Friday, 2009-06-12

../irclogs/#mantishelp.2009-06-12.log
--- scribe started ---00:00
siebrandhow can I get the current user's name? (the one executing the script)00:11
nuclear_eclipsesiebrand: user_get_name( auth_get_current_user_id() ) ?00:13
siebrandnuclear_eclipse: that's sounds precisely like what I am looking for :)00:14
nuclear_eclipsenote that user_get_name() will return the user's realname if that's enabled00:14
siebrandthat sounds fine.00:15
siebrandadministrator has created an account for you with username "t5". In order to complete your registration, visit the following URL (make sure it is entered as the single line) and set your own access password:00:16
siebrandnuclear_eclipse: are you sure it does that? Looks like it does not, actually.00:18
siebrandnuclear_eclipse: ow, that's probably a config setting I am not using.00:18
CIA-61Mantisbt: s.mazeland * r755a5de6ba57 / (4 files in 3 dirs): Fix: #0010578: Admin created user accounts should get different confirmation e-mail00:33
CIA-61Mantisbt: s.mazeland * r3b0e73e47849 /plugins/MantisCoreFormatting/lang/strings_russian.txt: Follow-up to master 51e4567d : remove untranslated message from strings_russian.txt00:33
nuclear_eclipsesiebrand: depends on $g_show_real_name iirc00:51
* siebrand nods.00:51
devnethey everyone...how do I allow public viewing of my mantis install without authentication?  I'd like to have everyone be able to browse and view but not report bugs02:45
mantisbotNew bug: Bug 10585 - Kirill - open - new05:32
mantisbotNew bug: string not found manage_import_issues_link - http://www.mantisbt.org/bugs/view.php?id=1058505:32
RonaldWhat about plugin priority? I just added a dependency on jQuery from git.mantisforge to mu plugin; and unless I put my plugin at higher priority; its EVENT_REPORT_BUG_FORM output is not there?07:57
[KK]Kirillsiebrand: lo08:07
dhx_mhey08:07
[KK]Kirilldhx_m: lo08:13
dhx_m[KK]Kirill: hi08:14
[KK]Kirilldhx_m: nice look mantisbt.org with national character support08:20
[KK]Kirillhttp://www.mantisbt.org/bugs/plugin.php?page=ManTweet/index.php08:20
dhx_myep that shows well on my screen :)08:20
dhx_mI'm getting a patch ready for bug 4772 at the moment08:21
mantisbotBug 4772 - hugopedersen - open - acknowledged08:21
mantisbotChanging username doesn't advice user - http://www.mantisbt.org/bugs/view.php?id=477208:21
dhx_mjust picking random old small fixes to do :)08:21
[KK]Kirillit's good08:22
[KK]Kirillupload patch08:22
[KK]KirillI include in my version08:22
Ronaldhi; did you guys see my Q just before you became active?08:31
[KK]Kirillyes08:32
[KK]Kirillby I don't  know08:32
[KK]Kirillbut*08:32
RonaldAh well thats a shame :)08:32
[KK]KirillRonald: try ask nuclear_eclipse after 5 hours08:33
RonaldWill do; doesn't currently hold me up08:33
siebrandmogge08:56
siebrandCan anyone tell me where the 'ManTweet' code is? It has some i18n issues. If I fix an issue, it will log in Dutch. That should be the tracker language.08:58
[KK]Kirillgit.mantisbt.org09:00
[KK]Kirillhttp://git.mantisbt.org/?p=mantweet.git;a=summary09:00
siebrandHmm, I guess the bug is already somewhere in core.09:04
[KK]Kirillsiebrand: I upload new file.09:04
[KK]Kirillabout core plugin09:05
siebrand[KK]Kirill: yep, saw that.09:05
[KK]KirillThanks09:06
siebrandRonald: what question?09:08
Ronaldwas just before you entered; will repeat09:08
RonaldWhat about plugin priority? I just added a dependency on jQuery from git.mantisforge to mu plugin; and unless I put my plugin at higher priority; its EVENT_REPORT_BUG_FORM output is not there?09:08
siebrandRonald: no idea.09:11
siebrandRonald: I assume priority determines when it is processed, and if you process it too late, it may not be doing what you expect it to do...09:11
Ronaldpossibly, but jquery doesn't use that event; and it only went weird when jquery was actually added to my plugins dependencies (with jquery still loaded)09:12
