Sunday, 2009-05-03

« Saturday, 2009-05-02 Index Monday, 2009-05-04 »
../irclogs/#mantishelp.2009-05-03.log
--- scribe started ---00:00
giallupaulr, why don't you leave tests alone... now I've to fix conflicts :/12:57
paulrI didn't change much12:59
paulrI changed enough to fix the fact it broke build12:59
paulr:)12:59
giallupity I touched the same lines ;)12:59
paulr;/12:59
giallumerge not hard fortunately13:00
paulrI just did what phpunit do13:00
paulrto allow config of tests :)13:00
paulraka define a global variable in xml config13:00
gialluok, I'm just adding a check for the soap extension13:01
CIA-31Mantisbt: paul * rcb23b913521b /admin/check.php: Reorganise check.php a bit more13:01
CIA-31Mantisbt: paul * r60d5c864760b / (admin/check.php config_defaults_inc.php): Add check for valid variables_order13:01
CIA-31Mantisbt: paul * r36e64b319026 / (admin/check.php core/adodb/adodb.inc.php): Add WARN option for checks13:01
paulrgiallu: phpunit themselves seem to define global variables to specify test settings13:02
paulrput them in a config.xml13:02
gialludidn't read that part of their manual13:02
paulrso atm, I have an xml file that looks like:13:02
paulr<?xml version="1.0" encoding="UTF-8"?>13:02
paulr<phpunit>13:02
paulr  <php>13:02
paulr    <var name="MANTIS_TESTSUITE_SOAP_ENABLED" value="false" />13:02
paulr  </php>13:02
paulr</phpunit>13:02
gialluright now, I'm trying to imagine a workaround for the massive brokenness of the core.php concept13:03
gialluoitherwise testing just strings is a pain13:03
gialluif possible at all13:04
paulryea, I started looking at that before13:04
paulras I wanted to include all files13:04
paulras part of the test13:04
paulrfor code-coverage13:04
aptituz_uargh13:06
paulr?13:06
aptituz_yet another change in adodb :D13:06
paulr?13:06
paulraptituz_: what you whinging about *now*?13:07
aptituz_i'm not whinging13:08
paulrthe change should probably help you ;p13:08
paulrpotentially13:08
paulraka for pgsql/mssql where we've needed to modify adodb, require bundled one13:09
paulrfor mysql dont require bundled one, but do a version check13:09
paulrwhere i've set version check at 5.05 or something (which is like 2 years old)13:10
aptituz_and you modified core/adodb/adodb.inc.php for this?13:10
paulrto add a "MantisBT Version" string to adodb version13:10
aptituzah, ic13:10
paulr(the only other option I could think of was use the bundled adodb by default over the system one and add a config file optino you can specify if you want to use the system one over the bundled)13:11
aptituzso nothing changed for me. it still hurts me bad, that I cannot support pgsql ;)13:11
paulrwell, the other option is13:12
paulrwe copy/paste the adodb datadict code that we've needed to modify13:13
paulrinto a mantis function that we call instead13:13
paulrbut i'm not really sure that is any better13:13
paulrtake core/13:13
paulradd core/adodb13:14
paulrcp core/adodb/datadict core/datadict13:14
paulrthen depending on how you look at it13:14
paulryou've got 2 copies of adodb in mantis to 'package'13:14
aptituzwell the best would be to push the changes upstream or whatever is neded to not diverge from upstream13:15
paulrADOdb 5.08 Released - modified: 17-04-2009 02:27:30 PM13:15
paulron the 18th of April (1 day later) I posted:13:15
paulrhttp://phplens.com/lens/lensforum/msgs.php?id=1794613:15
paulrhttp://phplens.com/lens/lensforum/msgs.php?id=1794713:15
paulrhttp://phplens.com/lens/lensforum/msgs.php?id=1794913:15
paulrhttp://phplens.com/lens/lensforum/msgs.php?id=1794813:16
paulrof those posts13:16
paulr1 got a reply on the 30th april13:16
paulrthe other 3 have no reply13:16
paulrI decided to check adodb.sf.net download page13:16
aptituzhmm13:16
paulrand saw that there'd be a 5.08a release crediting me for pointing out 2 of the problems i'd whigned about13:17
paulraka a parse error in datadict13:17
paulrand fact adodb 5.08 didn't modify version string13:17
giallupaulr, haven't they a development mailing list wqere you can beat them with a cluestick?13:18
aptituzyeah I also wonder if that forum is the best way to contact upstream13:18
giallupaulr, suggest them to use a bug tracker...13:19
aptituzbut the page says "yes"13:19
aptituz"For bug reports, feature requests and questions, use the ADOdb Forums. The author monitors this forum."13:19
gialluTBH13:19
paulrfrom what I can tell13:19
paulradodb is one guy13:19
gialluthe fact "author" is singular, I wonder if we are really smart continuing to rely on adodb...13:20
gialluyeah13:20
paulr'yes' because it's fairly stable (kinda)13:20
