| ../irclogs/#mantishelp.2009-04-08.log | ||
| --- scribe started --- | 00:00 | |
| slestak | i am looking at teh wiki and the manual for info on installing plugins. I see plenty of info for developing plugins, but not installing. | 02:08 |
|---|---|---|
| slestak | i am trying to install cvsimport plugin, and the readme says to "install like typical mantis plugins" lol | 02:09 |
| slestak | wait, think i have too new a version. lemme make sure i have the right csvimport for 1.1.6 | 02:10 |
| nuclear_eclipse | slestak: plugins for 1.1.x are created by a third party, and are unsupported by us - we actually recommend that yo unot run any plugins for 1.1.x because they replace core code and are known to cause problems with upgrading mantis | 02:29 |
| slestak | hmm, can you give me any advice on converting data from my existing bugtracker? | 02:30 |
| nuclear_eclipse | 1.2.x is the first version to have an official plugin system, and it is incompatible with the third party system for 1.1.x | 02:30 |
| nuclear_eclipse | slestak: what bugtracker? | 02:30 |
| slestak | trac | 02:30 |
| slestak | i have not looked at your roadmap very closely. Is 1.2a anywhere near rc or release? | 02:31 |
| nuclear_eclipse | we don't have any official importing system, but you can a) try looking through our tracker and forums to see if anyone has posted a solution, or b) find a way to export data from Trac in XML to match the format expected by the ImportExport plugin included with 1.2.x | 02:31 |
| slestak | is the existing csv import (3rd party) a dangerous hack? | 02:32 |
| nuclear_eclipse | slestak: we are trying to solidify a release, but we have a small dev team with little spare time, so there's no definite timeframe | 02:32 |
| nuclear_eclipse | slestak: I can't comment on the CSV import, as I've never used it | 02:33 |
| slestak | how about this, I may just run these issues out in trac. Can you tell me a way to start my mantis numbering out at an arbitrary value. I'd like to start with a number higher than my existing issue count | 02:34 |
| nuclear_eclipse | slestak: set your database's autoincrement value to what you want it to start at | 02:35 |
| slestak | nuclear_eclipse: thx. makes sense | 02:35 |
| nuclear_eclipse | you're welcome | 02:35 |
| slestak | would you recommend against using 1.2 now? | 02:36 |
| nuclear_eclipse | I would recommend against it for anything highly critical, but we use 1.2.x for our official tracker, and I use it in multiple places of my own without much trouble | 02:38 |
| * nuclear_eclipse is off to bed | 02:39 | |
| nuclear_eclipse | cheers all | 02:40 |
| slestak | gnite | 02:40 |
| [KK]Kirill | hi all. John, you wake up? | 07:38 |
| [KK]Kirill | one question about git | 07:38 |
| [KK]Kirill | How I merge 2 branch with conflict? | 07:39 |
| CIA-18 | Mantisbt: paul * rea8cb633033e /core/filter_api.php: Empty ?> <?php block can be removed. | 10:17 |
| CIA-18 | Mantisbt: paul * rf34ebf2ba7c7 /bug_revision_view_page.php: Update for new top/bot naming | 10:17 |
| fialar | hi.. I want to upgrade an old Mantis from v1.0.7 to v1.1.6. Is there anything special I need to do for the databases? | 10:32 |
| dhx_m | hey paul, just a heads up that I'm getting this error with the latest git master: | 11:43 |
| dhx_m | Parse error: syntax error, unexpected '}' in /var/www/localhost/htdocs/mantis-master/bug_view_advanced_page.php on line 95 | 11:43 |
| paulr | bah ;/ | 11:45 |
| paulr | lets see | 11:45 |
| paulr | dhx_m: i've been trying to set up http://www.mantisforge.org:8080/cruisecontrol/buildresults/mantisbt?tab=metrics | 11:46 |
| paulr | with some 'rules' for codeformatting | 11:46 |
| dhx_m | very nice :) | 11:46 |
| dhx_m | that's a great idea | 11:47 |
| paulr | it also means I break everything | 11:47 |
| dhx_m | yep no problems :) | 11:48 |
| paulr | whats in your branch? | 11:48 |
| paulr | (on mantsiforge) | 11:48 |
| dhx_m | I just thought I'd give you a heads up because it is going to be hard to test all parts of the code | 11:48 |
| dhx_m | I have 1 branch per 'feature' or issue | 11:48 |
| dhx_m | just a few fixes and improvements | 11:48 |
| CIA-18 | Mantisbt: paul * r9d6ac1168534 /bug_view_advanced_page.php: missing ; | 11:49 |
| dhx_m | the most useful one outstanding is probably a new 'stick'/'unstick' button | 11:49 |
| dhx_m | commit messages are fairly in depth if you're interested in checking out the branches :) | 11:49 |
| dhx_m | that commit fixed it :) | 11:50 |
| dhx_m | I'm interested mostly in jreese's commit that added revision control to the description field and bug notes | 11:52 |
| * paulr nods | 11:53 | |
| paulr | i think there's probably some work to do | 11:53 |
| dhx_m | yep, I'm finding a few minor things | 11:54 |
| dhx_m | for instance, the "View Revisions" link shows even when there is only 1 revision | 11:54 |
| paulr | er | 11:54 |
| paulr | view revisions is only in history right? | 11:54 |
| dhx_m | and it seems you can edit a bug note (without changing anything) and it'll record a new revision and history log entry | 11:54 |
| dhx_m | yep | 11:54 |
| paulr | yes, in fact it will | 11:55 |
| paulr | lets see if i'm right in thinking we didn't move something | 11:55 |
| dhx_m | I suppose you add more queries if you start checking how many revisions a note has had | 11:55 |
| paulr | if ( bug_revision_count( $p_bug_id, REV_DESCRIPTION ) < 1 ) { | 11:56 |
| paulr | $t_revision_id = bug_revision_add( $p_bug_id, $t_current_user, REV_DESCRIPTION, $t_old_data->description, 0, $t_old_data->last_updated ); | 11:56 |
| paulr | } | 11:56 |
| paulr | --- | 11:56 |
| paulr | if( $t_old_data->description != $p_bug_data->description ) { | 11:56 |
| paulr | $t_revision_id = bug_revision_add( $p_bug_id, $t_current_user, REV_DESCRIPTION, $c_bug_data->description ); | 11:56 |
| paulr | history_log_event_special( $p_bug_id, DESCRIPTION_UPDATED, $t_revision_id ); | 11:56 |
| paulr | } | 11:56 |
| paulr | that should be within the loop | 11:56 |
| paulr | I spotted that last night when reviewing it | 11:57 |
| paulr | but didn't remember to tell john/remind john :) | 11:57 |
| dhx_m | ;) | 11:57 |
| paulr | try pulling now and seeing if that issue is fixed. | 11:58 |
| CIA-18 | Mantisbt: paul * r7a19a5ced0dd /core/bug_api.php: Fix issue with revisions: we only add the initial revision if we are making a change and adding a new revision. | 11:58 |
| dhx_m | it still shows up the "View Revisions" link next to the history entry for "Bugnote added" | 11:59 |
| dhx_m | let me check manually to make sure my bug_api.php file is correct | 11:59 |
| dhx_m | ah, I think you may have fixed the description field | 12:02 |
| dhx_m | but I was editing the bugnote | 12:02 |
| dhx_m | let me check if the description field is ok | 12:02 |
| paulr | also, it won't fix 'stuff you've already done' | 12:03 |
| dhx_m | yep, no duplicate description change history | 12:03 |
| dhx_m | but I hadn't tested that before anyway, so I don't know if it was originally a problem | 12:04 |
| paulr | but I think bugnotes needs fixing anyway | 12:04 |
| paulr | for the same thing I had john fix in bugs last night | 12:04 |
| dhx_m | ideally I recon Mantis should behave more like Medawiki | 12:05 |
| dhx_m | where you keep a revision history of the entire page | 12:05 |
| dhx_m | *entire bug | 12:05 |
| dhx_m | and you can diff between revisions | 12:05 |
| dhx_m | it is annoying having to click on "View Revisions" for each bugnote that has been modified, for the description field and other text fields, etc | 12:06 |
| dhx_m | the "View Revisions" page only shows the differences for one field at a time | 12:06 |
| dhx_m | oh my mistake | 12:12 |
| dhx_m | there is a "View All Revisions" button that shows everything | 12:12 |
| dhx_m | I take that back | 12:13 |
| paulr | dhx_m: gitsync | 12:14 |