dhx_mok my patch for bug 4772 is complete and tested, will push it to mantisforge now :)09:41
mantisbotBug 4772 - hugopedersen - open - acknowledged09:41
mantisbotChanging username doesn't advice user - http://www.mantisbt.org/bugs/view.php?id=477209:41
gialluuhm. is it possible to change the username???09:50
siebrandgiallu: manage_user_edit_page.php allows this.10:19
siebrandgiallu: non-admins cannot do it.10:19
siebranddhx_m: nothing pushed? http://git.mantisforge.org/w?o=age10:21
dhx_mwill push it now sorry11:07
dhx_mjust working out how to convert access levels to strings11:08
dhx_mah get_enum_element11:09
dhx_mhmmm well this is a problem...11:17
dhx_m$g_enable_email_notifications in config_defaults.inc says that signup emails are turned off11:18
dhx_min the documentation, it says they aren't turned off11:18
dhx_mhttp://git.mantisforge.org/w/mantisbt/dhx.git?a=shortlog;h=refs/heads/manage-user-email-notify11:38
nuclear_eclipseRonald: a) is the jQuery plugin installed correctly in your Mantis install, and b) did you set an appropriate dependent version in your own plugin?12:36
Ronaldnuclear_eclipse, manage plugin screens shows jquery as green/correct.12:37
nuclear_eclipseand what about your plugin?  does it show that its dependencies are met?12:37
Ronaldyes12:37
Ronaldboth mantis core ans jquery as green12:37
nuclear_eclipseRonald: how many plugins do you have installed, and what are they?12:39
Ronaldjquery, mantis formatting (came default) and core12:39
Ronaldand my own obviously12:39
[KK]Kirillnuclear_eclipse: lo12:40
nuclear_eclipsecan I see the source for your plugin?  or at least the ->register(), ->hooks() and such setup?  mantis.pastebin.com12:41
nuclear_eclipsehi [KK]Kirill12:41
[KK]Kirillnuclear_eclipse: http://git.mantisforge.org/w/mantisbt/dhx.git?a=commitdiff_plain;h=d62c08e3b6828c6f4a227f97e664528d1759eb67;hp=44719eb9f8d7043cb90c816ed8e345ca0263473912:41
[KK]Kirillis it patch-file?12:41
Ronaldnuclear_eclipse, http://mantis.pastebin.com/m33ff336a12:42
nuclear_eclipse[KK]Kirill: what about that?12:43
[KK]Kirillit's dhx_m modification12:43
nuclear_eclipseright12:44
[KK]Kirillok12:44
dhx_myep12:45
nuclear_eclipseRonald: that all looks correct -- honestely I'm not sure what the priority would have to do with whether your plugin hook got executed...12:45
dhx_m[KK]Kirill: did you test it?12:45
nuclear_eclipseRonald: are you running the official a3 release, or something from git?12:45
Ronaldmantisbt-1.2.0a3-2009-06-10-master-d9978e512:46
Ronaldfrom the nightly dir12:47
nuclear_eclipsehmm12:47
Ronaldthe jquery is git from tiday12:49
RonaldJust verified; either need to raise my plugin to priority 2; or drop depend on jquery. (keeping jquery loaded in mantis; AND its code shows up in page source)12:54
Ronaldfor all 3 pages (report, update, view bug)12:56
nuclear_eclipseactually, priority 2 is a lower priority =\12:57
Ronaldalright ;) all plugins are at 3; cept my own. needs to be 2 or 1 to work12:59
Ronaldplugin order as listed on manage plugins: Mine, jquery, formatting, core13:00
nuclear_eclipseI'm still boggled as to how that could be happening -- I've got *lots* of plugins that depend on each other, and never had to change their priorities to make them work...13:02
RonaldI think i can give you the complete code; nothing in house yet in there13:02
nuclear_eclipseok13:02
Ronaldwhere do you want it?13:03
nuclear_eclipseon pastebin should be fine13:03
Ronaldmultiple files13:04
nuclear_eclipseah, right13:04
nuclear_eclipseumm, email?13:04
Ronaldsure13:04
nuclear_eclipsejreese@leetcode.net13:04
Ronaldusing lang and some small/light template engine(used to working with it)13:04
mantisbotNew bug: Bug 10586 - rickb - open - new13:04
mantisbotNew bug: Add burn-down charts to roadmap report - http://www.mantisbt.org/bugs/view.php?id=1058613:04
* nuclear_eclipse uninstalls on plugins on his test install....13:05
Ronaldsend 10K tar.bz2 your way13:06
nuclear_eclipseok13:06
dhx_m10kB? :o13:06
Ronaldit does include jotti.org/openlibs/jtpl ;)13:07