paulrand there's not that many other alternatives13:20
paulr(that work for what we do)13:20
paulrotoh13:21
paulrgiven we only use a subset of adodb13:21
paulrand we only use a subset of database engines13:21
paulrif we were to 'copy/paste adodb'13:21
paulrand delete all the bits we dont use13:21
paulrwe'd probably speed it up13:21
paulrand reduce the ram requirements of it13:21
paulrbut then we could do the same to phpmailer and whatever else13:22
paulrand there's a reason why we dont :)13:22
paulr(remember there's someone that came up with 'adodb lite' for that purpose - http://adodblite.sourceforge.net/benchmark.php13:23
paulrincidently13:36
paulrthink you save a few milliseconds by specifying13:36
paulr$ADODB_CACHE_CLASS = '';13:36
paulr$ADODB_CACHE = true;13:36
paulrbefore starting adodb13:36
CIA-31Mantisbt: paul * r965959b05a05 /core/print_api.php: Remove extra charset param13:41
paulrgiallu: how you getting on making the url stuff into phpunit?13:42
gialluI prepared the code, but until now I can't make it work properly.13:51
paulroh?13:51
gialluyou know the "include a single file and that pull in the entire kitchen sink" sindrome13:51
paulrignore that a sec13:52
paulrdid you move the thingies to tests?13:52
paulri.e. test_string_sanitise_url file contents13:52
gialluyes13:52
paulri'd be inclined to commit those then13:52
paulr$t_lang_var = preg_replace( '/^s_/', '', $t_var );13:53
paulrif( $t_lang_var != $t_var ) {13:53
paulr$g_lang_strings[$p_lang][$t_lang_var] = $$t_var;13:53
* paulr wonders why we dont just store s_ in lang files and add s_13:53
paulrthe kitchen sink syndrome as you put it is something  that we've already look at a bit13:55
gialluwe need some more bits then ;)14:01
paulrbut I know the problem you are hitting14:02
paulras I observed it whilst trying to get phpuc set up14:02
paulrsome things e.g. phpunit can be told to include code14:02
paulrsome of our core classes execute errors/code14:02
giallupaulr, did you think about any strategy to fix it?14:03
paulrsledgehammer14:03
paulr:14:03
paulrcore\bug_api.php(53): * requires twitter_api14:04
paulrcore\bug_api.php(55): require_once( $t_core_dir . 'twitter_api.php' );14:04
paulri'd like to fix things like that initially14:04
paulr( I dont use twitter, i'm not sure why bug api would require it at work14:04
paulrbut then something like twitter should be a plugin14:05
paulrand the plugin api would handle it for us14:05
gialluAFAIK twitter is a plugin14:06
gialluthat should be something to remove14:07
giallubut the problem is not moving stuff to plugins14:07
gialluthe problem is writing better, self contained stuff14:07
paulrno, but the fact that as well as a twitter plugin14:07
paulrwe have a twitter api14:08
paulrthat's in bug/news api14:08
paulrturns into a mess quickly14:08
paulrgraphs(jpgraph)14:08
paulrgraphviz14:08
gialluoki, so it seems we don't have any plan right now14:08
paulrtwitter14:08
paulrneed to be plugins14:08
giallutbh, I think the first step is to trim core.php, it should really avoid including core/*14:10
gialluand then I'm grateful you nuked the "let's be smart and cache all projects now" from there14:10
paulri'm gonna do a commit that'll make you happy ;p14:10
giallugo14:11
gialluso I'll rebase my local changes now14:11
paulrthere14:11
CIA-31Mantisbt: paul * r6e73eeecb3df /core/print_api.php: I don't think it does require... :)14:11
paulr1 less :P14:11
gialluplease... no riddles in commit messages ;)14:12
gialluyay14:12
paulrto be fair though14:12
paulri'm wondering if I added that by mistake14:12
paulrerror_api doesn't have any dependencies btw14:13
paulror has the absolute minimum14:13
gialluwell. IMHO the goal is to remove all "require" lines from core.php first14:14
paulrthat's not really feasible14:15
paulrat least14:15
gialluthen the same from individual core/*.php files14:15
paulrat least at some point, something needs to include config_api14:15
gialluyes, php will do14:15
gialluor better14:16
* paulr wonders if we really need timer_api14:16
gialluwe'll do that from the core/*.php files after step 114:16
gialluphp will do that after step 214:17
* paulr is tempted to delete timer_api14:17
gialluuhm14:25
gialluwhat different between:14:25
gialluif ( file_exists( dirname( __FILE__ ).DIRECTORY_SEPARATOR.'custom_constant_inc.php' ) ) {14:25
giallurequire_once( dirname( __FILE__ ).DIRECTORY_SEPARATOR.'custom_constant_inc.php' );14:25
giallu}14:25
gialluand14:25
gialluinclude_once( dirname( __FILE__ ).DIRECTORY_SEPARATOR.'custom_constant_inc.php' );14:26
