Monday, 2009-06-22

../irclogs/#mantishelp.2009-06-22.log
--- scribe started ---00:00
mantisbotNew bug: Bug 10610 - eliasn - open - new03:02
mantisbotNew bug: Include information about test execution - http://www.mantisbt.org/bugs/view.php?id=1061003:02
mantisbotNew bug: Bug 10611 - Kirill - open - new07:33
mantisbotNew bug: Attach not recognize correctly - http://www.mantisbt.org/bugs/view.php?id=1061107:33
dhx_mlooking towards the 1.2 release, there seems to be quite a few TODO things specifically marked for 1.2 in database_api08:08
CIA-17Mantisbt: hickseydr * rb77fcd42ffb6 /docbook/developers/en/contributers.sgml: Update contributor guide for pending 1.2 release08:10
[KK]Kirilldhx_m: lo08:12
dhx_m[KK]Kirill: hi08:13
[KK]Kirilldhx_m: Something strange08:13
[KK]KirillCan you take me your time?08:13
dhx_mwhats up? :)08:14
[KK]Kirillabout bug 1061108:14
mantisbotBug 10611 - Kirill - open - new08:14
mantisbotAttach not recognize correctly - http://www.mantisbt.org/bugs/view.php?id=1061108:14
dhx_mhmm that's odd08:14
[KK]Kirilllook08:15
dhx_mthe live site at mantisbt.org uses the latest version08:15
[KK]Kirillcore/file_api.php08:15
dhx_mso I'd expect it to be a problem there too08:15
[KK]Kirillfunction file_get_extension( $p_filename ) {08:15
[KK]KirillI add string echo $p_filename;08:16
[KK]Kirillresult SQL стратам.png08:16
[KK]Kirillafter $i = utf8_strlen( $p_filename ) - 1;08:16
[KK]KirillI place string08:17
[KK]Kirillecho $p_filename[$i];08:17
[KK]Kirillresult ?08:17
[KK]Kirillsomething strange with one char of string08:17
[KK]Kirilland other chars too08:18
dhx_myou're right, that function is broken with respect to UTF808:18
dhx_mI'll see if I can fix it08:19
[KK]KirillI think more function broken. I have problem such this in other function, but I forgot where..08:19
[KK]KirillIf you can solve this, I try found and fix other promblem08:20
dhx_mok08:20
CIA-17Mantisbt: hickseydr * r3758e6f12788 /library/README.libs: Update libraries README08:35
CIA-17Mantisbt: hickseydr * r96814088450f /lang/langreadme.txt: Remove deprecated information from langreadme08:50
[KK]KirillOh.. I remember where I can see that08:56
[KK]KirillI see in geshi08:56
dhx_mwe use geshi? :D08:56
[KK]KirillI think it's in my server broke :)08:57
[KK]Kirillnot08:57
[KK]KirillI use geshi in pastebin08:57
paul_+ezc             | ez Components   | 2008.2.3 | unpatched09:00
paul_dhx_m: please add correct version09:00
paul_dhx_m: a) replace file_get_extension with a regex09:03
[KK]Kirillpaul_: hi09:04
paul_b) it's probably safe to use strlen itself given how we're trying to use it there09:04
paul_hi09:06
[KK]Kirillpaul_: why it's can be?09:07
[KK]Kirillstring show normal, but each char broken in utf-8?09:07
paul_I think i'd expect that to break09:10
paul_instead of getting strlen, we're getting utf_strlen09:10
paul_then recursing through each char of the string09:10
paul_(which if php isn't treating as utf or something09:10
paul_either way09:11
paul_strikes me as something that could be written as a regex09:11
paul_we only *need* the last N chars before a '.'09:11
paul_normally 2 / 3 / 4 chars e.g. .html /.htm09:11
[KK]Kirillpaul_: I ask in russian forum09:22
[KK]Kirillthey answer that utf have 2 byte in string09:22
[KK]Kirill$p_filename[$i] - 1 byte09:23
[KK]KirillI test in bug 1061109:26
mantisbotBug 10611 - Kirill - open - new09:26
mantisbotAttach not recognize correctly - http://www.mantisbt.org/bugs/view.php?id=1061109:26
[KK]Kirillthis file equal, but diff names09:26
[KK]KirillOh I fix it09:30
[KK]KirillI change $p_filename[$i] to utf8_substr($p_filename,$i,1)09:30
dhx_mback10:03
dhx_mpaul_: what is the correct version?10:03
dhx_mpaul_: I just guessed you used the latest stable... I take it you used latest unstable?10:05
dhx_mdo we have utf8 regex?10:05
paul_erm10:05
paul_kinda10:05
paul_I used latest svn :P10:06
dhx_mahhh :)10:06
paul_which I think actually corresponds to 2009.1 beta110:06
dhx_msounds good to me :)10:06
paul_I got warned there was a commit in svn relating to large number of data points that I'd want)10:06
paul_but in terms of file extension10:07
paul_I reckon we'd be better doing a regex with /u or something10:08
paul_it strikes me as a complicated function to get what (normally) will be last 3 chars of a string10:08
dhx_mregex seems like high overhead10:09
dhx_manyway, I disagree with using the file extension10:10
dhx_mwhen we should be using the mime type10:10
paul_tbh, regex might be less10:18
paul_but yea10:18
paul_that's another point10:18
* paul_ thinks10:18
dhx_mI think Regex is probably the best solution10:19
dhx_mbecause it is easiest10:19
dhx_malthough technically it'd be better to do a backwards search through the string until you hit the first period (.)10:19
dhx_massuming the file has an extension! :)10:19
mantisbotNew bug: Bug 10612 - rombert - open - new10:24
mantisbotNew bug: mc_project_get_issue_headers returns an empty response (this bugtracker) - http://www.mantisbt.org/bugs/view.php?id=1061210:24
mantisbotNew bug: Bug 10613 - nuan burke - open - new12:04
mantisbotNew bug: Cannot Filter by Project - http://www.mantisbt.org/bugs/view.php?id=1061312:04
nuclear_eclipsemorning all12:23
[KK]Kirilllo12:23
[KK]KirillJohn, how I can see what diff curent file with commit state?12:23
[KK]Kirillin git12:24
nuclear_eclipse`git diff filename` ? :P12:24
[KK]Kirill:) it's work. thanks12:27
mantisbotNew bug: Bug 10614 - Kirill - open - new12:30
mantisbotNew bug: assigned issue not assigned - http://www.mantisbt.org/bugs/view.php?id=1061412:30
dhx_mhi nuclear_eclipse12:33
nuclear_eclipsehi dh12:33
nuclear_eclipsex12:33
paul_nuclear_eclipse: I HATE PAGE NUMBERS13:27
[KK]Kirillpaul_: You implement new bug class?13:29
[KK]Kirillor not?13:29
nuclear_eclipsehmm, what?13:32
paul_you dont want to know13:32
[KK]Kirillbug 10614. I see, somebody did hard work to change bug class use13:32
mantisbotBug 10614 - Kirill - open - new13:32
mantisbotassigned issue not assigned - http://www.mantisbt.org/bugs/view.php?id=1061413:32
[KK]Kirillmay be my issue is result of this work?13:34
nuclear_eclipse[KK]Kirill: I can't replicate the issue13:35
[KK]KirillCan you create test project and give me manager right on it?13:37
[KK]Kirillin your issue?13:37
[KK]Kirillin your mantis13:37
nuclear_eclipsegimme a bit to update my live site to latest git master13:40
[KK]Kirillnuclear_eclipse: may be I create test project?13:40
[KK]KirillI always use last git13:41
dhx_mit's paul_ 's fault for sure13:42
dhx_m:p13:42
nuclear_eclipse[KK]Kirill: perhaps it's a configuration issue?13:43
dhx_mnuclear_eclipse: what do we do with UTF8 regex matching?13:54
nuclear_eclipsedepends13:54
dhx_madd /u ?13:54
nuclear_eclipsethat's only part of it13:54
nuclear_eclipsea) /u will allow regexs to work on multibyte strings, b) to actually match utf8 characters, we need to be matching against more than a-zA-Z0-9...13:55
dhx_mtrue13:55
dhx_malthough I'd say your regex is wrong if you're using ranges like a-Z13:56
nuclear_eclipsewell, it all depends on the context...13:57
dhx_mI just don't see why you'd want to match against \d instead of 0-913:58
dhx_mI mean... match against 0-9 instead of \d13:58
nuclear_eclipseI'm lost13:59
nuclear_eclipseoh, nvm13:59
nuclear_eclipseyour wording bugged out my brain :P13:59