dhx_mah ok13:07
Ronaldhtml inline got messy real quick, so pulled in the template thing I normally use13:08
mantisbotNew bug: Bug 10587 - rickb - open - new13:09
mantisbotNew bug: UI improvements to Roadmap report - http://www.mantisbt.org/bugs/view.php?id=1058713:09
nuclear_eclipsewow, I can replicate the problem, but I have absolutely no clue why that's happening...13:15
RonaldThere is only one thing worse then a problem you have no clue about13:16
Ronalda problem you cannot replicate AND have no clue about ;)13:16
nuclear_eclipselol13:16
RonaldI have a few of those lingering, very hard to fix13:17
mantisbotNew bug: Bug 10588 - rickb - open - new13:19
mantisbotNew bug: Rename 'assigned' state as 'approved' or 'actioned' - http://www.mantisbt.org/bugs/view.php?id=1058813:19
paul_nuclear_eclipse: regex13:22
nuclear_eclipseI'm stumped; I'm gonna have to pull out the Eclipse PDT debugger and step through this thing, see what's going on;  my only thought is that somehow your plugin is never getting past the dependency-check phase of initialization...13:22
Ronaldtake the time; i'm not blocked cus the priority workaround works13:23
nuclear_eclipseyeah13:23
paul_Ronald: john's busy fixing regex13:25
paul_:P13:25
Ronaldask the author of the template thing I use, he dreams in regex13:26
* nuclear_eclipse smacks paul_ 13:26
nuclear_eclipseregexen are very wonderful :D13:26
Ronaldregex have character!13:27
Ronaldyou either get along..... or /not/13:28
paul_nuclear_eclipse: i'm fixing mantis for 5.313:28
paul_hence the regex query13:28
nuclear_eclipsepaul_: so are ereg* going away in 5.3?13:30
paul_deprecated13:30
paul_as is split13:30
paul_etc13:30
paul_I might clone your utf8 branch and work on13:32
nuclear_eclipsepaul_: I added work for mb_substr yesterday13:33
paul_I saw13:33
paul_but i'm still not sure about it13:34
paul_substr is a function i've think i've got a utf8 version (using regex) of13:34
paul_I partly wonder if we should add mb_* stuff as a performance add on once we're happy with our  compat-versions13:35
nuclear_eclipseand I profiled utf8_substr vs other options, and for 40,000 calls on an ever-growing string, it was 0.7s for utf8 versus 0.3 for regular, so the performance cost is negligible to mantis13:35
nuclear_eclipsepaul_: I'd rather add compat on top what I already have13:35
paul_you can probably see my logic though i.e.13:36
paul_if we drop strlen13:36
paul_use mb_Strlen if available13:36
paul_use custom_strlen if not available13:36
paul_and use strlen if phpv613:37
paul_we've got 2(3) different branches13:37
paul_with out custom_strlen most likely to be incorrect13:37
nuclear_eclipseright, but the good part is that those branches only need to be evaluated once, at the time of defining our custom utf8_* functions13:37
paul_yep13:38
nuclear_eclipseand utf8_strlen *will* be correct because it uses utf8_decode13:38
paul_you'd hope so at least13:38
nuclear_eclipsewell, I can't see why it wouldn't be...13:39
paul_although I did wonder last night if we should call utf8_strlen when adding a BLOB to database (e.g. a file upload)13:39
paul_btw, did you look at http://git.mantisforge.org/w/mantisbt/paul.git?a=shortlog;h=refs/heads/11June13:39
paul_or more13:40
paul_specifically:13:40
paul_http://git.mantisforge.org/w/mantisbt/paul.git?a=commitdiff;h=34e2ecbd4660fd559e285d758bf7efb46d00e34013:40
paul_http://git.mantisforge.org/w/mantisbt/paul.git?a=commitdiff;h=8189a1ceb01bcae72798c4a0ea15030d265eabe113:40
dhx_mwhy is there a need for utf8_strlen? UTF8 is guaranteed not to have any 0x00 characters13:40
dhx_munless you're trying to count actual characters (not buffer size)13:41
nuclear_eclipseRonald: ah ha!13:43
nuclear_eclipsethe logic in plugin_init_installed() for how many retries to perform is flawed - it doesn't make enough passes at initializing everything13:44
RonaldI guess the good thing about that is it negates the phase of cluelessness ;)13:45