| CIA-18 | Mantisbt: paul * r95ff20e990a9 /core/bugnote_api.php: 1. Handle bugnote revisions in same way as bug revisions | 12:14 |
| dhx_m | that works ;) thanks | 12:15 |
| dhx_m | you fixed all the bugnote problems I have at once, nice :) | 12:16 |
| dhx_m | saving a bugnote edit without actually making changes no longer leaves a revision/history item behind | 12:16 |
| dhx_m | and the "View Revisions" only shows for bug notes that have been edited | 12:17 |
| dhx_m | another parse error: Parse error: syntax error, unexpected T_FOR in /var/www/localhost/htdocs/mantis-master/roadmap_page.php on line 303 | 12:18 |
| dhx_m | and: | 12:18 |
| dhx_m | Parse error: syntax error, unexpected '<' in /var/www/localhost/htdocs/mantis-master/account_page.php on line 177 | 12:18 |
| dhx_m | (I have pretty much every feature turned on, which is why I'm finding a lot of these problems) | 12:19 |
| dhx_m | also: | 12:20 |
| dhx_m | Fatal error: Call to undefined function ctype_space() in /var/www/localhost/htdocs/mantis-master/core/rssbuilder/class.RSSBase.inc.php on line 64 | 12:20 |
| dhx_m | I think I've clicked on every link/button I can find now :D | 12:21 |
| paulr | Fatal error: Call to undefined function ctype_space() | 12:22 |
| paulr | huh? | 12:22 |
| paulr | what generates that | 12:22 |
| dhx_m | the "RSS" link on the News page | 12:23 |
| dhx_m | I don't have any news items, which may be an issue | 12:23 |
| dhx_m | I'll create one and see if that resolves the problem | 12:23 |
| dhx_m | ah nope, still an issue | 12:23 |
| paulr | news_rss ? | 12:24 |
| dhx_m | yep | 12:24 |
| paulr | http://uk.php.net/ctype_space | 12:25 |
| paulr | hmmm | 12:25 |
| paulr | Beginning with PHP 4.2.0 these functions are enabled by default. For older versions you have to configure and compile PHP with --enable-ctype. You can disable ctype support with --disable-ctype. | 12:25 |
| paulr | The Windows version of PHP has built-in support for this extension. You do not need to load any additional extensions in order to use these functions. | 12:25 |
| paulr | ok | 12:25 |
| paulr | what version of php you have exactly and where from? :) | 12:25 |
| dhx_m | ah | 12:25 |
| dhx_m | 5.2.8-r2 Gentoo Linux | 12:25 |
| dhx_m | and yes, I have the ctype USE flag disabled | 12:26 |
| paulr | why? | 12:26 |
| dhx_m | which is the cause of that problem it seems | 12:26 |
| paulr | and before you fix it | 12:26 |
| dhx_m | that is the default state on Gentoo | 12:26 |
| paulr | really? | 12:26 |
| * paulr wonders why | 12:26 | |
| dhx_m | they must consider it non-essential | 12:27 |
| dhx_m | the PHP ebuild for Linux has most things disabled by default | 12:27 |
| dhx_m | I know that RHEL4 is ok with ctype_space | 12:27 |
| dhx_m | I wonder if this is something to put in the installation documentation... or installation script | 12:29 |
| nuclear_eclipse | howdy all | 12:30 |
| dhx_m | I guess it doen't affect many people though (I consider Gentoo users fairly crazy!) | 12:30 |
| dhx_m | hey | 12:30 |
| paulr | nuclear_eclipse: fail | 12:31 |
| giallu | dhx_m, well, yes I think we are technically able to tell if a required extension is missing | 12:31 |
| dhx_m | I always thought PHP had an is_space function... I must be mistaken | 12:32 |
| paulr | we could probably wrap ctype_space to is_blank | 12:33 |
| dhx_m | not worth it IMO given this will really only affect people who compile their own version of PHP (most people won't) | 12:34 |
| paulr | i'll add a check | 12:34 |
| dhx_m | in the installation? | 12:35 |
| dhx_m | core/adodb/adodb-datadict.inc.php also uses ctype_alnum | 12:35 |
| dhx_m | other than that, these functions are unused | 12:36 |
| dhx_m | I think the reason Gentoo doesn't include this feature in PHP is that it relies on having the standard C library installed | 12:37 |
| dhx_m | and maybe under very specialised circumstances, you might be using a strange C library (not GNU libc)... or maybe there is 32bit vs 64bit problems with PHP/libc | 12:38 |
| CIA-18 | Mantisbt: paul * r3f2bc1a5ed7d /account_page.php: sigh | 12:38 |
| paulr | dhx_m: can you go to /admin/check.php and check the new test I just added fails | 12:38 |
| CIA-18 | Mantisbt: paul * r3db07add07cf /admin/check.php: Add test to check that ctype in php is enabled for rss feeds. | 12:38 |
| dhx_m | ok | 12:38 |
| dhx_m | check fails correctly, well done :) | 12:41 |
| dhx_m | account_page also works ok now | 12:43 |
| paulr | has john broken anything else? | 12:43 |
| dhx_m | the only other thing I noted above was: | 12:43 |
| dhx_m | Parse error: syntax error, unexpected T_FOR in /var/www/localhost/htdocs/mantis-master/roadmap_page.php on line 303 | 12:43 |
| dhx_m | $t_count_ids = count( $t_issue_set_ids ) | 12:44 |
| * paulr isghs | 12:44 | |
| dhx_m | line 302 | 12:44 |
| paulr | next job | 12:44 |
| dhx_m | missing ; | 12:44 |
| paulr | make a test script that visits every page | 12:44 |
| dhx_m | yep, that'd be very handy :) | 12:45 |
| dhx_m | although is the idea of this testing to test branches before merging them into the master? | 12:49 |
| dhx_m | as opposed to doing the merge and then making sure nothing broke? | 12:49 |
| paulr | tbh, we tend to test branches | 12:50 |
| paulr | it's when we do tidyups that stuff breaks :) | 12:50 |
| dhx_m | yep | 12:51 |
| dhx_m | not exactly major breakages though :) | 12:51 |
| dhx_m | which is good | 12:51 |
| * paulr pokes nuclear_eclipse | 12:53 | |
| dhx_m | I suppose a test would need to be create for each page the user might land on | 12:55 |
| dhx_m | and the test would require all user input parameters are hard coded into the test | 12:56 |
| dhx_m | and any database/file modifications are 'muted'? | 12:56 |
| * paulr shrugs | 12:56 | |
| [KK]Kirill | paulr: John talk with me ;) | 12:57 |
| paulr | hi | 12:57 |
| [KK]Kirill | hi | 12:57 |
| nuclear_eclipse | hi paulr | 12:57 |
| dhx_m | hey | 12:57 |
| paulr | hi john | 12:57 |
| * paulr confused now | 12:57 | |
| paulr | nuclear_eclipse: talk to kirill | 12:57 |
| nuclear_eclipse | I am | 12:58 |
| paulr | wow | 13:02 |
| paulr | nuclear_eclipse: did siebrancd modifify some libs? | 13:04 |
| nuclear_eclipse | I don't remember | 13:06 |
| dhx_m | sorry, don't want to nag... but just wanted to point out the missing ; on line 302 of roadmap_page.php again before it is forgotten ;) | 13:08 |
| dhx_m | it is one of those things I can't be bothered making a ticket/patch over :p | 13:08 |
| dhx_m | and I've gtg soon | 13:08 |
| nuclear_eclipse | dhx_m: no worries, it's good to bitch at paulr when he makes these mistakes; he does it enough to ours :P | 13:09 |
| CIA-18 | Mantisbt: paul * r4b4e98c7c671 /roadmap_page.php: missing ; | 13:10 |
| paulr | nuclear_eclipse: new phpmailer version! | 13:10 |
| paulr | version 5! | 13:10 |
| nuclear_eclipse | yay.... | 13:10 |
| dhx_m | haha, just as long as your code isn't front page thedailywtf.com you're ok! | 13:10 |
| dhx_m | thanks paulr :) | 13:11 |
| dhx_m | I've tested most of the major functionality of Mantis now with the cleanups | 13:12 |
| dhx_m | there are a few things left over to check such as plugins, editing/updating information (so far I've mostly only tested viewing) | 13:12 |
| dhx_m | SVN integration and a few other minor things | 13:13 |
| nuclear_eclipse | dhx_m: you're testing help is greatly appreciated :) | 13:13 |
| dhx_m | nuclear_eclipse: no problems, I am trying to follow development as much as possible | 13:15 |
| paulr | dhx_m: does your test box having working mail? | 13:16 |
| dhx_m | unfortunately not at the moment | 13:16 |
| dhx_m | but I can set that up on another host easily enough | 13:17 |
| dhx_m | actually I can fix that now | 13:18 |
| paulr | dhx_m: wasn't you going ?;p | 13:26 |
| dhx_m | in a moment, yep :) | 13:26 |
| dhx_m | if I don't setup mail now, I'll most likely forget | 13:27 |
| dhx_m | or not be bothered to do it later heh | 13:27 |