dhx_mwell I would assume that matching against \w would match any character that is a "word character" in the native UTF8 locale14:00
dhx_mwhereas a-z specifically only matches a-z in the English locale14:00
nuclear_eclipseit's complicated, let's just put it that way :P14:00
[KK]Kirilldhx_m: I write how I fix it14:01
[KK]KirillI use utf8_substr()14:01
dhx_mso if we're using regex on utf8 strings... it'll be bad?14:01
nuclear_eclipseyes/no - depends on what we're trying to do with it14:02
dhx_mnuclear_eclipse: take for example plugin_file.php:3614:03
dhx_mthat would prevent someone from using a Unicode filename14:03
nuclear_eclipseright14:03
nuclear_eclipsebut at the same time, since many filesystems don't handle utf filenamse correctly anyways, I'm not considering that a giant loss atm...14:04
dhx_mfor no real reason at all? (unless PHP file functions aren't UTF8 compatible... which I doubt)14:04
dhx_mon Windows I imagine it all completely breaks14:04
nuclear_eclipseright14:04
nuclear_eclipsethat's the main reason I made that decision14:04
dhx_mbut then we'd need to prevent all attachment names from containing Unicode characters?14:05
nuclear_eclipseWindows tends to break everything, so if we simply say that all plugins must be using ascii names, we ensure they'll work on any install14:05
nuclear_eclipseand we don't enforce utf8 on attachments because we store them in the database by default14:05
[KK]Kirillnuclear_eclipse: I talk about attach event14:06
nuclear_eclipseif someone wants to put attachments on disk, they will probably understand the problems involved14:06
nuclear_eclipsedhx_m: http://www.php.net/manual/en/regexp.reference.php14:06
nuclear_eclipsesearch for "utf"14:06
dhx_myep14:07
nuclear_eclipseit would seem to me that \w is not utf-8 compatible14:07
nuclear_eclipsewe'd have to start doing \pL and so forth14:07
dhx_mA "word" character is any letter or digit or the underscore character, that is, any character which can be part of a Perl "word". The definition of letters and digits is controlled by PCRE's character tables, and may vary if locale-specific matching is taking place. For example, in the "fr" (French) locale, some character codes greater than 128 are used for accented letters, and these are matched by \w.14:08
nuclear_eclipseright, that's my point14:08
dhx_malso they provide \C specifically to match single bytes, so I imagine \w would be UTF-8 compatible?14:08
nuclear_eclipseyou can't rely on the locale14:08
nuclear_eclipseeg, if I set my server locale to iso8859, pcre won't match utf8 characters with \w14:09
nuclear_eclipsePHP's a sticy mess when it comes to locales and such14:09
dhx_m"Matching characters by Unicode property is not fast, because PCRE has to search a structure that contains data for over fifteen thousand characters. That is why the traditional escape sequences such as \d and \w do not use Unicode properties in PCRE. "14:10
dhx_mah14:10
nuclear_eclipselike I said, it's complicated :P14:10
dhx_mthat documentation is confusing... I guess the part about French accents was in reference to old school code pages for characters > 0d12714:10
nuclear_eclipseright14:11
* Reapazor looks nervously around -- i see were getting ready for a release...but api still is broken ? :/14:11
nuclear_eclipsewhat api is still broken? soap you mean?14:11
Reapazorya14:12
Reapazoreverything still coming back with access denied14:12
Reapazor(works in stable release 1.1)14:13
Reapazorkeeps throwing an access denied exception to a user with full rights :)14:13
dhx_mnuclear_eclipse: what about $g_user_login_valid_regex = '/^[\w \-]+$/';14:13
nuclear_eclipseReapazor: I think either a) something is broken in your configuration, or b) your client is not up to date with any changes that may be in 1.214:13
dhx_mnuclear_eclipse: is that a candidate for UTF-8 regex?14:14
nuclear_eclipsedunno14:14
dhx_mI imagine you need mbstring though :(14:14
nuclear_eclipseReapazor: bug 10612 acts as though everything works fine except for a couple functions themselves, so I can't imagine that it's totally broken...14:15
mantisbotBug 10612 - rombert - open - assigned14:15
mantisbotmc_project_get_issue_headers returns an empty response (this bugtracker) - http://www.mantisbt.org/bugs/view.php?id=1061214:15
nuclear_eclipsedhx_m: iirc, pcre's utf8 does not use mbstring, like mentioned, it has its own lookup tables for multibyte characters14:15
dhx_mnuclear_eclipse: ugh, what a mess :p14:16
nuclear_eclipseutf8 is always a mess in any language that wasn't built from scratch to make it not be a mess14:16
dhx_mReapazor: does 1.2.0a3 work at all?14:17
nuclear_eclipseeven in python 3, which supports utf8 from the ground up, utf8 is still a bit tricky at times14:17
dhx_myep14:17
dhx_monce the world moves to UTF-8 it'll be a much happier place though :)14:17
dhx_mfor instance, being able to match things like \Sc for a currency symbol is very cool :)14:18
Reapazori just pulled from git repo this morning14:19
nuclear_eclipseno, once the world move completely to ASCII, it'll be amuch happier place :P14:19
Reapazorstill no dice14:19
Reapazorhave things changed in the auth process /14:19
Reapazorfor soap ?14:19
nuclear_eclipseReapazor: not afaik14:19
dhx_mReapazor: 1.2.0a3 is from months ago, so I was wondering if any of the recent fixes caused the problem14:19
dhx_mReapazor: can you also try with $show_detailed_errors = ON in your config file?14:20
Reapazoroh i tried the dev release on the site as well14:21
Reapazorno it did not work14:21
Reapazorand i have show detailed erros on :)14:21
Reapazortho do the detailed errors pass back thru soap?14:22
dhx_mjust a blank output?14:22
dhx_mactually they don't... so ignore that request :)14:22
Reapazorwhy oh why wont aptana search thru files ...14:24
Reapazormc_projects_get_user_accessible14:24
Reapazorgotta find where that is14:24
Reapazorthats the last function listed before the exception14:24
dhx_m[Mon Jun 22 13:24:13 2009] [error] [client 127.0.0.1] PHP Fatal error:  Cannot use object of ty14:27
dhx_mpe BugData as array in /srv/www/htdocs/users/robert/mantis12x/api/soap/mc_filter_api.14:27
dhx_mphp on line 15914:27
nuclear_eclipseReapazor: bug 10612 is reported against that function14:27
mantisbotBug 10612 - rombert - open - assigned14:27
mantisbotmc_project_get_issue_headers returns an empty response (this bugtracker) - http://www.mantisbt.org/bugs/view.php?id=1061214:27
dhx_mwas that the exception you reported?14:27
dhx_mah I see the problem14:28
Reapazori dont know if its exactly tht14:29
Reapazorill paste the exception from .net14:29
Reapazorsec14:29
dhx_mfilter_api.php:1986 returns the result of bug_row_to_object14:29
dhx_mso we're actually returning an array of bug objects now14:29
dhx_mfrom what I gather14:29
Reapazorhttp://pastebin.ca/146989714:29
dhx_mwhereas in 1.1 we would of been returning a normal array from the database14:29
Reapazorahhh....14:29
dhx_mand it seems not all part of Mantis have been updated to work with the new bug objects14:30
nuclear_eclipsepaul_ said he'd be investigating the issue, so hopefully we'll be seeing a fix before I roll the rc release14:31
dhx_malthough you must be experiencing another problem if it also occurs on 1.2.0a3?14:32
Reapazorwill it change the outward facing functions14:32
dhx_mthe conversion to use objects was made on 2009-06-1114:32