nuclear_eclipseyep13:46
CIA-61Mantisbt: jreese * rc40a83d192c5 /core/plugin_api.php: Fixed bug with plugin initialization retries.13:49
nuclear_eclipseRonald: if you pull from latest git, or pull tonight's snapshot, you'll have the fix13:49
Ronaldwill do!13:49
nuclear_eclipseRonald: if you don't mind, I have a couple tips13:49
Ronaldsure, fire away13:49
nuclear_eclipseif you implement the class function ->init() and put your lib inclusion there, it will simplify your code so you don't have to load it in multiple places13:51
nuclear_eclipsealso, if you use `config_get_global( 'plugin_path' ) . 'Fadiro/<lib>'` in your requires calls, it will work properly even if you later decide to put plugins elsewhere for some reason13:52
nuclear_eclipseand lastly, you don't need to explicitly require MantisPlugin.class.php -- the plugin system will have it loaded by the time it ever gets around to executing your code13:53
nuclear_eclipsenow, all that said, it's just recemmendations, everything you did is still perfectly fine, so cheers :)13:54
RonaldIts welcome.... The mantisplugin require is from XmlImportExport (I copied that to have some guidance)13:55
Ronaldif its not needed you may want to remove it there ;)13:57
nuclear_eclipseI honestly haven't really looked at that plugin, as it was created by giallu13:57
RonaldAh well, seems he was unfamiliar too ;)13:58
gialluwhat's that?14:00
RonaldA assumed the db_prapare_* functions are there to sanitize untrusted input; so it becomes safe from exploits to the db?14:00
nuclear_eclipseRonald: correct14:01
gialluRonald, I believe that has changed lately with the query_bound stuff14:01
nuclear_eclipsegiallu: XmlImportExport plugin14:01
giallunuclear_eclipse, yeah what about it?14:01
nuclear_eclipseright, db_query_bound( $query, array( $param, ... ) ) is preferred14:02
RonaldI use that; but ALSO db_prepare_*14:02
Ronaldyou say the latter is not needed with the first?14:02
nuclear_eclipsegiallu: it explicitly requires the MantisPlugin class file, which isn't necessary as the plugin system will include it before it ever starts including plugins themselves14:02
nuclear_eclipseRonald: correct, it can actually result in double-escapes for strings14:03
Ronaldvery well, will clean up14:03
gialluah ok14:03
giallunuclear_eclipse, I think that was written because of "class XmlImportExportPlugin extends MantisPlugin"14:05
gialluI've the habit of including the base class file14:05
nuclear_eclipseright, it generally makes sense, but as plugin_api is guaranteed to load it first, each require_once after that is just a (very slight) performance hit, hence the reason I never put it in mine14:06
nuclear_eclipsebut it's really a non-issue IMO14:06
nuclear_eclipseit makes so little difference that I wouldn't even devote a commit to it :P14:06
RonaldThe only reason would be it sets an example for clueless people like me ;)14:07
nuclear_eclipsepaul_: do you need me to replace any other ereg* calls in the code?14:46
CIA-61Mantisbt: jreese * rf523d3c1372f /core/string_api.php: Migrate from ereg* to preg* for PHP 5.3 compat.15:35
CIA-61Mantisbt: jreese * ra583c8a1d534 / (5 files in 2 dirs): Add events for including/excluding users from notification lists.15:35
CIA-61Mantisbt: jreese * ra556dc0d55c5 / (4 files in 3 dirs): Fix #9744: Allow users to turn off session validation at login time.15:35
CIA-61Mantisbt: jreese * r0ca1c657c0e5 / (9 files in 4 dirs): Merge branches 'preg', 'notify' and 'proxy-login'15:36
paul_nuclear_eclipse: probably15:48
nuclear_eclipsepaul_: `ack " ereg"` found four files with ereg* that aren't in /library15:51
paul_1-2 of them are trivial right?16:07
vb123Good morning16:16
vb123nuclear_eclipse: ?16:16
vb123We should have associated issue #s for checkins like new plugin events and ereg->preg for 5.3 compatibility.16:17
vb123This is to make such enhancements / fixes discoverable for the users and to make it easier for us to find / resolve duplicates in the future.16:18