paulrI think include once will generate a warning for missing file14:28
gialluright...14:30
CIA-31Mantisbt: paul * r317e06dca504 / (5 files in 2 dirs): Remove timer api (we dont use it), and I think core dev's tend to use xdebug to look at performance issues now.14:30
gialluyay for xdebug14:30
CIA-31Mantisbt: paul * r734dc2c32707 / (4 files in 2 dirs): Remove microtime_float function.14:33
CIA-31Mantisbt: paul * rbd9b85446f01 /core.php: Remove microtime_float function14:33
giallustringTest runs! now let's see ho much damage I made14:47
siebrand[KK]Kirill: ping?15:15
[KK]Kirillpong15:16
siebrand[KK]Kirill: can you please test http://www.mantisbt.org/bugs/view.php?id=8763 ?15:16
siebrand(on git master)15:16
[KK]Kirillone sec15:16
[KK]KirillI'l test this with john string-branch15:17
[KK]Kirillone sec.. I'l try pure master-branch15:17
siebrand[KK]Kirill: trunk should be fine... Can only mark it fixed when fixed there...15:17
* siebrand nods.15:17
[KK]Kirillall ok15:31
[KK]Kirillsiebrand: close this task.15:33
siebrand[KK]Kirill: great.15:34
[KK]Kirillsiebrand: can you push in master last russian translate?15:34
siebrand[KK]Kirill: have patience :)15:35
siebranddarn, trying to set up roadmapping. No idea what to do... I though this would work out of the box for an admin level account, but it appear nwo.15:35
siebrandnot15:35
* paulr gets back15:54
CIA-31Mantisbt: s.mazeland * rf2c854a349dd / (3 files in 3 dirs): Remove 'charset' message that is no longer used. Replaced by 'utf-8'.16:02
CIA-31Mantisbt: s.mazeland * r1096c049774b / (changelog_page.php roadmap_page.php): * remove some unneeded strtolower()16:02
* paulr looks at siebrand commits16:03
* paulr ponders16:03
paulrsiebrand: are those in trunk??????16:03
siebrandpaulr: what exactly?16:03
siebrandstrtolower? Yes, there are a bunch left...16:03
paulrlang_get( 'word_separator' ) ? :)16:04
paulrwhat's that16:04
siebrandpaulr: well, that's a hardcoded space in most languages, nothing in some.16:04
paulrbut yea well spotted I must have missed that charset :)16:05
CIA-31Mantisbt: paul * ra2146d2a8b13 / (core.php core/email_api.php): autload disposable + phpmailer classes16:14
CIA-31Mantisbt: s.mazeland * rc81c3171c79f / (core.php core/email_api.php): Merge branch 'master' of git://mantisbt.org/mantisbt16:14
CIA-31Mantisbt: s.mazeland * r2fbf92c477f6 /lang/ (40 files): Localisation updates from http://translatewiki.net (2009-05-03 16:04 UTC)16:14
siebrandpaulr: hmm, where does rc81c3171c79f come from?16:15
paulryou :)16:15
siebrandpaulr: but I didn't change anything.16:15
paulr17:17 < CIA-31> Mantisbt: paul * ra2146d2a8b13 / (core.php core/email_api.php): autload disposable + phpmailer classes16:15
paulrbut I did16:15
siebrandpaulr: did I somehow accidentally revert your commit?16:15
paulrdunno lets see :)16:16
paulrI think it's jsut merged them16:16
siebrandhmm, I didn't make a change there.16:17
* siebrand is confused.16:17
paulryou made a change16:18
paulri made a change16:18
paulrand we probably both pushed close together16:18
paulrso it's merged16:18
paulrwould be my guess ;p16:18
siebrandok. just as long as it works, and there are no surprises :)16:19
siebrand[KK]Kirill: Russian was updated (only 4 messages updated, though)16:19
[KK]Kirillsiebrand: thanks16:26
[KK]KirillToday I have celebration17:13
[KK]KirillMy daughter 7 month old17:14
paulrcongrats :)17:17
CIA-31Mantisbt: paul * r49a5c45ac7f2 /core/ (database_api.php session_api.php): Session api requires gpc, database api does not require gpc.17:18
[KK]Kirillpaulr: http://yfrog.com/b5img6460p17:33
paulr:)18:54
paulrsiebrand: ?19:23
paulrbug 1043419:23
mantisbotBug 10434 - Chi-Yu - open - acknowledged19:23
mantisbotallow localized decimal seperator - http://www.mantisbt.org/bugs/view.php?id=1043419:23
paulryou seen that?19:23
siebrandpaulr: yes, I think I requested a patch19:24
paulroh heh sorry19:24
siebrandpaulr: guess I may look into adding that one day... MediaWiki has formatNum() in the Language object.19:25
paulri'm wondering whether we could do another 1.2 release soon19:25
siebrandpaulr: are we feature comp[lete yet? Then we could go into beta...19:26
paulrI want/need to break dates19:27
siebrandpaulr: roadmap shows only at 37%19:27
siebrandpaulr: or maybe plan 1.2.0 with only the criticals from the current 1.2.x planning?19:27
siebrand(afk)19:28
paulr4084 = semi-fixed19:30
paulr9081 = we've changed code not not use email links but link to user mainly19:30
paulrhmmmmmmmmmmmmmmmmm22:19
« Saturday, 2009-05-02 Index Monday, 2009-05-04 »

Generated by irclog2html.py