Reapazorala will the code im using need changes ?14:32
nuclear_eclipseshouldn't, afaik14:32
dhx_mnope14:32
dhx_mI might even be able to fix it... shouldn't be hard14:32
nuclear_eclipsebut monitor bug 10612, and see what paul fixes14:32
mantisbotBug 10612 - rombert - open - assigned14:32
mantisbotmc_project_get_issue_headers returns an empty response (this bugtracker) - http://www.mantisbt.org/bugs/view.php?id=1061214:32
dhx_mjust use bug->variable14:33
dhx_minstead of bug['variable']14:33
dhx_mReapazor: in fact, can you please try that... at mc_filter_api.php:15914:33
dhx_mplease change $t_id = $t_issue_data['id'];14:33
dhx_mto14:33
dhx_m$t_id = $t_issue_data->id;14:33
Reapazorsure second ;)14:35
dhx_mthen check your PHP error logs on the server to see what the next exception is14:35
dhx_mhopefully it is mc_filter_api.php:16414:36
dhx_mI'm not too sure how to test SOAP... otherwise I'd test myself14:36
Reapazorwell i could give you a login to a test page14:37
Reapazorif that helps ya ?14:37
dhx_mreally need to see PHP error logs14:38
dhx_mis there some sort of command line utility I'm meant to use for SOAP?14:38
Reapazornaw ;/14:39
Reapazorironically easiest way is to use .net :)14:39
Reapazorso its not generating an error14:39
Reapazorfor php at all14:39
Reapazorits just coming back with access denied still14:39
Reapazorso maybe i should say this ...14:39
Reapazorthe user is "api"14:39
Reapazorhe has a nice long passwor14:39
Reapazord14:39
Reapazori've logged in with him to make sure he can14:39
Reapazor... any other things i should have setup ?14:39
dhx_mI think you might have a different problem then14:40
dhx_mif you're getting access denied14:40
Reapazorhttp://pastebin.ca/146989714:40
dhx_mare you sure you're sending the username and password correctly?14:40
Reapazorthats the stack return14:40
Reapazor(its a commercial extension for a wiki we use)14:40
Reapazorit worked in prior versions14:40
Reapazorala 1.1 wasnt a problem14:40
Reapazorbut here i m wanting 'bleeding edge'14:41
dhx_m.NET = wall of text! :p14:41
Reapazorlol :)14:41
Reapazorhere basically what the stack exception shows is i go to use that 'user accessible' function14:41
Reapazorand it is coming back with no14:41
Reapazor:)14:41
dhx_mI take it they don't usually have line length limits in coding standards heh14:41
dhx_mah, I'm looking at a different function completely14:42
Reapazorhehe ya im trying to find that function14:42
Reapazorso i can sorta 'cheat'14:42
Reapazorand see whats causing it14:42
Reapazorbut cant ;/14:42
Reapazor*sigh* aptana = fail14:42
Reapazori know you guys have an index online somewhere ...gonna look for that14:42
Reapazorfound the ref :)14:43
dhx_mReapazor: have you set mc_readonly_access_level_threshold and etc to the right values?14:43
Reapazorwell the api user has adminstrator rights atm14:43
Reapazorone would imagine :)14:43
Reapazorhttp://pastebin.ca/146992114:44
Reapazortheres my config :)14:44
dhx_mReapazor: we don't actually define mc_* defaults in config_defaults.php14:45
dhx_mso by default, an administrator user won't have access to SOAP API14:45
dhx_myou have to actually specify it in config_inc.php first14:45
dhx_mwhy it's not in config_defaults_inc.php I'm not sure14:45
Reapazorwell i know i have tried him as just a viewer of the project as well14:46
nuclear_eclipsedhx_m: that's because they're specified in api/soap/mc_config_defaults_inc.php :P14:46
dhx_mah :p14:46
Reapazorif thats the case...14:47
dhx_mand we even have our own config_get in there too :o14:47
nuclear_eclipsenaturally!14:47
Reapazorwhy isnt it working ... when i've adjusted the rights14:47
Reapazorall over the place14:47
Reapazor(the user has gone all over the map - from just a viewer to admin)14:47
Reapazorim adding some debug stuff into get_user to find out :)14:50
Reapazorhehe this is why i love opensource :)14:50
Reapazor$t_user_id = mci_check_login( $p_username, $p_password );14:51
Reapazoris coming back false14:51
dhx_mpaul_: from 10612, I think it'd be a good idea to grep the source code like: grep -Rn "\['id'\]" *14:51
dhx_mpaul_: and make sure all uses have been updated to work with bugobjects (where applicable)14:52
Reapazori ve verified that username/password are being sent correctly14:52
dhx_mhmm ok14:52
dhx_mtry placing a debug output after the auth_attempt_script_login call14:54
dhx_mand before the return14:54
Reapazorworking on it ;)14:55
Reapazorim stepping thru the mci_check_login14:55
Reapazoratm14:55
dhx_m:)14:55
dhx_myou only need to place checks before return statements14:55
Reapazormci_check_login14:55
Reapazoryes :)14:55
Reapazorwoops14:55
Reapazorauth_attempt_script_login( $p_username, $p_password )14:55
Reapazoris returning false...14:55
Reapazoroff to there i go :)14:55
dhx_mok14:55
dhx_mI guess you just keep working down to the final branch that causes the problem?14:57
dhx_mat the moment it is down to: user doesn't exist, user is disabled, passwords don't match14:58
CIA-18Mantisbt: jreese * r30debf44a82e /doc/ (ChangeLog RELEASE): Added release notes and dropped the old changelog14:58
CIA-18Mantisbt: jreese * rc69753cdb4eb /core/bug_api.php: Fix: Auto-assign not working on bug submission14:58
paul_dhx_m: ya, thinking about it14:58
dhx_mhooray for no old changelog :)14:58
paul_nuclear_eclipse: mbstring seems old functionality anyway14:59
dhx_mpaul_: do you know how to test SOAP API easily?14:59
paul_sure14:59
paul_with a sledgehammer14:59
* nuclear_eclipse beats paul_ 14:59
dhx_mpaul_: lol14:59
paul_nuclear_eclipse: I meant the lib's internally14:59
dhx_mmaybe I'll venture into the dark realms of tests/soap14:59
nuclear_eclipseno, I'm beating you for breaking obvious things with your bugobject update...14:59
paul_like?15:00
paul_I dont touch soap api15:00
nuclear_eclipse CIA-18> Mantisbt: jreese * rc69753cdb4eb /core/bug_api.php: Fix: Auto-assign not working on bug submission15:00
nuclear_eclipsehi Glenn15:00
Reapazorim almost there :)15:00
paul_nuclear_eclipse: well that was due to http://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=a5f7dbcd540c533546fc80533866d808d4c64f9d15:01
nuclear_eclipsewell, my point being that when you did the giant bugobject migration, you didn't update everything...15:02
* nuclear_eclipse scolds paulr15:03
paul_that code was added to bugapi after the bugobject migration :P15:03
paul_blame git merge :P15:03
dhx_mtests/soap isn't very useful... it's not a blackbox test15:03
nuclear_eclipseI blame paulr for not fixing merge conflicts appropriately...15:03
dhx_mI've just got to work out this XML format stuff15:04
nuclear_eclipsejust toss in a bunch of angle brackets, and you'll be fine :P15:04
Reapazorwtf15:04
dhx_mhaha15:05
Reapazorits erroring on a password mismatch ...15:05
Reapazorthe password is 100% right15:05
Reapazorso ... now we work with that ;)15:05
dhx_mcheck login_method in config_inc.php15:05
Reapazormd5...15:05
dhx_mbefore going further, can you double check the login by logging into mantis manually with that user/pass?15:06
dhx_mI assume you have already though15:06
Reapazoryes :)15:06
Reapazorim tripple checking right now15:06
Reapazorat that level what it thinks the pw is15:06
Reapazor:)15:06
[KK]KirillJohn, close this bug 1061415:07
mantisbotBug 10614 - Kirill - open - new15:07
mantisbotassigned issue not assigned - http://www.mantisbt.org/bugs/view.php?id=1061415:07