vb123Relating to the notification events, it would be useful to add an extra parameter.  For example, in case of adding a note, include the note id.  In case of adding a relationship include the relationship id.16:21
vb123We could have this set to 0 in the initial implementation, but it will avoid us having to change the plugins when we add this in the future.16:21
vb123I think we should eventually have a mode where MantisBT would craft the email with some basic issue info + the new note or the new relationship, rather than including all the issue details.16:22
vb123This will also allow plugins to potentially react differently to add new private note vs. add new public note.16:23
[KK]Kirillvb123: vboctor?16:39
nuclear_eclipsehi vb12316:51
vb1234hi nuclear_eclipse16:51
nuclear_eclipsewhy would we want users to know what's in a new release? :P16:52
vb1234:)16:52
vb1234if we are to add an extra parameter to an event, does this break existing plugins?16:53
nuclear_eclipseno16:53
nuclear_eclipsethe only way we break existing plugins is by changing or dropping existing parameters16:54
vb123that's good.16:59
vb123it would be useful to include in the documentation a sample for an implementation of each event.17:02
nuclear_eclipsevb123: how extensive are you thinking for the sample?17:04
vb123just something to show the arguments, how to return the results.17:06
vb123return an array vs a single element.  What parameters to take in, etc.17:06
nuclear_eclipseisn't that what the documentation of parameters and return value is for? :P17:06
vb123let me check again...17:06
nuclear_eclipseevery event is documented with a list of parameters passed and return value expected17:07
vb123for example, EVENT_MENU_MAIN_FRONT17:09
vb123it is not clear whether to return <a href....> or http://....  ---17:09
vb123it will be clearer if there is an example, or you can include the sample as part of the existing documentation.17:10
nuclear_eclipseah, I think that portion might be a special case :P17:10
vb123ok, I guess what we have may be OK, we should just make sure it is not ambiguous.17:13
nuclear_eclipseagreed17:13
nuclear_eclipsemaybe I can take care of that before any final release of 1.217:13
vb123sounds good. thanks.17:13
vb123are we any closer to the rc?17:13
nuclear_eclipseI think so17:14
nuclear_eclipseafaik, all the dates stuff is solidified17:14
vb123ok, are we able to do it this weekend?17:14
nuclear_eclipsealthough paul_ was wanting to get his bug object updates committed to master as well...17:15
vb123isn't that a big change?  DO we need it for 1.2?17:15
vb123if we get it in, we have to wait to stabilize it.17:15
* nuclear_eclipse looks for paul_'s link to the diff..17:15
vb123also it wasn't one of the issues we thought was necessary a couple of weeks ago.17:15
nuclear_eclipsewell, I could have sworn paul had pushed it in at the same time as his dates change, which is why I hadn't brought it up... =17:16
nuclear_eclipsehmm, can't find the likn17:20
nuclear_eclipseanyways, I don't even remember what it was that it was "fixing", so I'm assuming it wouldn't be too terrible to release without it, but the more stable our API is after making a plugin release, the better...17:21
dhx_mbefore 1.2, can rm-hardcoded-enum (bug 10330) be addressed?17:26
mantisbotBug 10330 - dhx - open - assigned17:26
mantisbotRemove all hardcoded enum levels from within Mantis code - http://www.mantisbt.org/bugs/view.php?id=1033017:26
dhx_mrm-hardcoded-enum is the branch I was working on17:26
dhx_mthere are still a few hardcoded values left which I'm not 100% sure how to fix17:26
nuclear_eclipseright, that's one of the things I was gonna take care of today :)17:27
dhx_mthis will close off a whole number of long standing bugs (5 years in some cases)17:27
dhx_mbug 3973 I'll fix soon (it is part of what 10330 aims to fix)17:27
mantisbotBug 3973 - Wanderer - open - assigned17:27
