | 01:10 |
| * paul__ goes to bed | 01:13 | |
| paul__ | anyway | 01:13 |
| paul__ | you will notice that admin/check.php | 01:13 |
| paul__ | generates a warning about an obsolete config option | 01:13 |
| paul__ | that has been added back in | 01:13 |
| nuclear_eclipse | paul__: | 01:14 |
| paul__ | . | 01:14 |
| nuclear_eclipse | can you look at a patch real quick? | 01:14 |
| paul__ | k | 01:14 |
| paul__ | can you quickly run admin/check.php on your instance as wel | 01:14 |
| nuclear_eclipse | http://mantis.pastebin.com/m6cdf6f64 | 01:15 |
| nuclear_eclipse | sure | 01:15 |
| paul__ | i've got a branch where i've done that | 01:15 |
| nuclear_eclipse | http://bugs.steelwow.com/mantis/admin/check.php | 01:16 |
| paul__ | mm | 01:16 |
| paul__ | you dont get an obsolete warning | 01:16 |
| paul__ | mmmm | 01:16 |
| nuclear_eclipse | that's running 1.2.x tho | 01:16 |
| paul__ | http://bugs.steelwow.com/mantis/admin/check.php?showall=1 | 01:16 |
| paul__ | look at the first entry on that list | 01:17 |
| paul__ | (with the incorrect 'good') | 01:17 |
| nuclear_eclipse | ah | 01:17 |
| paul__ | - print_test_warn_row( $t_description, $t_info, false ); | 01:17 |
| paul__ | + print_test_warn_row( $t_description, false, $t_info ); | 01:17 |
| nuclear_eclipse | anyways paul__ if that patch looks good, I'm ready to push | 01:17 |
| paul__ | the fix for that is in trunk | 01:18 |
| paul__ | i'd kinda rather you didn't ;/ | 01:18 |
| nuclear_eclipse | oh? | 01:18 |
| paul__ | as i've got that patched in a branch | 01:18 |
| paul__ | where I was trying to fix columns properly :( | 01:18 |
| nuclear_eclipse | well are you actually gonna get that pushed upstream anytime soon? | 01:18 |
| paul__ | we can't store the bug fields in config_defaults | 01:18 |
| paul__ | as if you turn off a field | 01:18 |
| paul__ | it picks it up as a change and will store user columns in config db even if they are the default | 01:19 |
| paul__ | i'm trying ;/ | 01:20 |
| paul__ | i've not been feeling that great + xmas + pc hdd playing up | 01:20 |
| paul__ | on top of that, whenever I start looking at something | 01:21 |
| paul__ | I end up finding all sorts of crap that's wrong ;/ | 01:21 |
| paul__ | for example | 01:22 |
| paul__ | as part of looking at the constants, then trying to make it not save to DB when there's no column change | 01:22 |
| paul__ | (i.e. effectively build a list of default columns) | 01:22 |
| paul__ | you realise that we have 2-3 overlapping api's that do 'similar' things e.g. helper_get_columns to view and the columns api stuff | 01:23 |
| nuclear_eclipse | paul__: maybe you try to focus on making small incremental changes and pushing those before getting side tracked down more rabbit holes.. :P | 01:23 |
| paul__ | take a look at custom_function_default_get_columns_to_view | 01:23 |
| paul__ | can you spot what's wrong with it? :) | 01:24 |
| nuclear_eclipse | trust me, I've already looked at all those when adding in filter and column support for plugins | 01:24 |
| paul__ | you didn't spot whats wrong with it then :P | 01:24 |
| paul__ | it's a 12 line function with 4 mistakes in | 01:24 |
| paul__ | spotted? :P | 01:25 |
| nuclear_eclipse | paul__: I usually try to assume that functions are doing things for a reason, rather than looking for bugs when I'm just seeing how they work ;) | 01:25 |
| paul__ | function config_get( $p_option, $p_default = null, $p_user = null, $p_project = null ) | 01:26 |
| paul__ | we call config_get with params wrong around ;/ | 01:26 |
| nuclear_eclipse | that's not something I would have ever considered looking for... | 01:26 |
| paul__ | it's also why column constants didn't get pulled out :) | 01:27 |
| paul__ | I can't recall now whether I modified higher up chain then ended up debugging | 01:28 |
| paul__ | or modified, tested then on off chance spotted that and realised my modifications/testing was invalid | 01:28 |
| paul__ | anyway, it's 01:28 i reallyneed to go sleep | 01:29 |
| paul__ | I'm trying to get ldap/columns stuff I have merged in, but I also don't really want to break it 10 mintues before 1.2.0 | 01:30 |
| paul__ | (at the same time, I consider ldap support storing md5 password as a security risk) | 01:31 |
| paul__ | nn | 01:31 |
| paul__ | nuclear_eclipse: http://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=8971ecfeb3f09ccbd46a222a962ae81c0a597107 | 14:28 |
| paul__ | can you recall why? | 14:28 |
| nuclear_eclipse | paul__: just woke up | 14:42 |
| nuclear_eclipse | gimme a few and ill take a look | 14:42 |
| * paul__ wonders if user_get_avater should be a custom function | 15:54 | |
| paul__ | siebrand: lo? | 18:37 |
| siebrand | paul__: sup? | 18:37 |
| paul__ | you repro'd http://www.mantisbt.org/bugs/view.php?id=11153 ? | 18:37 |
| siebrand | paul__: nope. | 18:42 |
| siebrand | paul__: just marked it critical for release because <some core developer> mentioned it as to be fixed before release. | 18:43 |
| paul__ | personally | 18:43 |
| paul__ | I can download >5MB files | 18:44 |
| siebrand | paul__: then the report may be invalid... | 18:44 |
| paul__ | we call file_get_contents( 0, 512kb) | 18:45 |
| paul__ | then we add chunksize | 18:45 |
| paul__ | wow | 19:29 |
| paul__ | php 5.1.6 is so old ;/ | 19:29 |
| paul__ | nuclear_eclipse: ? | 22:24 |
| nuclear_eclipse | hi paul__ | 22:24 |
| paul__ | out of interest, can you repro the download thing? | 22:25 |
| paul__ | I think i've fixed 1 'bug' in the code anyway | 22:25 |
| paul__ | albeit I dont think that particular code can get hit | 22:25 |
| * paul__ doesn't have php5.1.6 | 22:25 | |
| giallu | paul__, re 5.1.6, That guy with Fedora 6 should really really upgrade, but, if the bug is reproducible in a CentOS 5 installation (5.1.6 with a number of backported patches) I'd say we need to fix it | 23:16 |
| giallu | paul__, if you need to test, use http://mantisbt.org/bugs as it runs that php package | 23:18 |
| paul__ | that's not set up for uploading files | 23:24 |
| paul__ | iirc | 23:24 |
| paul__ | iirc, I already looked | 23:24 |
| * paul__ sighs | 23:27 | |
| paul__ | now i need john to help me fix my local git repo ;/ | 23:27 |
Generated by irclog2html.py