Reapazoryup can login via that ;/15:07
paul_dhx_m: I thought I make make soap something t play with for a bit15:07
Reapazordarn i was hoping it was that15:07
Reapazorok into that function i go15:07
dhx_mpaul_: if you figure out what XML I'm meant to be sending, let me know :p15:07
dhx_mpaul_: otherwise I consider it someone else's problem heh15:07
dhx_mReapazor: can you please pay special attention to the utf8_substr at authentication_api.php:36015:09
dhx_mor one second...15:09
dhx_mnah that is all I can think of15:10
dhx_mohh no15:10
Reapazorheh actually i was looking at that line15:10
Reapazor:)15:10
dhx_mauthentication_api.php:410 is probably the culprit15:10
Reapazorwell that section15:10
Reapazorthe password itself is long15:10
dhx_ma utf8_substr using PASSLEN as the limit doesn't sound good?15:10
Reapazorbut a hash is always the same15:11
Reapazorworking to get there ;)15:11
dhx_myeah15:11
Reapazorsoon i will have your answer15:11
* nuclear_eclipse hates the current auth_api ....15:11
dhx_mI love bugobjects :p15:11
dhx_mjust need to expand it IMO15:12
dhx_muserobjects sounds nice :)15:12
dhx_mget rid of all the 10000000000 user_ functions15:12
dhx_muser_get_* in particular15:12
dhx_mnuclear_eclipse: is PASSLEN the number of bytes or number of characters?15:13
nuclear_eclipsenumber of characters15:13
dhx_mah that's ok15:14
nuclear_eclipseit's the varchar() length15:14
dhx_myep15:14
dhx_mUTF8 safe then15:14
dhx_m... maybe?15:14
nuclear_eclipseafaik, yes15:14
ReapazorHTTP status 302: Found15:15
Reapazorits reporting that atm ...15:15
Reapazor*sigh*15:15
Reapazorstill working on it :)15:15
Reapazorlol15:15
dhx_mReapazor: check the output of the md5 function at authentication_api.php:40015:15
dhx_mReapazor: and compare it to the output of utf8_substr( $t_processed_password, 0, PASSLEN )15:16
dhx_mif you can15:16
Reapazorhmm interesting ...15:16
Reapazorya it is definately the auth part in there15:16
Reapazori just was able to get it to return 'stuff' without an errror by avoiding that function15:16
Reapazorso ... back in i go :)15:16
dhx_m:)15:16
Reapazor:(15:23
Reapazorso it definately is the password len15:24
Reapazor:)15:24
Reapazormy password is > 32 chars15:25
Reapazor:)15:25
Reapazorwhich its getting cut down too15:25
dhx_maren't you using md5?15:25
Reapazori m15:25
dhx_myou can still have passwords > 32 chars15:26
dhx_mas long as you use crypt or md515:26
Reapazorya ... but thats all i got for possible causes15:26
Reapazorim dead ended right now15:26
Reapazorlol15:26
Reapazorsomething with that ;/15:26
Reapazorim gonna 'shorten'15:26
Reapazorthe password15:26
Reapazorjust to see15:26
dhx_mwhat line did it end up failing at?15:26
Reapazordidnt fail15:27
Reapazorpreformed as it should15:27
Reapazorfrom what i can tell15:27
dhx_mthe check after auth_process_plain_password?15:27
dhx_mhmm15:27
Reapazorya it wasnt coming back authed15:27
dhx_mI'd expect it work OK from what I can see15:27
dhx_m(the password checking)15:27
Reapazorya ;/15:29
Reapazorshorter password didnt work15:29
Reapazor*sigh*15:29
Reapazori checked the two functions as well15:29
Reapazorthe utf8 matches the md515:29
Reapazorwait what ...15:30
Reapazorhold on a second here15:30
Reapazorthis doesnt makesense15:30
nuclear_eclipseis soap api perhaps expecting your app to pre-md5 the password?15:30
Reapazorthats what im thinking :)15:30
Reapazorcause hte password is plaintext15:31
Reapazorand line 35015:31
Reapazoris comparing the hash == plaintext15:31
Reapazor(authentification_api.php15:31
Reapazorlooks like we've found the problem :)15:31
Reapazorso ..the question is ...should the soap apps be hashing already15:32
Reapazor ?15:32
Reapazorid wager against that15:32
Reapazorbut thats not a problem15:32
Reapazorif it has to be :)15:32
Reapazor(im testing the theory right now)15:32
nuclear_eclipseI personally think the soap client should pre-hash the password for security reasons15:33
nuclear_eclipsebut then again, if you're just sending the hash, it's not much different :P15:33
nuclear_eclipseand when we get to the point of having pluggable authentication methods, I'm not sure how that should or should not affect soap api, etc15:34
dhx_mnope, no security to be had unless you're using challenge-response15:34
kenguestnuclear_eclipse: even better: use https rather than http15:34
dhx_myep15:34
nuclear_eclipsewell, right15:34
dhx_mit should ideally use https with client certs15:34
nuclear_eclipsessl is always the answer for web security :P15:34
kenguestno.15:35
kenguestbut it helps15:35
dhx_mI guess just server https is also OK15:35
kleinweby(because you're currently talking about soap): Is there a way to "searchs" with the maintis soap api like "All assigned to me"?15:35
kleinwebymantis, sorry15:36
dhx_mnot too sure15:37
nuclear_eclipseironically, I think the only devs that actually use the soap api aren't here :P15:37
dhx_mlol15:37
kleinwebyoO15:38
Reapazorbad news15:39
Reapazorstill no dice with hash being sent15:39
Reapazorsigh ;/15:39
Reapazorill keep going ;/15:39
Reapazornow with the hash thing i gotta start at the top15:41
Reapazor:) good thing im not expected to do anything today lol15:41
dhx_mlol15:41
Reapazor*sigh* still no luck16:03
Reapazori just cant win16:03
ReapazorLOL16:03
Reapazor# Do not support migration to PLAIN, since this would be a crazy thing to do.16:19
Reapazor# Also if we do, then a user will be able to login by providing the MD5 value16:19
Reapazor# that is copied from the database.  See #8467 for more details.16:19
Reapazorif( $t_configured_login_method != PLAIN && $t_login_method == PLAIN ) {16:19
Reapazorreturn "should be true";16:19
Reapazorcontinue;16:19
Reapazor}16:19
Reapazor(yes thats my return)16:19
Reapazorthats my problem16:19
Reapazorauth_does_password_match16:19
Reapazorso you 'want' md516:19
Reapazorbut then when it gets there16:19
Reapazorit doesnt know whats going on16:19
Reapazorthere my problem...16:20
Reapazorbut now how to resolve that16:20
Reapazorlol16:20
Reapazoroh and the user_get_access_level also doesnt respect global access levels it seems ;)16:30
paul_nuclear_eclipse: oh john16:34
paul_that reminds me16:34
paul_did we talk about mssql? :)16:34
* paul_ forgets16:34
paul_Reapazor: I dislike the current soap api a little16:36
Reapazormeh .. i still havent got it workiing ;/16:37
Reapazori can see one global issue ...16:37
Reapazorim passing user/password to login...16:37
Reapazorbut post that its treating me as anoynmous16:37
Reapazorala when it gets access levels16:37
Reapazorits pulling for my anonymous user16:37
paul_nuclear_eclipse: can we branch?16:41
ReapazorMon Jun 22 12:42:03 2009] [error] [client 70.32.66.194] PHP Fatal error:  Cannot use object of type BugData as array in /var/www/vhosts/dotbunny.com/subdomains/bugs/httpdocs/api/soap/mc_project_api.php on line 3116:42
Reapazoryay! a core issue !16:43
Reapazorno more auth issues back to the bugdata16:43
paul_cool16:43
paul_what was it?16:43
Reapazorб═б═б═ б═б═б═ б═б═б═if( $t_configured_login_method != PLAIN && $t_login_method == PLAIN ) {16:44
Reapazorб═б═б═ б═б═б═ б═б═б═ б═б═б═continue;16:44
Reapazorб═б═б═ б═б═б═ б═б═б═}16:44
Reapazorsec getting line number16:44
Reapazorits just the auth part ...16:44
Reapazorthe bigger issue is my logged in user isnt having his access level rightly pulled16:44