| dhx_m | ok mail works ok | 13:31 |
| dhx_m | so I'll test that from now on as well | 13:31 |
| dhx_m | it seems I found an idea for a new patch... email notification when a bug note in a monitored issue is changed | 13:35 |
| dhx_m | also account_prefs_page.php has the string "E-mail on note Added" where the N should be capitalised to match the other strings | 13:35 |
| nuclear_eclipse | paulr: drop the mantisbt/revisions.git and mantisbt/bug9856.git ? | 13:36 |
| paulr | nuclear_eclipse: there? | 13:36 |
| paulr | heh | 13:36 |
| paulr | I take that as yes | 13:36 |
| paulr | http://git.mantisforge.org/w/mantisbt.git?a=commitdiff;h=be5f1dda9128ae5e994f827660b93d747c5c0c2c | 13:36 |
| paulr | http://git.mantisforge.org/w/mantisbt.git?a=commitdiff;h=2f3156bcb18bd56c2a178a4856e0bfe69768f7ca | 13:36 |
| paulr | whats your thoughts on those ? | 13:36 |
| * nuclear_eclipse looks | 13:38 | |
| nuclear_eclipse | hmm, if we're making changes to PHPmailer, shouldn't we be tossing them upstream for approval? | 13:39 |
| paulr | I did with mine | 13:39 |
| * paulr mailed | 13:39 | |
| * paulr never got a reply | 13:39 | |
| paulr | however | 13:39 |
| paulr | between 2.2->2.3 of php mailer | 13:39 |
| paulr | LE I think went to \r\n | 13:39 |
| paulr | as per: | 13:40 |
| paulr | +* changed default of $LE to "\r\n" to comply with RFC 2822. Can be set by the user | 13:40 |
| paulr | + if default is not acceptable | 13:40 |
| paulr | that I think is back to \n now | 13:40 |
| paulr | ;p | 13:40 |
| dhx_m | check RFC2822... it should be \r\n | 13:41 |
| dhx_m | not just \n | 13:41 |
| dhx_m | http://tools.ietf.org/html/rfc2822 page 5, section 2.1 | 13:41 |
| paulr | :) | 13:41 |
| dhx_m | so I'd agree with the comment about following upstream | 13:42 |
| paulr | dhx_m: you mis understand here actually ;) | 13:43 |
| paulr | upstream seem to be a bit confused ;p | 13:43 |
| dhx_m | well I guess they might not have replied to your emails because they no longer properly support PHPMailer 2.x | 13:44 |
| paulr | in 5.0: | 13:44 |
| paulr | * Provides the ability to change the line ending | 13:44 |
| paulr | * @var string | 13:44 |
| paulr | */ | 13:44 |
| paulr | public $LE = "\n"; | 13:44 |
| dhx_m | oh :o | 13:44 |
| dhx_m | that looks wrong | 13:44 |
| paulr | in 2.3: | 13:44 |
| paulr | * Provides the ability to change the line ending | 13:44 |
| paulr | * @var string | 13:44 |
| paulr | */ | 13:44 |
| paulr | public $LE = "\r\n"; | 13:44 |
| paulr | I *think* in 2.2 it was \n | 13:45 |
| dhx_m | ah | 13:45 |
| dhx_m | let's see what PHPMailer 6.x does | 13:45 |
| paulr | you mean 5.x | 13:45 |
| paulr | in 2.2.1 it was: | 13:46 |
| paulr | private $LE = "\n"; | 13:46 |
| dhx_m | yep oops, was getting confused | 13:46 |
| dhx_m | so it seems they haven't ported that fix into their 5.0 version | 13:46 |
| paulr | and when they changed from \r\n to \n | 13:46 |
| paulr | they broke stuff | 13:46 |
| paulr | erm | 13:47 |
| paulr | \n to \r\n | 13:47 |
| paulr | which is what I mailed about ;p | 13:47 |
| paulr | I made a possible fix for that | 13:47 |
| paulr | siebrand has since modified by fix | 13:47 |
| paulr | so I dont know if my attempt at fixing broke something | 13:47 |
| paulr | or :) | 13:47 |
| dhx_m | so 5.0 has \n as the line ending? | 13:47 |
| nuclear_eclipse | I'm confused; so is RFC2822 specifying '\r\n', and that causes everything to break? | 13:47 |
| paulr | no | 13:47 |
| dhx_m | nuclear_eclipse: RFC2822 will kill you if you use \n :p | 13:47 |
| paulr | RFC2822 specifies \r\n | 13:48 |
| paulr | most mail daemons I think accept/allow either \r\n or \n nowadays | 13:48 |
| paulr | iirc | 13:48 |
| nuclear_eclipse | so why is PHPmailer using just '\n', and why are we switching to '\n' ? | 13:48 |
| paulr | phpmailer sends \n (I think) - at least it has a variable defining LE's to \n | 13:48 |
| paulr | phpmailer: | 13:48 |
| paulr | 2.2.1 set the $LE variable to \n | 13:48 |
| dhx_m | which is against what RFC2822 defines | 13:48 |
| paulr | 2.3.0 set the $LE variable to \r\n (referencing RFC in changelog) | 13:48 |
| paulr | 5.0.0 sets the $LE variable to \n | 13:49 |
| dhx_m | but maybe they do that as a hack to get around a broken SMTP implementation from oh... Microsoft? someone else? :p | 13:49 |
| paulr | when phpmailer changed | 13:49 |
| paulr | we got a bug report, which I attempted to fix by using \n in an encode function | 13:49 |
| paulr | and mailed upstream who didn't respond | 13:49 |
| paulr | since then siebrand has made a subsequent patch to my patch to fix something else | 13:49 |
| paulr | http://www.mantisbt.org/bugs/view.php?id=10269 | 13:50 |
| paulr | Email subjects are generated incorrectly, up to 1.1.6 they were really wrong because you were splitting the resulting base64 line againts base64 rules, in 1.2.0a3 the situation is much better, you are apparently splitting the text before encoding it in base64 but you add an extra space for each encoded part even if there wasn't a space in the original text. | 13:50 |
| [KK]Kirill | bug 10303 | 13:53 |
| mantisbot | Bug 10303 - jreese - open - assigned | 13:53 |
| mantisbot | Create and use string_api wrappers for multibyte string functions - http://www.mantisbt.org/bugs/view.php?id=10303 | 13:53 |
| dhx_m | I thought maybe RFC822 (predecessor to RFC2822) didn't require CRLF... but it does | 13:55 |
| dhx_m | so I don't see a reason (unless there is another RFC spec I'm unaware of) why people shouldn't be using CRLF | 13:56 |
| nuclear_eclipse | I would agree | 13:56 |
| nuclear_eclipse | I think that using '\n' against spec is probably the "wrong" solution.... | 13:56 |
| nuclear_eclipse | and I never saw any problems at all in any of my mail readers when using '\r\n'... | 13:57 |
| dhx_m | their mailing list looks rather dead | 13:57 |
| dhx_m | it sounds like they aren't really setup nicely to receive input from the wider community? | 13:57 |
| nuclear_eclipse | IIRC, phpmailer isn't really a community effort anyways? | 13:58 |
| nuclear_eclipse | maybe I'm thinking of a diff lib though | 13:58 |
| dhx_m | the last bug they closed on their sourceforge bugtracker was from November 2008 | 13:59 |
| dhx_m | and given that sourceforge is basically now a historical code museum (does anyone still use it for active projects?), that doesn't surprise me | 13:59 |
| paulr | i'm going to reopen those 2 bugs I think | 14:00 |
| paulr | and commit new phpmailer | 14:00 |
| paulr | and leave dhx_m to test | 14:00 |
| paulr | whether I've just broken it | 14:00 |
| paulr | :) | 14:00 |
| dhx_m | well I'm not too sure what the original problem was... but I'll try ;) | 14:01 |
| paulr | 'it' being mailing in general | 14:01 |
| dhx_m | it might be worth nothing RFC2822 section 2.1 in your commit | 14:01 |
| paulr | I think we might need to initise class with exceptions = false | 14:01 |
| dhx_m | *noting | 14:01 |
| dhx_m | so that others know why it is being reverted | 14:02 |
| giallu | paulr, if we are upgrading phpmailer, please stick to upstream | 14:04 |
| paulr | lol | 14:05 |
| paulr | we do | 14:05 |
| dhx_m | giallu: even if they ignore bug reports? | 14:05 |
| paulr | upstream is just broken :) | 14:05 |
| dhx_m | I do agree it is best to follow upstream... but it doesn't sound like they're being responsive | 14:05 |
| * nuclear_eclipse agrees with giallu | 14:05 | |
| giallu | lol... an in their home page they sugget using the sf.net tracker.. | 14:05 |
| giallu | dhx_m, yeah. we planmned to change our mailer | 14:06 |
| giallu | but I never got to it really | 14:06 |
| dhx_m | is there another alternative to p... ah ok :) | 14:06 |
| paulr | well, to be fair giallu | 14:06 |
| paulr | they started doing releases again | 14:06 |
| giallu | paulr, who cares | 14:06 |
| CIA-18 | Mantisbt: paul * r6405283d96ac /core/phpmailer/ (24 files in 3 dirs): Update PHP Mailer to 5.0. | 14:07 |