mantisbot"Reporter effectiveness" on summary calculated wrongly in case of using custom-severity attributes - http://www.mantisbt.org/bugs/view.php?id=397317:28
dhx_mand the only other TODO that comes to mind with that bug is what to replace instances of the FEEDBACK constant with17:29
dhx_mI think it is only used in a few places... for example a preset filter that shows all bugs at feedback level17:30
nuclear_eclipseyeah, I noticed that17:31
dhx_motherwise branches manage-user-email-notify and 10268 are 'ready'... with succint-html-buttons and 10274 also 'getting there'17:33
dhx_mif I get some time, I can try squashing some more old feature requests/bugs... but no point holding up 1.217:34
dhx_mit's already jam packed with new stuff :)17:34
nuclear_eclipsedhx_m: was rm-hardcoded-enum recently rebased onto master?17:55
nuclear_eclipsewait for it...18:23
CIA-61Mantisbt: hickseydr * ra0525539884f / (11 files in 2 dirs): Replace CLOSED with $g_bug_closed_status_threshold18:23
CIA-61Mantisbt: hickseydr * r5d3edf63aec0 /core/email_api.php: Replace RESOLVED with $g_bug_resolved_status_threshold18:23
CIA-61Mantisbt: hickseydr * rc7ecb89dacb0 /bug_change_status_page.php: Replace ASSIGNED with $g_bug_assigned_status18:23
CIA-61Mantisbt: hickseydr * r7b940ad5f931 / (5 files in 2 dirs): Replace NEW_ with $g_bug_submit_status18:23
CIA-61Mantisbt: hickseydr * r313abf14e50c / (5 files in 3 dirs): Rename OPEN to $g_default_bug_resolution18:23
CIA-61Mantisbt: hickseydr * rd162566dc287 / (8 files in 3 dirs): Remove hardcoded references to FIXED resolution18:23
CIA-61Mantisbt: hickseydr * rc007dee836e2 / (3 files in 3 dirs): Custom priority & severity significance thresholds18:23
CIA-61Mantisbt: hickseydr * rfd57267a10e0 /core/constant_inc.php: Add custom constants for projection and ETA fields18:23
CIA-61Mantisbt: hickseydr * r7b57cd410f97 / (api/soap/mc_api.php api/soap/mc_issue_api.php bug_report.php): Fix instances of enum levels hardcoded to 1018:23
CIA-61Mantisbt: hickseydr * rf74dbc75bddf / (2 files in 2 dirs): Config options to set default projection & ETA18:23
CIA-61Mantisbt: jreese * r16677b026002 / (23 files in 4 dirs): Merge commit 'dhx/rm-hardcoded-enum' into enums18:23
CIA-61Mantisbt: jreese * r9531f78d316d /.gitignore: Updated .gitignore18:23
CIA-61Mantisbt: hickseydr * rb51b4435dd7e / (bug_change_status_page.php manage_config_workflow_page.php): Replace REOPENED with $g_bug_reopen_resolution18:23
vb123dhx_m: thanks for this nice patch.  I specially like that the documentation was updated too :)  Also a great quality work.18:39
nuclear_eclipsevb123: I think this might also make for a useful addition to 1.2: http://git.mantisforge.org/w/mantisbt/jreese.git?a=shortlog;h=refs/heads/utf818:41
vb123what is this work about?18:43
nuclear_eclipseusing utf8-capable string functions whenever possible, throughout the whole codebase18:43
nuclear_eclipseie, there's a lot of places where we use non-utf8 string functions on strings that may or may not contain utf8 data18:44
vb123on a high level it makes sense.18:45
vb123Do we see the change as risky?18:45
vb123is it just the top two checkins?18:45
nuclear_eclipseI don't -- it's really only adding an api that doesnnothing but call either the same function, or a multibyte version of the same function, depending on what's available18:46
nuclear_eclipseyeah, just the two right now18:46
nuclear_eclipsemost of the real work is contained in core/utf8_api.php; the rest is more or less `sed` work to convert all usages to the new api18:47
nuclear_eclipseit's not yet a "complete" solution, because it needs to be further extended to cover things like str_pad(), but it's a first-step18:47
nuclear_eclipseand further work can easily by committed before a final release18:48
vb123It looks like a good change.  I have a couple of comments:18:50
vb1231. We should cover str_replace as well.  Even if utf8_str_replace ends up being a direct mapping for str_replace.  i.e. change all calls and just fix implemenations as appropriate.18:51