Reapazor;/16:44
Reapazorbut in the end boils down to that bugdata object error16:45
Reapazorafter the auth stuff ...16:46
nuclear_eclipsepaul_: we haven't talked about mssql afiak, but I'd like to wait for you to look at the soap stuff before we branch16:53
dhx_mnuclear_eclipse: is RELEASE meant to list as many "main" features as possible?17:20
dhx_mguess not :)17:21
dhx_mthe changelog would be far too large now that I look at changelog_page.php17:22
paul_nuclear_eclipse: branch then port!17:31
nuclear_eclipsepaul_: really, have patience17:33
nuclear_eclipsewhat's so important that you can't keep it in a private branch for a couple more days at the most?17:33
paul_we've done that before :P17:33
nuclear_eclipseand if you really want 1.2 branched so bad, fix the soap api stuff from bug 10612 and so forth, and then I'll branch17:33
mantisbotBug 10612 - rombert - open - assigned17:33
mantisbotmc_project_get_issue_headers returns an empty response (this bugtracker) - http://www.mantisbt.org/bugs/view.php?id=1061217:33
paul_you know I need to do more then that17:34
paul_iirc, there's 2717:34
paul_well 28 now17:34
paul_soap api bugs to review17:34
paul_of which 4 contain patchsets17:34
* paul_ sets to work17:34
* nuclear_eclipse didn't realize soap api was that broken...17:35
paul_no but there's lots of stuff missing17:36
paul_also need to fix the utf8 bugs that seem to be coming up now17:36
dhx_myep I'd like some more testing before RC17:38
paul_huh17:38
paul_testing?17:38
dhx_mRC is meant to be "this should be stable"17:38
dhx_mthe bugs we're finding aren't just minor things17:38
dhx_malthough17:38
dhx_mI think branching is good for people that want to start working on new features again :)17:38
nuclear_eclipseRC imo means "representative of a final release"17:39
dhx_m... but I imagine we're all focussing on fixing at the moment17:39
dhx_mthat too17:39
nuclear_eclipseit doesn't need to be completely bug free, but it should be relatively ready for primetime use17:39
dhx_mbut users are going to hate us if we release an RC where the most obvious things break17:39
dhx_myep17:39
nuclear_eclipseat the same time, I'd rather have an RC release sooner where a bunch of users can submit bug reports than let it sit in Git forever while we don't get it out into the feild enough to even find those issues17:40
dhx_myep17:42
dhx_mpaul_: do you have a link to the test site you showed me ages ago?17:42
dhx_mpaul_: I really like that idea :)17:42
paul_hmm?17:42
dhx_mpaul_: the unit test stuff17:43
paul_you mean http://www.mantisforge.org/cruisecontrol/buildresults/mantisbt?tab=documentation17:43
paul_or well17:43
paul_http://www.mantisforge.org/cruisecontrol/buildresults/mantisbt?tab=testResults17:43
paul_for some reason17:43
paul_http://www.mantisforge.org/cruisecontrol/buildresults/mantisbt?tab=phpcs isn't showing output17:43
dhx_myep17:43
* paul_ needs to fix codesniffer output :(17:45
dhx_mnow I just need to learn how to add tests17:45
paul_make the config api test thing into a test17:45
paul_if you want something to do :)17:45
dhx_mdo the tests run on a live database?17:45
dhx_mie. it actually inserts and deletes from the database?17:45
paul_they would do17:46
paul_atm17:46
dhx_mok :)17:47
dhx_moh btw, did anyone see the deprecation notices in database_api.php?17:52
dhx_mcan we either remove the version if not fixing in 1.2 - or remove the deprecated stuff?17:52
paul_I swwear I already fixed this file17:54
nuclear_eclipsepaul_: have you seen bug 9798?18:01
mantisbotBug 9798 - bkraul - open - feedback18:01
mantisbotUnable to perform Mantis upgrade from 1.1.2 to 1.2.0a2 MSSQL(SQL 2005 Express) - http://www.mantisbt.org/bugs/view.php?id=979818:01
paul_Reapazor: stiill here?18:02
Reapazoryes18:04
Reapazorwhats shaking?18:04
paul_try pulling trunk and testing your thingy18:04
Reapazorokie give me a min18:05
Reapazorfixed the auth part ?18:05
paul_no the filter part18:05
CIA-18Mantisbt: paul * r76dbbcfb8307 /api/soap/mc_filter_api.php: #0010612: mc_project_get_issue_headers returns an empty response (this bugtracker)18:05
Reapazorok lemme see what i can do :)18:06
paul_nuclear_eclipse: well, I think we need to update install.php to cover what we actually support18:07
* paul_ uses mssql with odbc18:08
nuclear_eclipsewhat does install.php have to do with bad SQL syntax or what not during the schema install/upgrade?18:08
paul_there's multiple mssql drivers18:08
paul_i.e.18:09
paul_you can use18:09
paul_mssql18:09
paul_ado_mssql18:09
paul_odbc_mssql18:09
paul_pdo_mssql18:09
* paul_ thinks18:09
paul_the MS native mssql driver18:09
Reapazorhttp://pastebin.com/m4f6c2ef418:09
nuclear_eclipseoh, so you think that bug is a driver problem, not a syntax problem?18:09
Reapazornow i get someerrors back :)18:09
paul_install.php uses 'mssql'18:09
paul_I tend to use odbc_mssql at work18:09
paul_(as iirc, mssql truncates strings at the old sybase 8000 char limit18:09
paul_and MS have now done there new driver18:10
paul_which is version 1.018:10
paul_but won't be useful for us until version 1.1 gets released18:10
paul_On top of that mess18:10
paul_sql server stores utf8 data as ucs-2? (or similar typo)18:10
paul_in 1.1 *apparently* they going to support some sort of translation to/from utf818:11
paul_I'm not quite sure atm which of ado/odbc etc support utf8...18:11
nuclear_eclipseso, basically MSSQL is crap :P18:12
paul_on top of that18:13
paul_I made the mistake of asking about lowercase'ing some specific utf8 chars18:13
paul_in php18:13
dhx_mnuclear_eclipse: yeah... UCS-2 doesn't support surrogates and hence doesn't support the full range of valid Unicode characters :o18:14
nuclear_eclipsewhatever the hell that all means :P18:14
paul_It seems that mbstring doesn't use the same utf8 libs that the php6 uses18:14
paul_(or php5.3)18:14
paul_with it's intl extension18:14
paul_dhx_m: whats a surrogate?18:14
paul_isn't that a type of pregnancy?18:15
dhx_mpaul_: it is a second 16bit value that follows an initial 16bit value... this allows more characters to be represented18:16
dhx_mpaul_: basically with UCS-2 they decided "oh wait, 65536 characters is not enough to fit everything"18:16
nuclear_eclipsefacepalm18:16
dhx_mpaul_: and then they came up with UTF-16 which is quite similar to how UTF-8 works... except the base unit is a 16bit character (not 8bit)18:17
nuclear_eclipsedamn all those people and their pictograms and such :P18:17
dhx_m:)18:17
dhx_mUTF8 is great because you can just keep extending it as much as you like18:17
paul_anyway18:17
nuclear_eclipseif everyone would just learn ascii and english, programs would be so much easier to write!18:17
Reapazorpaul_:  http://pastebin.com/m4f6c2ef4 ... ;/18:17
dhx_mwell maybe not as much as you like... but anyway :)18:17
dhx_mhaha18:17
paul_i've taking unicode data18:17
paul_from unicode.org18:18
paul_run it over mbstring/compatlibs and so forth18:18
dhx_mjust grab it from zh.wikipedia.org and similar :p18:18
paul_and found that all the libraries I have (include phputf8 amongst others)18:18
paul_vary to mbstring by a bunch of characters18:18
paul_and i'm now trying to work out if mbstring or the libs are wrong18:19
dhx_mdid you set the locale correctly?18:19
paul_iirc18:19
paul_according to mbstring18:19
dhx_mthose utf8 functions have different outputs depending on the locale18:19