| giallu | doing releases is just _one_ metric for a healty upstream | 14:07 |
| dhx_m | that's nice... but they need to be responsive to their mailing list and errmmm "bug tracker" (if you'd call it that) | 14:07 |
| dhx_m | yep | 14:07 |
| dhx_m | Mantis is great because it is very open to suggestions, changes, bug reports, etc | 14:07 |
| nuclear_eclipse | wtf is this 'metric' garbage you peoples talk about? :P | 14:07 |
| dhx_m | ie. a very healthy development environment | 14:07 |
| paulr | nuclear_eclipse: http://www.mantisforge.org:8080/cruisecontrol/buildresults/mantisbt?tab=metrics | 14:08 |
| * nuclear_eclipse kicks paulr | 14:08 | |
| paulr | apparently our test framework covers 100% of the code it tests | 14:08 |
| nuclear_eclipse | lol | 14:08 |
| dhx_m | hah | 14:09 |
| paulr | and look at the number of codesniffer errors i've fixed today | 14:09 |
| nuclear_eclipse | paulr: so how does that tell a "good" build from "bad"? | 14:09 |
| paulr | it's gone from 1500 to ~0 | 14:09 |
| paulr | nuclear_eclipse: broken builds are when the scripts to execute tests+phpdoc etc fail to run :) | 14:09 |
| nuclear_eclipse | mmk | 14:10 |
| paulr | I should really work out how to clear old logs | 14:10 |
| dhx_m | ie. it tells you when you're missing a brace somewhere | 14:10 |
| paulr | :) | 14:10 |
| dhx_m | well email stills works after the upgrade :) | 14:11 |
| dhx_m | which sounds like the 2.x -> 5.x version change is just a marketing ploy | 14:11 |
| paulr | nod | 14:12 |
| dhx_m | if it required that little work to jump 3 whole versions | 14:12 |
| paulr | well from what i can tell | 14:12 |
| giallu | well that's strange | 14:12 |
| paulr | having just diff'd the whole thing | 14:12 |
| paulr | the main changes seem to be | 14:12 |
| giallu | since the tarball of 5.0 is twice as large | 14:12 |
| paulr | a) using exceptions internally | 14:12 |
| paulr | giallu: they added a test framework | 14:12 |
| nuclear_eclipse | dhx_m: coulda been like WinAmp: WA2 + WA3 = WA5 :P | 14:12 |
| giallu | ah | 14:12 |
| paulr | giallu: I didn't include it all | 14:12 |
| paulr | as some of it seemed to be a .swf | 14:12 |
| dhx_m | nuclear_eclipse: haha | 14:12 |
| paulr | so i'm assuming the test thing has an interface for sending mails... | 14:13 |
| * daryn_away is away: Gone away for now | 14:14 | |
| dhx_m | paulr: I checked with wireshark and it is indeed sending \r\n as line ends now | 14:14 |
| paulr | hmm | 14:15 |
| paulr | wonder what $LE=\n refers to then | 14:15 |
| * paulr shrugs | 14:15 | |
| paulr | unless thats 'line endings within body of mail' | 14:16 |
| dhx_m | unless \0d\0a refers to something else :p | 14:16 |
| nuclear_eclipse | holy crap! a tester who even knows what Wireshark is!!! | 14:16 |
| * nuclear_eclipse hugs dhx_m | 14:16 | |
| paulr | nuclear_eclipse: who is this dhx guy? :) | 14:16 |
| dhx_m | nah, it is \0d\0a for both header line endings and the ending of lines withint he message | 14:16 |
| dhx_m | haha | 14:16 |
| dhx_m | well yeah, I'm not your traditional PHP/web development person | 14:17 |
| dhx_m | in that I code in assembler, C, C++, etc | 14:17 |
| dhx_m | so TCP packets don't scare me :p | 14:17 |
| paulr | dhx_m: do you code flash? | 14:18 |
| dhx_m | paulr: I very much hate Flash, so nope | 14:18 |
| dhx_m | no offence hah | 14:18 |
| nuclear_eclipse | hooray! | 14:18 |
| nuclear_eclipse | I'm not the only one who hates flash! | 14:19 |
| dhx_m | lol | 14:19 |
| dhx_m | I've been testing Firefox 3.1 and nightly builds of Epiphany/Webkit solely for the new <video> and <audio> HTML5 tags | 14:19 |
| dhx_m | I can't wait for that to go mainstream | 14:20 |
| dhx_m | and replace Flash for good :p | 14:20 |
| nuclear_eclipse | +1 | 14:20 |
| dhx_m | paulr: core/phpmailer/class.smtp.php:60 = public $CRLF = "\r\n"; | 14:20 |
| nuclear_eclipse | and for Canvas to replace the remaining uses of flash.... | 14:20 |
| dhx_m | hmmm I haven't read much about that | 14:21 |
| dhx_m | sounds bad from the Wikipedia entry... if it uses Javascript to do the drawing | 14:21 |
| dhx_m | what was wrong with SVG? | 14:21 |
| nuclear_eclipse | dhx_m: performance | 14:21 |
| nuclear_eclipse | and SVG lacks the ablity for high-level logic | 14:22 |
| nuclear_eclipse | eg, Canvas can be used for games, animations, etc that SVG doesn't really support | 14:22 |
| dhx_m | while it does sound like a large step up from Flash/Silverlight(lol) | 14:23 |
| dhx_m | it still maintains the whole concept of breaking usability so people can invent their own stupid web UIs | 14:23 |
| nuclear_eclipse | hehe | 14:23 |
| dhx_m | ie. Flash based site navigation (GRRRRRRRRRR!) | 14:24 |
| paulr | right anyway | 14:24 |
| paulr | back to work! | 14:24 |
| dhx_m | :) | 14:24 |
| paulr | what we need to do next | 14:24 |
| CIA-18 | Mantisbt: paul * rf89f1a0fde01 /core/rssbuilder/ (13 files in 2 dirs): Rssbuilder updated to 2.2.1 | 14:24 |
| nuclear_eclipse | gotta take the bad with the good I reckon, but at least canvas will be built into every browser, rather than requiring an opaque third party plugin.... | 14:24 |
| paulr | canvas? | 14:24 |
| dhx_m | http://en.wikipedia.org/wiki/Canvas_%28HTML_element%29 | 14:24 |
| dhx_m | It is going to get abused heavily by web developers that think code should run on the client's machine | 14:25 |
| nuclear_eclipse | I think the only thing that canvas can't do that flash can is combine sounds within the animation | 14:26 |
| dhx_m | "Canvas consists of a drawable region defined in HTML code with height and width attributes. JavaScript code may access the area through a full set of drawing functions similar to other common 2D APIs, thus allowing for dynamically generated graphics. Some anticipated uses of the canvas include building graphs, animations, games, and image composition." | 14:26 |
| paulr | nuclear_eclipse: what's plan for 1.2.0? | 14:27 |
| nuclear_eclipse | are you still doing anything with pluggable authentication? | 14:28 |
| dhx_m | nuclear_eclipse: do you write patches for Mantis? | 14:29 |
| paulr | lol | 14:29 |
| paulr | nuclear_eclipse: I guess I should really | 14:29 |
| dhx_m | a different username for committing? | 14:29 |
| giallu | jrees | 14:29 |
| paulr | dhx_m: jreese | 14:29 |
| dhx_m | OOOOHHHH! :) | 14:30 |
| dhx_m | lol | 14:30 |
| dhx_m | that makes a lot of sense now hah | 14:30 |
| paulr | I guess I should look at: auth, cf, duedate/times and utf8 | 14:30 |
| dhx_m | if you're bored, tell me what you think of: | 14:30 |
| dhx_m | http://git.mantisforge.org/w/mantisbt/dhx.git?a=commit;h=67476ec9cef810b0fc6cf1b4adca63512aac1f0f | 14:31 |
| paulr | or something | 14:31 |
| dhx_m | that is just a minor fix | 14:31 |
| nuclear_eclipse | dhx_m: I'll look at it shortly | 14:31 |
| * paulr is already looking at it | 14:32 | |
| dhx_m | there are also a few other things I came up with... but I'm not 100% sure if they're correct/good: | 14:32 |
| nuclear_eclipse | is there an associated bug report? | 14:32 |
| dhx_m | http://git.mantisforge.org/w/mantisbt/dhx.git?a=commit;h=229a6848f6eb8a8847fa85c763b45451556b708f | 14:32 |
| dhx_m | yep, I'll give you those links instead | 14:32 |
| dhx_m | needs closing as "already fixed" / "invalid": http://www.mantisbt.org/bugs/view.php?id=10282 | 14:33 |
| dhx_m | Empty target version field is shown in advanced bug report screen when no versions exist for a project: http://www.mantisbt.org/bugs/view.php?id=10281 | 14:33 |
| nuclear_eclipse | closed | 14:33 |
| dhx_m | Add new stick/unstick button to html_buttons_view_bug_page: http://www.mantisbt.org/bugs/view.php?id=10268 | 14:33 |
| dhx_m | Display target release date next to target version: http://www.mantisbt.org/bugs/view.php?id=10274 | 14:34 |
| paulr | closed? | 14:34 |
| dhx_m | I reported it by accident | 14:34 |