vb1232. Move the function_exists() inside the function definition rather than outside it.  This gaurantees the function interface to not depend on functions being defined.18:52
nuclear_eclipse2 is a very bad idea18:52
vb123so I would add utf8_str_pad and utf8_str_replace, call it from everywhere, then fix the implementation later if necessary.18:52
vb123This scopes the changes to be merged later to smaller changes.18:53
nuclear_eclipsea) performance, between branching and function_exists, b) php_api defines a random implementation of mb_substr that would interfere18:53
vb123shoudn't we get rid of the str related stuff in php_api.18:56
nuclear_eclipseI believe it's there to work around problems in phpmailer/adodb18:56
nuclear_eclipsebecause one of them (can't remember which) makes a call to mb_substr without checking for its presence, which would break on any PHP without the mbstring extension (the whole reason we need utf8_api)18:57
paul_nuclear_eclipse: erm18:57
paul_how exactly did you merge dhx/rm-hardcoded enum ?18:58
nuclear_eclipsevia `git merge dhx/rm-hardcoded-enum`, why?18:58
paul_http://git.mantisbt.org/?p=mantisbt.git;a=shortlog;h=refs/heads/master19:00
paul_doesn't show the changes19:00
paul_http://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=16677b02600202b5ebbb3fff9fe0d5c445bffa6219:00
nuclear_eclipseyes it does19:00
paul_it only modifies:19:01
paul_bug_graph_bystatus.phpdiff1 |diff2 |blob | history19:01
paul_core/email_api.phpdiff1 |diff2 |blob | history19:01
paul_manage_config_workflow_page.phpdiff1 |diff2 |blob | history?19:01
paul_?19:01
nuclear_eclipsethat's just the merge commit19:01
nuclear_eclipsethe real commits are further down19:01
nuclear_eclipseeg, all the ones by "David Hicks" (dhx)19:01
paul_that's a mess19:02
nuclear_eclipsewhat do you mean?19:03
nuclear_eclipsethat's the way all DVCS systems work, including your beloved Hg19:03
paul_so anyway, I'll push the bug objects stuff19:06
nuclear_eclipseok19:06
vb123paul_: what is the goal of the bug objects work?  Was it reviewed?19:16
paul_yes19:16
paul_john iirc19:17
vb123I would like to understand what we are aiming for and make sure that we at least get a couple of devs to review a big change.19:18
paul_anyway, it's in branch on mantisforge19:19
vb123I would just raise the awareness of this and send to mantisbt-dev19:22
vb123Also we need to consider whether we want to do this for 1.2 or 1.3.19:22
paul_1.2 imo19:31
paul_vb123: did you clarfiy if we prefer if( or if ( ?19:37
nuclear_eclipsevb123: no need to make a utf8_str_replace(); it's utf8 safe as it is19:47
* paul_ slaps nuclear_eclipse 19:48
paul_did you look over branch again?19:48
nuclear_eclipsepaul_: no, it seemed like it had a lot of other stuff in it that wasn't related to bug objects...19:49
paul_probably19:50
paul_as it's now my local master ;p19:50
nuclear_eclipsepaul_: you scare me...19:56
paul_now what19:56
nuclear_eclipse15:50 < paul_> as it's now my local master ;p19:57
paul_? :)19:57
paul_erm20:00
paul_you get asked about the session thing on logon????20:00
paul_if a user turns that off, not knowing what it means, that's a security risk ;/20:00
nuclear_eclipseduh20:03
nuclear_eclipsethat's why it's called "Secure Session"20:03
paul_I thought you meant you'd put it uner user prefs to configures :P20:03
paul_-s+d20:03
nuclear_eclipsehow can you get into your preferences to disable it if your session won't validate after login?20:03
nuclear_eclipsebesides, this is exactly how other sites/applications handle it; see sf.net20:05
nuclear_eclipseor at least they used to have that option20:05
paul_so push?20:08
nuclear_eclipsesounded like vb123 wanted to review it first...20:09
* paul_ sighs20:10
vb123nuclear_eclipse: relating to utf8_str_replace, I think ideally we should have a simple rule, any str related function we should use utf8_* rather than everyone considering whether the operation is safe or not.20:17