paul_LATIN UPPERCAZE Z IN A CIRCLE18:19
paul_stays as LATIN UPPERCAZE Z IN A CIRCLE when lowercase'd iirc18:19
dhx_mfor instance, "a" might be lowercase in English but it might be uppercase in WeirdLanguage18:20
paul_the unicode data suggests it should become latin lowercase z in a circle18:20
dhx_mpaul_: a lot of these libraries have a very limited lookup table for each language as well18:20
paul_http://www.fileformat.info/info/unicode/char/24cf/index.htm18:21
paul_http://unicode.org/cldr/utility/character.jsp?a=24C918:23
paul_see what mb_strtolower() of char 24C9 returns for you18:23
paul_Reapazor: mc_api.php18:25
paul_line 19618:25
paul_Reapazor: try removing the & from function mci_null_if_empty( &$p_value ) {18:25
Reapazorsec :)18:26
Reapazorholy crap 1/3 tests passed!!18:27
paul_1/3?18:27
Reapazorhehe ya :) i am testing different api function calls18:27
Reapazorthe other ones are coming back with another .... message18:28
Reapazorsomething about content type being returned18:28
Reapazornot expected18:28
paul_function mci_get_severity_id( &$p_severity ) {18:28
paul_i'm tempted to remove all the &'s18:28
* paul_ pokes nuclear_eclipse 18:28
dhx_mpassing by reference is more efficient though?18:28
nuclear_eclipsehi paul_18:29
Reapazorim getting some Not supported Content-Type in the response: 'text/html'18:29
Reapazorerrors ;/18:29
Reapazoras well as System.Web.Services.Protocols.SoapException: Invalid Filter18:30
Reapazornow :)18:30
nuclear_eclipsetimezone chooser breaks on mantisbt.org because the timezone_identifiers_list() function is not in 5.1 as a core function18:30
paul_(PHP 5 >= 5.1.0)18:30
paul_Installation18:30
paul_There is no installation needed to use these functions; they are part of the PHP core.18:30
nuclear_eclipsenot to mention I want to stab the list that you're having it use anyways because it's a) horribly verbose, and b) not specific18:31
nuclear_eclipseread the first comment: "Please note that the timezone_identifiers_list function is not available in the most recent versions of PHP available for CentOS/RHEL as of this writing ( 5.1.6-23.2.el5_3 )."18:31
Reapazor^ that could be me ?18:31
paul_...18:31
paul_nuclear_eclipse: why not? ;/18:31
nuclear_eclipse"I think the function is labeled as >= 5.1.0 in this documentation because it's possible to get the Date::Time class installed in 5.1 if you're compiling from scratch (or it's a documentation error).18:32
nuclear_eclipse"I just spent a bunch of time trying to get it to work in 5.1 (installing timezonedb manually because the default PECL runs out of memory, etc.), but have decided to use a manually generated list instead. The actual timezone support seems to work fine in 5.1."18:32
dhx_mpaul_: you'll soon learn that the sole aim of RHEL is to frustrate sysadmins :p18:32
paul_nuclear_eclipse: right, let me find out18:32
paul_i'll ask the php-dev who updates the timezone db18:33
paul_whether RH fail18:33
nuclear_eclipsepaul_: I've confirmed via command line on the server that the funciton does not exist in our installation18:33
nuclear_eclipseanyways, I think that list is disingenuous anyways for the reasons I stated above -- we should present a much smaller list to the user, including +/- from GMT18:34
paul_I'd agree there18:34
paul_but I also think we should try to use the php functions to do the work for us18:34
paul_(They have a more accurate timezone db)18:34
nuclear_eclipse+1, but it doesn't seem to have anything like that18:35
paul_whilst we need a simple list18:35
paul_you can get the timezone stuff as an object and do whatever we want to18:35
paul_dinner brb18:35
dhx_musing +/- hours from GMT disregards daylight savings time though18:39
dhx_mwhich is wrong18:39
nuclear_eclipsegah, let's please not bring DST into this mess just yet :P18:40
dhx_m:p18:41
paul_well we first need to work out why RHEL doesn't have the function18:49
paul_as i'd rather use the php timezone stuff18:49
paul_then roll our own database18:49
paul_(especially given our names need to match theres anyway)18:49
Reapazorpaul_ ...seems the only problem i have left is mc_project_api.php:line 5418:50
ReapazorPHP Fatal error:  Cannot use object of type BugData as array in /var/www/vhosts/dotbunny.com/subdomains/bugs/httpdocs/api/soap/mc_project_api.php on line 5418:50
dhx_mI guess a fallback to using +/- hrs from GMT would be acceptable18:50
dhx_mbut too much work IMO18:50
paul_Reapazor: did you remove & to all ?18:51
Reapazornot in that file yet18:51
Reapazorin the other :)18:51
Reapazorhmm looks like i might have missed one18:52
Reapazor*sigh* back i go18:52
Reapazorah different file :)18:52
Reapazormc_enum_api18:52
Reapazoras i go thru you seem to be right on18:56
Reapazorall the &'s are breaking it18:56
paul_nuclear_eclipse: *sigh* ok having googled18:57
paul_there's 2 things here18:57
paul_a) it seems the datetime stuff was experimental in php 5.118:57
paul_and default from 5.2.018:58
paul_seperately to that18:58
Reapazorline 54 of mc_project_api  .... was $t_issue_data['handler_id'] needs to be $t_issue_data->handler_id18:58
Reapazorthat puts me up to 3 of 4 tests passed !18:58
paul_there's some whinging about php distributing timezones (to fix issues with distributions updating and breaking the timezonedb on applications behind the scene)18:58
Reapazorln 30 in mc_project_api should read: $t_id = $t_issue_data->id;18:59
Reapazorassuming someone did that, or thats left from my fix18:59
paul_my original plan with that drop down (Timezones)18:59
paul_was to improve it later on to put common ones at top of list or have a short list19:00
paul_then have a long list listing everything to pick up the 'random timezones' people might want to use19:00
nuclear_eclipseI think just having a short list would be the best option19:00
paul_I was just trying to avoid the case of people 'bug-reporting'19:01
paul_Please add Europe/Athens to timezone19:01
paul_as on one hand, the complete list that people need to use is at http://us.php.net/manual/en/timezones.php19:01
paul_if a Server is set to19:01
paul_Europe/Athens19:01
paul_but that isn't on our pre-defined dropdown list19:02
paul_we'd have to add it or something19:02
nuclear_eclipseif we don't really categorize timezones by their "city", then we shouldn't get those sorts of bug reports, no?19:02
Reapazorso where might i go to tell / write down the changes ?19:02
Reapazoras 4/5 test work19:02
paul_*maybe*19:02
Reapazor(i keep adding a few)19:03
Reapazor5'th being a filter error19:03
nuclear_eclipseReapazor: post them to a bug report on mantisbt.org please19:03
* paul_ slaps nuclear_eclipse 19:03
paul_nuclear_eclipse: the point is19:03
paul_a) we need to use a valid php timezone19:03
paul_b) users could use any valid phptimezone in their php.ini19:03
paul_so we'd effectively be saying19:03
nuclear_eclipsewe can't just map all the possible +/- settings to a single timezone?19:04
nuclear_eclipseit's not like London will all of a sudden move over +3 hours...19:04
paul_I was wondering about summer time in places but not others or similar19:04
nuclear_eclipsescrew DST?19:05
nuclear_eclipse:P19:05
paul_the db stores timezone offsets http://us.php.net/manual/en/datetimezone.getoffset.php19:05
paul_the only thing we'd need to support is:19:06
paul_if user sets timezone in php.ini as 'Australia/Lord_Howe'19:06