| dhx_m | Added ability to have the html buttons bar at the top, bottom or both of the bug view page: http://www.mantisbt.org/bugs/view.php?id=10217 | 14:34 |
| dhx_m | that last one (10217) is not my suggestion | 14:34 |
| dhx_m | and to be honest, I think it sucks | 14:35 |
| dhx_m | ... a bit | 14:35 |
| dhx_m | but it can be useful if the metadata of the bug scrolls off the screen | 14:35 |
| paulr | target_version makes sense | 14:36 |
| paulr | imo | 14:36 |
| dhx_m | (you have a lot of custom fields, attachments, a long description, etc) | 14:36 |
| dhx_m | even if you have no target versions to select? | 14:36 |
| CIA-18 | Mantisbt: paul * rf28189be5521 /bug_report_advanced_page.php: Fix: #0010281: Empty target version field is shown in advanced bug report screen when no versions exist for a project | 14:37 |
| paulr | nuclear_eclipse: and yes, imo we should | 14:38 |
| paulr | 6 days ago 10274 shortlog | log | tree (base | diff) | 14:38 |
| paulr | 9 days ago 10217 shortlog | log | tree (base | diff) | 14:38 |
| paulr | 10 days ago 10268 shortlog | log | tree (base | diff) | 14:38 |
| paulr | 10 days ago 10263 shortlog | log | tree (base | diff) | 14:38 |
| paulr | dhx_m: those as well? | 14:38 |
| dhx_m | 10263 is already merged | 14:38 |
| dhx_m | those numbers refer to the issue # on the mantis bugtracker | 14:39 |
| paulr | nod | 14:39 |
| dhx_m | I've rebased all my commits into just one | 14:39 |
| dhx_m | so it is cleaner (no "oops... forgot a ;" | 14:39 |
| paulr | :( | 14:39 |
| * paulr goes off to sulk | 14:39 | |
| dhx_m | :) | 14:40 |
| dhx_m | some of those other patches may need work | 14:40 |
| dhx_m | they're more the "feature changes" | 14:40 |
| dhx_m | rather than bug fixes | 14:40 |
| dhx_m | example... for 10274, I was uncertain about whether the <td> cells should be printed for each html_button even if that button won't be displayed | 14:42 |
| dhx_m | I'm not sure the code was already consistent in that respect | 14:43 |
| dhx_m | grangeway = paulr? | 14:47 |
| dhx_m | nuclear_eclipse: very nice work with the revisions on bug notes... I was really looking for that sort of feature :) | 14:50 |
| * daryn is back. | 15:02 | |
| nuclear_eclipse | dhx_m: thanks, so was my employer :) | 15:21 |
| nuclear_eclipse | and yes, paulr == grangeway | 15:21 |
| dhx_m | :) | 15:21 |
| nuclear_eclipse | I'm actually lucky enough that most of my work on mantisbt is part of my job :) | 15:22 |
| nuclear_eclipse | although unfortunately, most of my time lately has been to spend time on proprietary in-house plugins | 15:22 |
| dhx_m | interfacing with in-house stuff? | 15:23 |
| nuclear_eclipse | yeah | 15:23 |
| dhx_m | the plugin system is a nice direction to head towards (at least I think so) | 15:24 |
| nuclear_eclipse | like using a set of perl scripts on top of SCCS for source control of our products (ugh) | 15:24 |
| dhx_m | is the aim to try and move core Mantis features into plugins where possible? | 15:24 |
| dhx_m | for example... SVN integration, time tracking, the status page, emailing? | 15:24 |
| nuclear_eclipse | no, core Mantis features should generally stay in core Mantis | 15:24 |
| nuclear_eclipse | plugins are designed for a) modifying core outputs, and b) adding new features to Mantis | 15:25 |
| dhx_m | so is there still room for adding new features into the Mantis core? | 15:25 |
| dhx_m | or will most of that be forced into plugins now? | 15:25 |
| nuclear_eclipse | depends on the feature, and how integral it is to the point of tracking bugs/issues/projects | 15:26 |
| dhx_m | I guess IRC notification would be considered non-core? | 15:26 |
| nuclear_eclipse | correct | 15:26 |
| dhx_m | and email notification is still 'core'? | 15:27 |
| nuclear_eclipse | right | 15:27 |
| [KK]Kirill | nuclear_eclipse: öðí, | 15:27 |
| [KK]Kirill | why? | 15:27 |
| nuclear_eclipse | although, we are "planning" to build a new pluggable notification system after 1.2.0, which will allow plugins to hook new notification systems beyond just emails | 15:27 |
| [KK]Kirill | Yes!!!! | 15:28 |
| dhx_m | can more than one plugin use a hook at a time? | 15:28 |
| nuclear_eclipse | eg, a plugin could be created to hook notifications, and then you could use that for an IRC bot, Twitter, IM, etc | 15:28 |
| dhx_m | yep | 15:28 |
| nuclear_eclipse | and yes, events can be hooked by any number of plugins | 15:28 |
| dhx_m | great | 15:28 |
| dhx_m | the main thing I think I'd like to see changed is restructuring the database schema to be more orthogonal | 15:29 |
| nuclear_eclipse | I've already created a rather comprehensive source control integration framework that is probably the best example of what the plugin system can do in its current state | 15:29 |
| dhx_m | :) | 15:29 |
| [KK]Kirill | John | 15:30 |
| [KK]Kirill | $ git checkout Kirill | 15:30 |
| [KK]Kirill | error: You have local changes to 'core/phpmailer/README'; cannot switch branches. | 15:30 |
| nuclear_eclipse | it's the same source integration used on our official tracker, although our tracker is using an old copy, and not many of the extra features it contains | 15:31 |
| [KK]Kirill | John? | 15:33 |
| dhx_m | I've actually noticed that commit f28a89be (http://www.mantisbt.org/bugs/plugin.php?page=Source/view&id=13900) doesn't show up in the corresponding bug 10281 | 15:33 |
| mantisbot | Bug 10281 - dhx - fixed - resolved | 15:33 |
| mantisbot | Empty target version field is shown in advanced bug report screen when no versions exist for a project - http://www.mantisbt.org/bugs/view.php?id=10281 | 15:33 |
| nuclear_eclipse | you need to do `git checkout <filename>` to drop the local changes to that file before you can switch branches | 15:34 |
| dhx_m | bug 10282 | 15:34 |
| mantisbot | Bug 10282 - dhx - unable to reproduce - closed | 15:34 |
| mantisbot | Advanced bug report screen doesn't factor in new options $g_enable_profiles and $g_enable_build - http://www.mantisbt.org/bugs/view.php?id=10282 | 15:34 |
| dhx_m | ahhh, I was wondering why that appeared all of a sudden :) | 15:34 |
| [KK]Kirill | :( | 15:34 |
| * nuclear_eclipse pats mantisbot | 15:34 | |
| [KK]Kirill | kkrasnov@FCN0062 /c/Pers/SVN/MantisBTGit (master) | 15:35 |
| [KK]Kirill | $ git checkout core/phpmailer/README | 15:35 |
| [KK]Kirill | kkrasnov@FCN0062 /c/Pers/SVN/MantisBTGit (master) | 15:35 |
| [KK]Kirill | $ git checkout Kirill | 15:35 |
| [KK]Kirill | error: You have local changes to 'core/phpmailer/README'; cannot switch branches | 15:35 |
| nuclear_eclipse | ! | 15:36 |
| nuclear_eclipse | I have no clue | 15:36 |
| * giallu heading home | 15:37 | |
| nuclear_eclipse | bye giallu | 15:37 |
| [KK]Kirill | bye | 15:37 |
| [KK]Kirill | I go home too | 15:37 |
| dhx_m | [KK]Kirill: rm README | 15:38 |
| dhx_m | and check it out again | 15:38 |
| nuclear_eclipse | paulr: your ':' after 'fix' killed the ability for source integration to detect the link.... | 15:38 |
| dhx_m | then switch branches | 15:38 |
| [KK]Kirill | tnx | 15:38 |
| dhx_m | I think | 15:38 |
| nuclear_eclipse | btw paulr, when you use somebody's fix, you should do it in a way that gives them credit ;) | 15:39 |
| [KK]Kirill | not work | 15:39 |
| * paulr stops sulking | 15:39 | |
| dhx_m | hmmm | 15:40 |
| dhx_m | I'm not a git expert (yet) | 15:40 |
| nuclear_eclipse | eg, if you had merged his branch into master, it would have kept his author info and commit message | 15:40 |
| nuclear_eclipse | and then you could have ssigned off on his commit | 15:41 |
| dhx_m | at least I don't get the blame now :p | 15:41 |
| nuclear_eclipse | lol | 15:41 |
| dhx_m | I had *nothing* to do with it :p | 15:41 |
| nuclear_eclipse | well, with a signoff in Git, paulr would have still gotten the blame ;) | 15:41 |
| dhx_m | haha | 15:41 |
| nuclear_eclipse | I guess this just my queue to improve the developer guide with the needed info, and then rub everyone's nose in it... | 15:42 |
| dhx_m | one question I had is whether I need to continually merge the master back into my branches? | 15:43 |
| nuclear_eclipse | I'd generally say no | 15:44 |
| paulr | nuclear_eclipse: or you could not improve the guide and we just carry on using you | 15:44 |
| nuclear_eclipse | yeah, but then you go and do things without even asking me... | 15:44 |
| paulr | like what? | 15:44 |
| nuclear_eclipse | like using dhx_m's fix without giving him author credit... | 15:45 |
| paulr | oh | 15:45 |
| dhx_m | it's OK... no hard feelings! :) | 15:45 |
| nuclear_eclipse | the whole point of Git is that we can pull changes from remote branches to maintain author information | 15:45 |
| paulr | well, I was gonna do that last night with your stuff | 15:45 |
| paulr | :) | 15:45 |
| paulr | btw, git bundle exists too | 15:45 |
| nuclear_eclipse | right, the diff between bundle and format-patch (IIRC) is that `git bundle` requires parent objects to already exist in the receiving repo, and `git am` just applies the commits on top of your current branch | 15:47 |
| nuclear_eclipse | and bundle is more or less a low-level call meant to be used by git fetch and friends, not as a method of passing around commits/branches by hand | 15:48 |
| paulr | besides | 15:49 |
| paulr | if dhx is gonna be a regular commitor | 15:49 |
| paulr | should add to authors | 15:49 |
| dhx_m | well I wouldn't expect that straight away | 15:49 |
| dhx_m | I'm still a stranger here ;) | 15:49 |
| dhx_m | besides, do you have any form of review/sign-off so that more than 1 person reviews a patch before it is committed? | 15:50 |
| dhx_m | ie. person X authored it, person Y reviewed it | 15:50 |
| dhx_m | and they both signed off on the patch | 15:51 |
| paulr | dhx_m: asl? | 15:51 |
| dhx_m | haha | 15:51 |
| paulr | not that formally tbh | 15:52 |
| dhx_m | 23/m/en | 15:52 |
| paulr | the process is: | 15:52 |
| paulr | if john writes a patch | 15:52 |
| paulr | and asks for some feedback | 15:52 |
| paulr | he might get some | 15:52 |
| paulr | when he commits it | 15:52 |
| paulr | I read through commit | 15:52 |
| paulr | and whinge about anything I dont like :) | 15:53 |
| nuclear_eclipse | note the *might* part of that... | 15:53 |
| dhx_m | haha | 15:53 |
| paulr | I read through the commits | 15:53 |
| paulr | what this means is | 15:53 |
| paulr | you can code any feature and commit it | 15:53 |
| dhx_m | shouldn't that process happen in branches instead? | 15:53 |
| paulr | if it's broken, it gets pulled ;) | 15:53 |
| paulr | (that was in svn) | 15:53 |
| dhx_m | so the git master always builds? | 15:53 |
| paulr | mantis+git is still quite 'new' | 15:53 |
| dhx_m | yep | 15:53 |
| paulr | I tend not to code big features often | 15:54 |
| nuclear_eclipse | dhx_m: our general "policy" is that if you're committing anything large enough that changes how things work, ask for a review first, otherwise, try your best not to break master :P | 15:54 |
| dhx_m | it is new to most people in the open source community | 15:54 |
| paulr | but fiddle with things | 15:54 |
| dhx_m | yep | 15:54 |
| paulr | which means i normally do something (e.g. add replace space with tabs | 15:54 |
| paulr | fuck something up | 15:54 |
| dhx_m | when did you upgrade from SVN? | 15:54 |
| nuclear_eclipse | dhx_m: if we waited for formal reviews to get anything into master, we'd still be on version 1.1.2.... | 15:54 |
| paulr | and follow up with 1-2 fixes to fix the commit i did | 15:54 |
| dhx_m | nuclear_eclipse: I fully agree | 15:54 |
| nuclear_eclipse | we make the conversion around novermber-ish | 15:55 |
| nuclear_eclipse | made* | 15:55 |
| dhx_m | wow, that really is recent then | 15:55 |
| nuclear_eclipse | for the most part, it gets a lot like SVN, except that we can more easily handle asynchronous development with git, like pulling patches/branches | 15:55 |
| nuclear_eclipse | s/gets/gets used/ | 15:56 |
| dhx_m | the main advantage is branching is very cheap | 15:56 |
| nuclear_eclipse | right | 15:56 |
| * daryn loves git... | 15:56 | |
| dhx_m | which is why I've been doing 1 branch per feature so far | 15:56 |
| nuclear_eclipse | that, and merging actually works | 15:56 |
| dhx_m | although I have been rebasing my commits into a single patch | 15:56 |
| dhx_m | to hide the fact that I continually forget closing braces, etc | 15:57 |
| nuclear_eclipse | branching is easy in SVN; it's getting those branches to interact with each other that's the hard part :P | 15:57 |
| dhx_m | yep | 15:57 |
| nuclear_eclipse | dhx_m: that's fine, we tend to do the same thing | 15:57 |
| dhx_m | 1 branch per 3 line fix is going a bit too far | 15:57 |
| nuclear_eclipse | eg, the commit for 9856 was about 10 commits before I pushed it into master | 15:57 |
| dhx_m | but for major cleanups, etc... that is really something for a branch | 15:57 |
| dhx_m | yep | 15:57 |
| dhx_m | I guess the real problem is do you keep it in a branch for 3 months, or place it in the master ASAP and work on it there | 15:58 |
| dhx_m | Linux kernel development seems to say that you'd put it in the kernel ASAP so as many people as possible will see it, test it, develop on it | 15:58 |
| paulr | nuclear_eclipse: you need to tell me what to do here | 15:59 |
| nuclear_eclipse | right, that was the big trouble with 9856, it was too big a change to dump on master without review, but it sat for monthes on a branch waiting for review.... | 15:59 |
| paulr | and then we got bored | 15:59 |
| paulr | :) | 15:59 |
| nuclear_eclipse | gotta get to lunch, be back shortly | 15:59 |
| paulr | jfdi! | 15:59 |
| nuclear_eclipse | cheers | 15:59 |
| dhx_m | yep, but what you committed worked fine | 15:59 |
| dhx_m | according to the original design specs | 16:00 |
| dhx_m | paul just fixed a few minor things I picked up when I tested it | 16:00 |
| dhx_m | but that isn't necessarily something you'd always find in a review of code style, etc | 16:00 |
| dhx_m | you can have 5 people review a patch and completely miss some bugs | 16:01 |
| dhx_m | only to find someone else reports a bug or two in the patch 2 weeks after it is committed | 16:01 |
| paulr | dhx_m: well, what's interesting is I spotted what you picked up or well | 16:07 |
| paulr | the bugapi part | 16:07 |
| dhx_m | yep | 16:07 |
| dhx_m | I remember you mentioned that before | 16:08 |
| paulr | it's more why I knew what it was when you mentioned it | 16:08 |
| paulr | [12:54:56] <dhx_m> and it seems you can edit a bug note (without changing anything) and it'll record a new revision and history log entry | 16:09 |
| paulr | [12:54:59] <dhx_m> yep | 16:09 |
| paulr | [12:55:24] <paulr> yes, in fact it will | 16:09 |
| paulr | [12:55:37] <paulr> lets see if i'm right in thinking we didn't move something | 16:09 |
| paulr | 25 seconds to remember that | 16:09 |
| dhx_m | yep very quick :) | 16:10 |
| paulr | [12:58:08] <CIA-18> | 16:10 |
| paulr | ~3 minutes to fix | 16:10 |
| paulr | but I guess when you've already reviewed patch, spotted it needed fixing and forgot to tell the person who you was talking to when reviewing it... | 16:10 |
| paulr | dhx_m: could you test the phpmailer stuff with an invalid mailer setup | 16:22 |
| dhx_m | wrong smtp host, username, etc? | 16:22 |
| paulr | yea | 16:22 |
| paulr | i.e. throw an exception | 16:23 |
| dhx_m | ok | 16:23 |
| dhx_m | invalid password = silent... no hint at failure | 16:24 |
| paulr | as it's queueing mails probably | 16:25 |
| dhx_m | invalid host = same | 16:25 |
| dhx_m | I'm using the direct SMTP method | 16:25 |
| dhx_m | not php mail() or the other way (sendmail I think?) | 16:25 |
| paulr | ok so it might be fine | 16:26 |
| dhx_m | I just get no warning that it is not sending an email | 16:26 |
| dhx_m | oh actually | 16:26 |
| dhx_m | it does queue them | 16:27 |