paul_that would be default for people that haven't set a timezone (aka server timezone)19:06
paul_so whilst I knew it was a long list19:06
paul_i'd be kinda hoping we could 'tidy' it as time goes by ;/19:07
paul_In this case it doesn't help that19:07
paul_php manual is wrong19:07
paul_my windows boxes run php5.3 and my linux boes run 5.219:07
paul_(linux box runs stock ubuntu stuff)19:07
paul_looks like wordpress have the same issue http://wordpress.org/support/topic/27959319:09
paul_albeit it looks like the solution they have is not to allow users to change timezone unless they are on 5.219:09
nuclear_eclipsehmm19:10
dhx_mpreg_replace() [function.preg-replace]: Compilation failed: PCRE does not support \L, \l, \N, \P, \p, \U, \u, or \X at offset 9819:10
paul_the http://gophp5.org/ date was like Feb 200819:10
paul_and that was saying 'go php5.2 as 5.0 was crap and 5.1 was slow'19:11
nuclear_eclipsedhx_m: did you use the /u modifier+19:11
dhx_mstring_api.php:47319:11
dhx_mnuclear_eclipse: we already are19:11
paul_what are you changing ?19:11
mantisbotNew bug: Bug 10615 - reapazor - open - new19:11
mantisbotNew bug: API Failure - http://www.mantisbt.org/bugs/view.php?id=1061519:11
nuclear_eclipsepaul, look at 10615 and commit the fixes19:12
dhx_mnuclear_eclipse: mind you, this is a warning... I have all values for $g_display_errors set to 'halt'19:12
paul_dhx_m: we shouldn't be getting any warnings19:13
paul_nuclear_eclipse: already have19:13
dhx_mnuclear_eclipse: using CentOS 4 with custom upgraded packages19:13
paul_or well19:13
paul_i've fixed a ocuple more19:13
CIA-18Mantisbt: paul * r996615c7f346 /api/soap/ (mc_api.php mc_project_api.php): Fix additional soap api issues19:13
paul_Reapazor1: did i miss anything?19:14
dhx_mpaul_: the great folks at RHEL/CentOS must have decided no one needs Unicode support in PCRE19:14
paul_hmm?19:14
paul_isn't it /u?19:14
dhx_myep19:14
nuclear_eclipsedhx_m: upgrade to at *least* CentOS 5 :P19:15
dhx_mnuclear_eclipse: I can't :(19:15
paul_?19:15
dhx_mthis isn't my server19:15
nuclear_eclipsebad dhx_m19:15
nuclear_eclipseno biscuit!19:15
paul_dhx_m: are you on php5.1?19:16
paul_on that box?19:16
Reapazor1paul_ ill do a trunk pull19:16
Reapazor1and see how she works19:16
paul_mantis is male19:16
Reapazor1lol fine ...19:16
dhx_mpaul_: yes, 5.1.619:16
Reapazor1ill see how he works19:16
paul_dhx_m: what pcre version?19:16
paul_in phpinfo?19:16
paul_dhx_m: let me guess19:17
paul_4.519:17
dhx_mpaul_: 4.5 01-December-200319:17
dhx_myep19:17
paul_wget php5.3 source19:17
paul_./configure19:18
paul_make19:18
paul_make install19:18
paul_and move on with life19:18
dhx_mpaul_: I wish I could... it's a shared server I have no control over19:18
Reapazor1paul_ seems to be working :)19:18
Reapazor1now there is one 'problem' that i dont know if i really care to address19:18
paul_Reapazor1: I guess your not on php5.1 then19:18
paul_:P19:19
Reapazor1LOL19:19
Reapazor1the filters19:19
Reapazor1they changed in 1.2 ?19:19
paul_unknown19:19
paul_in fact probably19:19
paul_lots cfhanges19:19
Reapazor1" Invalid Filter "19:19
Reapazor1when i do a soap search by filter19:19
dhx_mpaul_: A non well formed numeric value encountered19:20
dhx_mpaul_: custom_function_api.php:33319:20
paul_what's php5.1 fail on now?19:20
dhx_mpaul_: arguments to date(): ( <string>'Y-m-d', <string>'2009-06-19 23:45:43' )19:21
Reapazor1paul_ : ya 5.2.6 atm :)19:21
paul_$t_function( $p_bug, $p_columns_target );19:21
paul_is custom function api 33319:21
dhx_myep19:22
paul_thats not a date....19:22
dhx_msorry19:22
dhx_mwrong line of the trace19:22
dhx_mcolumns_api.php:107919:23
dhx_mit's grabbing $p_bug->last_updated19:23
dhx_mand it seems to be a string?19:23
paul_what had you done to generate that ;/19:25
dhx_mpaul_: doh never mind me...19:25
paul_?19:26
dhx_mpaul_: hadn't fully upgrade the database yet *headslam*19:26
paul_although is tempting to add if(!is_numeric( to __Set for bug19:27
dhx_myep19:27
dhx_mI was hoping that in the future all checks are placed in __set/__get19:28
dhx_mie. who can update fields, etc19:28
CIA-18Mantisbt: jreese * r6ee263adfacc / (account_prefs_update.php core/print_api.php): Fix timezone list on some PHP 5.1 installations19:29
nuclear_eclipse^ temporary measures19:30
dhx_mnuclear_eclipse: can we check for PCRE Unicode support?19:34
* nuclear_eclipse stabs dhx_m 19:34
dhx_mnuclear_eclipse: it'd be easy :p19:34
nuclear_eclipsetalk to paul_, he's the overmind behind admin/checks.php19:35
dhx_mnah I meant inline19:35
dhx_mso...19:35
* nuclear_eclipse stabs dhx_m 19:35
dhx_m$t_valid_unicode_chars = is_unicode_supported() ? '\pL' : '';19:36
dhx_msomething like that19:36
dhx_mnuclear_eclipse: actually I'm of the opinion now that old crap like CentOS 4 should be providing their own patches to revert Unicode/etc support19:40
dhx_mat least for small things like I'm experiencing now19:40
nuclear_eclipseI'm of the opinion that anyone still running CentOS 4 should be taken out back and beaten or shot19:41
nuclear_eclipse:P19:41
paul_dhx_m: i'm thinking about your /u thing19:41
dhx_mnuclear_eclipse: I hate it too!19:41
paul_so19:47
paul_remind me19:47
paul_on centos 419:47
paul_utf8 fails?19:47
paul_so we're telling people to update? :P19:47
dhx_myeah19:48
dhx_mRHEL 5 is ok then?19:48
dhx_mthe fix seems really simple at the moment19:48
nuclear_eclipsedhx_m: it seems so, mantisbt.org runs it19:48
paul_RHEL seems to be as bad :P19:48
dhx_mbut as we use more and more unicode regex queries...19:48
nuclear_eclipseor rather, live site runs CentOS 519:49
dhx_myep, same thing I guess19:50
dhx_mif we're going to release this19:51
dhx_mactually never mind ;)19:52
dhx_mwas going to say the CentOS/RHEL fix is easy19:52
nuclear_eclipsepaul_: any other immediate fixes to soap api available/needed?19:52
dhx_mbut it isn't really... you need a custom PHP/Apache/1000 other things too19:52
dhx_mnuclear_eclipse: what is the reason for http://git.mantisforge.org/w/source-integration.git?a=commitdiff;h=67994d425f8a81ec6a0767a3f4de7b930225bb8119:55
dhx_mall throughout19:56
nuclear_eclipsebecause if/when the API function gets called from a different plugin, the plugin API will look in the wrong place for the configuration option due to the way the plugin stack works19:56
dhx_myep... but what calls the Source plugin?19:56
nuclear_eclipsea) other Source-based plugins, b) the SourceIntegration sub-plugin19:57
dhx_msounds like something we should fix with plugin_config_get so that it takes an optional "plugin" argument?19:57
dhx_mthen again... may as well use the raw thing19:58
nuclear_eclipseit would be preferred, but there are already multiple, optional parameters already, that you'd have to call plugin_config_get( 'value', null, null, null, 'Source' );19:58
dhx_mI see your point now though :)19:58
dhx_mah ok19:59
paul_re-order params :P19:59
nuclear_eclipseat which point, I figured it easier to make a call to the core API :P19:59
nuclear_eclipsere-order params would break anything tha'ts been relying on the API up to this point...19:59
nuclear_eclipseif PHP had named params like Python, I would have added it in a heartbeat20:00
dhx_myep20:00
nuclear_eclipseeg, plugin_config_get( 'value', plugin='Source' ) type of thing20:00
nuclear_eclipseI <3 Python20:00
dhx_myeah that is what I was thinking too (but in terms of Ada :p)20:00
dhx_mUndefined property: SourceChangeset::$comitter20:04
dhx_mSource/Source.ViewAPI.php:12020:05
nuclear_eclipsehah, gotta love typos... :P20:06
nuclear_eclipsethat one's *really* bad.... :(20:06
dhx_mlol20:07
nuclear_eclipsefix pushed20:07
dhx_mthanks :)20:08
nuclear_eclipsenp20:10
nuclear_eclipse*that* is why I haven't yet made a real release of source-integration yet :P20:10
dhx_m:)20:11
dhx_mfixed thanks20:12
dhx_mpaul_: with the new graphs plugin... I can't specify the path to jpgraph anymore20:13
paul_you'd put jpgraph in libs dir20:13
dhx_mah ok, it's forced now?20:13
dhx_mI don't mind that :)20:13
dhx_mthe other ezc thing was complaining about missing fonts and I couldn't be bothered debugging :p20:15
nuclear_eclipsepaul_: any other immediate fixes to soap api available/needed?  anything to add to the release notes?20:16
paul_what release notes?20:16
nuclear_eclipsedoc/RELEASE20:16
dhx_mminimum system requirements need updating20:19
CIA-18Mantisbt: jreese * rd11b55e8c96d /doc/INSTALL: Updated documented minimum PHP version.20:21
dhx_mis MySQL really a requirement?20:23
nuclear_eclipseyes :P20:23
paul_no :P20:23
dhx_mhah20:23
nuclear_eclipsejust look at the crap paul's going through for pgsql and mssql support :P20:24
paul_since when did we have a docs/RELEASE20:24
dhx_moh pgsql doesn't work :o20:24
nuclear_eclipsesince this morning20:24
dhx_mI don't use it... but it'd be my choice for database if I was installing one anytime soon20:24
* nuclear_eclipse notes that paul_ should really pay attention to git update emails...20:24
paul_well, hmm20:27
dhx_mmanage_user_create.php:82 $t_lang is undefined20:30
paul_mm?20:31
nuclear_eclipseneeds to be " = lang_get_current()"20:32
dhx_mhow did that happen? :o20:33
dhx_mthat fixes is20:34
dhx_m*it20:34
CIA-18Mantisbt: jreese * r6a862b67e037 /manage_user_create.php: Fix warnings about $t_lang and such.20:34
nuclear_eclipse^ even better fix20:35
dhx_myep20:36
nuclear_eclipseanywho, time to head home -- will be back on later20:36
nuclear_eclipsecheers20:36
dhx_mok cya :)20:36
paul_we miss some big chunks of work in those release notes20:39
dhx_mI'm more excited about what to do in 1.3 :p20:46
paul_I think we should pick a date for 1.320:48
paul_for the release20:48
dhx_mquarterly releases would be nice20:52
dhx_mmaybe a bit fast though?20:52
dhx_mI guess we have branching now anyway20:52
paul_well20:52
paul_if we RC now20:53
paul_historically that would mean we'd release in ~40 days ;p20:53
paul_(or do an rc2)20:53
paul_1.0 has 5 rc's20:53
paul_1.1 had 320:53
dhx_mlol20:53
paul_each rc takes about 40 days20:53
paul_once we hit a release20:54
paul_i.e a .X20:54
paul_we seem to do a point release every 60 days20:54
paul_i.e.20:54
paul_1st August would be 1.2 final20:55
paul_30th september would be 1.2.120:55
paul_now in terms of 1.320:55
dhx_m6 months for major releases?20:56
dhx_mand what about monthly updates?20:56
dhx_mor whenever they're needed (in terms of security fixes)20:56
paul_1.3a120:56
dhx_mI see no point delaying them20:56
paul_we normally see 218 days after the first 1.2 release20:56
paul_although we've speed up recently between 1.1.0 and 1.2a1 it was only 121 days20:57
paul_so any time betweeen 20th October and Feb 2010 should be 1.3a120:58
dhx_mdo we really need alpha releases?20:58
paul_sure20:59
paul_if we release an alpha on 20th October 200920:59
dhx_mIMO it'd be better to get early adopters to use git checkouts20:59
dhx_m*snapshots20:59
paul_after the alpha on 20th October 200920:59
dhx_mit makes it easier to say "fixed, please update"20:59
paul_we'd be for a final 1.3.0 during christmas vacation in January 201121:00
paul_it could be quite nice to do a december 2010 1.3 release of mantis :)21:00
paul_with snow and stuff21:00
dhx_mhaha21:00
dhx_mwell graphics are an area I want to improve in 1.321:01
dhx_mwas thinking of Tango icons but they don't provide everything we need21:01
paul_for the Xmas 2010 release, we'd need to support php 5.2+ if we allow 3 years of php versions21:02
dhx_myep21:02
paul_so we're set :)21:02
paul_so now we've got a release schedule for 1.321:02
dhx_mwhat are the aims?21:02
paul_need to decide contents21:03
dhx_mI recon a better automated test setup would be great21:03
paul_so lets see...21:03
paul_1.3 alpha: 20th October 200921:03
dhx_mit can't take that long to write tests :)21:03
paul_1.3.0: Dec/Jan 2010/1121:03
paul_min php version 5.221:03
dhx_mbetter database and Unicode support21:03
paul_contents21:03
paul_a) tests21:03
dhx_mnew graphics21:03
paul_b) graphics21:03
paul_c) auth plugins21:04
dhx_m:)21:04
dhx_mnotification plugins?21:04
dhx_mor is that pushing it hah21:04
dhx_mactually documentation needs a lot of work too21:05
paul_anyway21:06
paul_I think the point is21:06
paul_we need to decrease timescale21:06
dhx_myep21:06
paul_dhx_m: in terms of db_query and stuff you was asking about earlier21:07
paul_I need to work out about binary files21:08
paul_I think uploaded files wouldn't go through db_query_bound on one db21:08
dhx_mpaul_: ok21:08
dhx_myeah I'm a little worried about PHP script execution time too21:08
paul_hmm?21:09
dhx_mif someone is on a slow connection uploading a file21:09
sshirleyHi all. In v1.1.0 can anyone tell me off the tops of their heads in which configuration file the "editing" threshold is maintained. By editing threshold, I mean under "resolved" issues, I don't have edit icon. I need to be able to open resolved issues.21:19
nuclear_eclipsesearch for read_only21:20
paul_nuclear_eclipse: you missed the next release planning :)21:25
nuclear_eclipseno, I saw it21:25
sshirleyUnfortunately that didn't help. I couldn't find that. I know I've asked you this before. But right now I just copied files over to the production server and that icon is now missing. Everything works fine on my development. I copied over 4 files (core/bug_api.php, core/constant_inc.php, config_defaults_inc.php, and custom_strings_inc.php). I have until 6pm. :-)21:27
paul_nuclear_eclipse: it's statistically accurate :P21:28
nuclear_eclipsesshirley: $g_bug_readonly_status_threshold = RESOLVED;21:33
sshirleyThanks. JUST found it in config_defaults_inc.php.21:34
nuclear_eclipsesshirley: make sure you set the new value in config_inc.php so it doesn't get overwritten during an update21:35
sshirleySo if it's in config_inc.php, then it overrides what's in config_defaults_inc.php?21:38
sshirleyIf I want everything to be "editable", I set it to the highest value? Meaning if 'Closed' in 90 and I set this variable, then 10,20,30,...,90, etc will be editiable?21:39
nuclear_eclipseyes, config_inc overrides config_defaults; that's its entire purpose :P21:39
sshirleyMakes sense :-)21:39
nuclear_eclipseand no, whatever is *at* the threshold will not be editable; anything below it will be21:40
paul_bug 852621:41
mantisbotBug 8526 - vboctor - open - assigned21:41
mantisbotversion_api should return date_order in unixtimestamp format - http://www.mantisbt.org/bugs/view.php?id=852621:41
sshirleyAgain in v1.1.0, where is the coloration in 'my views' set? Since I made my modifications, the green under Resolved is not showing (it's showing white)22:07
sshirleyI think I found it :-) It's an array in config_defaults_inc.php, right?22:08

Generated by irclog2html.py