| dhx_m | when I go back to the correct settings, I get all the emails again | 16:27 |
| paulr | WOW | 16:35 |
| paulr | nuclear_eclipse: mo? | 16:36 |
| CIA-18 | Mantisbt: paul * rbe736c3b3d61 /core/user_api.php: Fix: PHP Notice: Undefined variable: p in user_api.php on line 851 generated on live site. | 16:39 |
| * paulr wows more | 16:43 | |
| paulr | i've found something to optimise | 16:45 |
| * paulr bounces | 16:45 | |
| * nuclear_eclipse is back now | 16:49 | |
| paulr | nuclear_eclipse: our email processing is retarded | 16:50 |
| paulr | incidently | 16:51 |
| paulr | is the wiki broken atm? | 16:51 |
| nuclear_eclipse | I dunno | 16:51 |
| paulr | $t_access_level_string = strtoupper( get_enum_to_string( config_get( 'access_levels_enum_string' ), $t_access_level ) ); | 16:51 |
| paulr | we will get_enum_to_string iirc | 16:51 |
| nuclear_eclipse | wiki looks fine to me | 16:51 |
| paulr | [Wed Apr 08 12:30:32 2009] [error] [client 77.54.75.184] PHP Fatal error: Call to undefined function get_enum_to_string() in /var/www/html/wiki/inc/auth/mantis.class.php on line 80 | 16:52 |
| nuclear_eclipse | oh | 16:52 |
| nuclear_eclipse | probably the live site was never updated after the MantisEnum class was added | 16:52 |
| paulr | nod | 16:52 |
| nuclear_eclipse | had the same problem on our local install | 16:52 |
| paulr | but that's wiki | 16:52 |
| paulr | is there a quick fix? | 16:52 |
| nuclear_eclipse | yeah, find the cooresponding MantisEnum::function() and replace one line of code :P | 16:53 |
| paulr | oh | 16:53 |
| paulr | get_enum_to_string | 16:53 |
| paulr | just becomes | 16:53 |
| paulr | MantisEnum::get_enum_to_string ? | 16:53 |
| nuclear_eclipse | I dunno | 16:54 |
| paulr | MantisEnum::getValues | 16:54 |
| nuclear_eclipse | it's been a while since I had to fix it | 16:54 |
| paulr | adm_permissions_report.php(39): $t_access_levels = MantisEnum::getValues( config_get( 'access_levels_enum_string' ) ); | 16:54 |
| paulr | that'll do | 16:54 |
| * paulr copies and pastes | 16:54 | |
| nuclear_eclipse | so, what about our email processing is retarded, aside from the usual | 16:56 |
| paulr | [Wed Apr 08 00:03:41 2009] [error] [client 65.55.208.196] PHP Fatal error: Unknown function: config_get_global() in /var/www/html/bugs/core/database_api.php on line 59 | 16:56 |
| paulr | hmm | 16:56 |
| paulr | that's a strange error | 16:56 |
| nuclear_eclipse | agreed | 16:57 |
| paulr | seems to happen quite a lot too | 16:57 |
| paulr | hmm | 16:57 |
| paulr | 65.55.106.231 - - [06/Apr/2009:09:57:42 -0400] "GET /bugs/core/database_api.php?g_db_type=mysql:/root@localhost HTTP/1.0" 200 - "-" "msnbot/2.0b" | 17:00 |
| paulr | lol? | 17:00 |
| Enisseo | hi everyone! | 17:00 |
| nuclear_eclipse | hello | 17:00 |
| paulr | nuclear_eclipse: .htaccess it *or* can we do it in php ;p | 17:01 |
| nuclear_eclipse | paulr: I assume that's some sort of "hack" attempt trying to attack PHP's register_globals setting | 17:01 |
| nuclear_eclipse | ie, harmless | 17:01 |
| * daryn_away is away: Gone away for now | 17:02 | |
| nuclear_eclipse | and we could do it in code similar to what other apps do, eg, in core.php, define( 'MANTISBT', 1 ), and in all the API's, check for if !defined( MANTISBT ) die() type of thing | 17:02 |
| nuclear_eclipse | but I don't think it really matters | 17:03 |
| Enisseo | and it's overwritten anyway, no? | 17:03 |
| nuclear_eclipse | well, assuming that register_globals is off, then it wouldn't ever make it into the code we use | 17:04 |
| paulr | i can just .htaccess it out | 17:05 |
| Enisseo | yes, but if it wasn't (like on other websites), the value of $g_db_type would be replaced by the one that is in the config file? | 17:05 |
| paulr | Enisseo: iirc, our install requires registerglobals off | 17:05 |
| paulr | [Tue Apr 07 20:19:08 2009] [error] [client 65.55.208.190] PHP Parse error: syntax error, unexpected '<' in /var/www/html/bugs/core/filter_api.php on line 3709 | 17:06 |
| paulr | daryn_away broke filter api again :( | 17:06 |
| nuclear_eclipse | paulr: those errors are my fault; I updated the live site forgetting that master was "broken" | 17:07 |
| paulr | oh today? | 17:07 |
| nuclear_eclipse | but I ended up updating it back to Siebrand's last translation update | 17:07 |
| paulr | revisions are on live site? | 17:07 |
| nuclear_eclipse | no, that was last night | 17:07 |
| nuclear_eclipse | not including revisions | 17:07 |
| nuclear_eclipse | I was coming across the portugese language misnaming issue | 17:08 |
| nuclear_eclipse | so I was updating the live site to fix that problem | 17:08 |
| paulr | nod | 17:08 |
| paulr | but that issue is erm | 17:08 |
| paulr | we dont deal with language renames right? | 17:09 |
| paulr | i.e. you can't just rename a language file | 17:09 |
| paulr | and not update database | 17:09 |
| nuclear_eclipse | I dunno about that part, I was just bringing the codebase up to date - it's last update was middle of January :P | 17:09 |
| paulr | [Sun Apr 05 05:30:30 2009] [error] [client 72.30.79.119] PHP Fatal error: Allowed memory size of 52428800 bytes exhausted (tried to allocate 151 bytes) in /var/www/html/bugs/core/database_api.php on line 357 | 17:10 |
| nuclear_eclipse | yay | 17:11 |
| nuclear_eclipse | sometimes I think Mantis caches a bit *too* much; we eat a lot of memory with caching everything in static variables | 17:11 |
| paulr | 1) we need to fix issues from renaming languages | 17:12 |
| paulr | 2) we need to fix email | 17:14 |
| paulr | nuclear_eclipse: have you ever seen quickskin? | 17:14 |
| paulr | from phpmailer guys | 17:14 |
| nuclear_eclipse | nope | 17:15 |
| * paulr wants mail templates ;p | 17:15 | |
| paulr | 3) optimise print_all_bug_page | 17:17 |
| * paulr starts on 3 | 17:17 | |
| * nuclear_eclipse pokes paulr about pluggable auth for a 1.2.x release | 17:17 | |
| * paulr fixin broken shit first | 17:18 | |
| nuclear_eclipse | lies | 17:18 |
| paulr | [Mon Apr 06 03:45:22 2009] [error] [client 65.55.208.191] PHP Fatal error: Allowed memory size of 52428800 bytes exhausted (tried to allocate 10240 bytes) in /var/www/html/bugs/print_all_bug_page_word.php on line 342 | 17:18 |
| paulr | google bot keeps hitting it! | 17:18 |
| * paulr hates the browsersearchpage stuff | 17:19 | |
| * daryn is back. | 17:27 | |
| daryn | so how did I break filter_api? I haven't committed in ages! | 17:28 |
| nuclear_eclipse | I was kind of wondering the same thing.... | 17:29 |
| paulr | wow | 17:29 |
| paulr | print all bug page is actually broken too | 17:29 |
| daryn | paulr I think you imagine broken things | 17:29 |
| daryn | and then they break | 17:29 |
| paulr | daryn: it's ok, john worked out i broke it | 17:30 |
| daryn | ic | 17:30 |
| daryn | apology accepted | 17:30 |
| daryn | :P | 17:30 |
| paulr | does the mantis logo get downloaded via file_download or something | 17:31 |
| nuclear_eclipse | not to my knowledge | 17:33 |
| paulr | I've worked it out | 17:33 |
| nuclear_eclipse | it's just a straight HTTP ref | 17:33 |
| paulr | chrome = i'll browse some pages you might browse to | 17:34 |
| paulr | IE = i'll download search page logic to see if it's changed | 17:34 |
| paulr | whats mod in php | 17:37 |
| paulr | nuclear_eclipse: do you plan to update mantisbt.org to live code in 2 minutes? | 17:42 |
| nuclear_eclipse | why? | 17:42 |
| paulr | +if ( $t_row_count % 50 == 0 ) { | 17:42 |
| paulr | +# to save ram as report will list data once, clear cache after 50 bugs | 17:42 |
| paulr | +bug_text_clear_cache(); | 17:42 |
| paulr | +bug_clear_cache(); | 17:42 |
| paulr | +bugnote_clear_cache(); | 17:42 |
| paulr | +} | 17:42 |
| paulr | should fix print_all_bugs_word thing | 17:42 |
| nuclear_eclipse | fancy | 17:42 |
| paulr | or at least reduce some ram | 17:42 |
| paulr | -$t_last_modified = date( config_get( 'normal_date_format' ), ( db_unixtimestamp( $t_bugnote->last_modified ) ) ); | 17:42 |
| paulr | +$t_date_submitted = date( $t_date_format, $t_bugnote->date_submitted ); | 17:42 |
| paulr | even fixed a bug too :) | 17:42 |
| paulr | so i want to test and see if it does enough ;p | 17:44 |
| CIA-18 | Mantisbt: paul * rcf16fba204af / (core/bugnote_api.php print_all_bug_page_word.php): Fix: bugnote date was incorrect when printing bugs to word | 17:44 |
| nuclear_eclipse | has the filter_api problem been fixed ? | 17:44 |
| paulr | I fixed known whitespaces problems hours ago :P | 17:44 |
| nuclear_eclipse | ok, I wasn't sure if that fix was already committed or not | 17:45 |
| nuclear_eclipse | I blame you if live breaks now.... | 17:45 |
| paulr | [Wed Apr 08 12:33:20 2009] [error] [client 77.103.37.178] PHP Fatal error: Maximum execution time of 30 seconds exceeded in /var/www/html/bugs/core/email_api.php on line 1123, referer: http://www.mantisbt.org/bugs/view.php?id=4286 | 17:45 |
| paulr | I want to fix that ) | 17:45 |
| paulr | :) | 17:45 |
| paulr | but i'm not sure if I *really* want to | 17:45 |
| nuclear_eclipse | helper_begin_long_process() ? ;) | 17:46 |
| paulr | it's not that | 17:46 |
| paulr | we rebuild all the bug/note information | 17:46 |
| paulr | for each user a bug gets emailed to | 17:46 |
| paulr | the whole purpose of repeating that work seems to be | 17:46 |
| paulr | a) so we can localize the header type strings | 17:46 |
| paulr | b) so we can check permissions | 17:47 |
| paulr | site hasn't been updated yet? | 17:47 |
| paulr | [Wed Apr 08 13:46:41 2009] [error] [client 70.24.13.71] PHP Notice: Undefined variable: p in /var/www/html/bugs/core/user_api.php on line 851 | 17:47 |
| paulr | pretty sure I fixed that earlier | 17:47 |
| paulr | :) | 17:47 |
| nuclear_eclipse | paulr: updated | 17:47 |
| paulr | ahh youd alrady done db update? | 17:47 |
| nuclear_eclipse | I have now ;) | 17:48 |
| nuclear_eclipse | I still haven't 'fixed' the database for the portugese problem though | 17:48 |
| paulr | that requires an upgrade step | 17:49 |
| paulr | or | 17:49 |
| paulr | we fix code to fallback to english | 17:49 |
| paulr | depends if we want to be nice to users or just work | 17:49 |
| nuclear_eclipse | I thought it already did.... | 17:49 |
| paulr | well we'll see I guess | 17:49 |
| nuclear_eclipse | lang_api should fall back to english for any language string not available in the requested language | 17:50 |
| paulr | [Wed Apr 08 13:49:29 2009] [error] [client 200.93.147.154] script not found or unable to stat: /var/www/git/gitweb.cgiinclude | 17:50 |
| paulr | [Wed Apr 08 13:50:08 2009] [error] [client 87.104.14.44] PHP Notice: A feed could not be found at http://twitter.com/statuses/user_timeline/7199732.rss in /var/www/html/main/files/simplepie.inc on line 1780, referer: http://projekt.kelacom.com/view.php?id=2 | 17:50 |
| paulr | i need a large database | 17:56 |
| paulr | hmm | 17:56 |
| paulr | nuclear_eclipse: you gone? | 18:38 |
| nuclear_eclipse | nope | 18:39 |
| paulr | nuclear_eclipse: have you seen the mantisenum issue stuff? | 18:42 |
| nuclear_eclipse | no | 18:43 |
| paulr | nuclear_eclipse: http://www.mantisbt.org/bugs/view.php?id=10170 | 18:43 |
| nuclear_eclipse | oh that one | 18:45 |
| nuclear_eclipse | I remember it now | 18:45 |
| nuclear_eclipse | that is a *giant* workflow.... | 18:46 |
| * paulr is gonna look at that now though | 18:46 | |
| paulr | nuclear_eclipse: ok, we *need* that patch ;p | 18:51 |
| paulr | nuclear_eclipse: are the custom function stuff 'legacy'? | 18:53 |
| CIA-18 | Mantisbt: paul * r35759cea4669 /core/columns_api.php: Perf: use strncmp over strpos | 19:05 |
| CIA-18 | Mantisbt: paul * r17093830cee6 /print_all_bug_page_word.php: cache config_get calls | 19:05 |
| CIA-18 | Mantisbt: paul * rb6b90f672bd9 /core/classes/MantisEnum.class.php: Fix: #0010170: MantisEnum class is inefficient when dealing with more than a few tuples | 19:05 |
| CIA-18 | Mantisbt: jreese * r8940888b8f51 /core/string_api.php: Fix #10228: Email linking boundaries changed to whitespace, to prevent linking in the middle of a URL. | 19:33 |
| paulr | hmm | 19:43 |
| nuclear_eclipse | hi giallu | 19:45 |
| giallu | hi nuclear_eclipse | 19:57 |
| giallu | so, what's next? | 20:05 |
| daryn | hi giallu | 20:16 |
| giallu | hey daryn :) | 20:21 |
| --- scribe started --- | 20:43 | |
| paulr | I type xm restart then went off to play game an hour ago ;p | 20:43 |
| daryn | see, i told you! | 20:44 |
| nuclear_eclipse | lol | 20:44 |
| nuclear_eclipse | paulr! | 20:44 |
| nuclear_eclipse | jreese@iowa[/space/workspace/plugins/source-integration] git push < 16:31 | 20:44 |
| nuclear_eclipse | ssh: connect to host git.mantisforge.org port 22: Connection refused | 20:44 |
| nuclear_eclipse | fatal: The remote end hung up unexpectedly | 20:44 |
| nuclear_eclipse | you really need to set that up to start automatically.... | 20:44 |
| Enisseo | does anyone know on which page the notice.gif is used? | 21:12 |
| moto-moi | Enisseo: try 'grep' ? | 21:33 |
| Enisseo | moto-moi: tried on eclipse, no result (but I'm sure I've seen this icon somewhere...) | 21:33 |
| Enisseo | anyway, I've made some changes to my graphic design of mantis (http://enisseo.net/mantisbt/ login/mdp=test/test) | 21:34 |
| moto-moi | nice :P | 21:34 |
| moto-moi | don't forget to delete the admin directory ;) | 21:34 |
| Enisseo | if you want a patch one day... ;) | 21:34 |
| moto-moi | I don't think it will ever meet out current design -> http://tweakers.net/ | 21:35 |
| Enisseo | doing that right now | 21:35 |
| moto-moi | ;) | 21:35 |
| Enisseo | moto-moi: it's mostly css | 21:35 |
| Enisseo | just a few minor changes to the html | 21:36 |
| moto-moi | well, try to get it upstream :) | 21:36 |
| Enisseo | i should get git to do so, I guess | 21:37 |
| moto-moi | not a clue, but I guess they'll take ordinary patches as well :) | 21:37 |
| Enisseo | like svn patches? | 21:38 |
| moto-moi | no, diff :) | 21:38 |
| Enisseo | what does this format look like? | 21:39 |
| moto-moi | http://en.wikipedia.org/wiki/Diff :) | 21:39 |
| Enisseo | yey, what I thought: I use Windows :S | 21:40 |
| moto-moi | Ah | 21:40 |
| moto-moi | poor you :P | 21:40 |
| moto-moi | but afaik eclipse can also make these kind of diffs | 21:40 |
| moto-moi | or else someone like paulr can tell you what kind of patches they like best | 21:40 |
| Enisseo | And then I send this patch to someone like paulr? Or is there a special way? | 21:41 |
| daryn | unified diff | 21:42 |
| daryn | attached to a bug | 21:42 |
| Enisseo | okay, thank you daryn | 21:43 |
| Enisseo | I'll try to do so after a long sleep | 21:43 |
| daryn | :) | 21:43 |
| Enisseo | bye everyone! | 21:44 |
| moto-moi | bye :) | 21:44 |
| paulr | nuclear_eclipse: isn't it @reboot /root/sshd.sh | 21:46 |
| nuclear_eclipse | I don't know, you're the one who set up the box.... | 21:47 |
| nuclear_eclipse | I have no clue how you set up anything on it | 21:47 |
| paulr | i mean, I have that in a crontab | 21:47 |
| nuclear_eclipse | why do you have that in a crontab rather than a some sort of init level system service? | 21:48 |
| paulr | true | 21:54 |
| paulr | I was lazy | 21:54 |
| paulr | :) | 21:54 |
| paulr | i'm also used to not being root needing to run something | 21:54 |
| paulr | nuclear_eclipse: trying to work out how to improve print_all_bug_page_word | 22:03 |
| paulr | should username's be hyperlinked when going to a word report? | 22:12 |
| paulr | ;/ | 22:28 |
| nuclear_eclipse | paulr: doesn't matter to me how the word file is created, but I doubt it matters enough to hypelink usernames | 22:29 |
| paulr | why did we use 1970 for null due dates ( | 23:16 |
| paulr | :( | 23:16 |
| paulr | nuclear_eclipse: i need to fix date handling ;/ | 23:40 |
Generated by irclog2html.py