paul_vb123: you wanted to review bugobject changes from a month ago?20:18
nuclear_eclipsevb123: I would generally agree, but I've actually found an external library that will be complete/stable/tested than rolling our own utf8 stuff;  I'm in the middle of reworking what I have20:19
nuclear_eclipses/be/be more/20:19
* paul_ sighs20:20
nuclear_eclipsecalm down paul_, it's an identical interface to what I would have implemented20:21
paul_you know i've already said i've worked out utf8 compat functions for most things already?20:21
nuclear_eclipsehttp://sourceforge.net/projects/phputf8/20:21
paul_seen that20:21
vb123we should make sure dependencies makes sense from linux distro's point of view.20:21
paul_it's not been updated for 2 years iirc20:22
nuclear_eclipsewell, no, but does it really need to be?  it's not like utf-8 is a moving target...20:22
paul_I looked at that a month ago and concluded no20:22
vb123I think we should put together our utf8_api, then maybe in the future, change it to utilize a library.20:23
vb123same like db_api20:23
paul_nuclear_eclipse: iirc, the utf8_substr is the 3rd or 4th slowest of the 5 utf8_substr functions I tested20:24
paul_in addition, iirc the licensing for that library is no good for us20:26
nuclear_eclipseit's LGPL, how is that no good?20:28
paul_it's been a month since I looked at it :P20:28
paul_I just concluded a month ago, not to touch it ;p20:29
nuclear_eclipseand why implement our own if the library is identical to what we would have created?20:29
paul_less hassle :P20:29
nuclear_eclipseand why is raw performance the determining factor?20:29
paul_a library that's not been updated for 3 years isn't a good choice of library20:30
nuclear_eclipsehow is implementing everything ourself *less*  hassle?20:30
nuclear_eclipseit doesnt need to be updated !20:30
nuclear_eclipseits utf820:30
paul_it needs to be updated to support improvements in php for example20:30
nuclear_eclipsebesides? it's not like we couldnt maintain it ourselves just as well as maintaining our own utf8 lib?20:31
paul_as then distributions whinge20:32
paul_which means it's more hassle20:32
paul_for example20:32
paul_I plan to patch nusoap for php5.320:32
nuclear_eclipsewhy must we always reinvent the wheel every freakin time?20:33
paul_it's fine to use a library that's kept up to date20:33
paul_imo20:33
paul_and is active20:33
nuclear_eclipseand btw, phputf8 is not packaged by debian/fedora, so it wouldn't matter20:33
paul_yea, but they'll want to package it20:33
nuclear_eclipseif they haven't packaged it in the three/four years it's been used by lots of other apps, why would they all of sudden want to package it now that we use it?20:34
paul_anyway most of phputf8 is from dokuwiki20:35
nuclear_eclipseand seriously why is rolling our own lib any better than just starting with an unsupported lib?  we have to maintain both ourselves, why not use the one that's already implemented and has plenty of testcases against it?20:35
nuclear_eclipsethe whole NIH syndrome makes me want to bash my head against the wall....20:35
paul_i've spent a chunk of time working out the best utf_* functions from various sources (e.g. mediawiki/dokuwiki and so forth over the last month20:35
paul_which I plan to dump in20:36
nuclear_eclipsepaul_: I still don't see how that's any better than a single documented library, just because the lib is unmaintained...20:36
paul_because we've already spent a bunch of time comparing the functions in that library to other implementations of the functions20:37
nuclear_eclipsefine, whatever20:38
vb123If the library is not bloated and does exactly what we need and we understand it, then we should use it.  That is my opinion as well.  However, I didn't look at it.  This is just my 2c.20:38
paul_you've just picked a topic where20:39
paul_i've been through code base20:39
paul_evaluated what str functions we have that dont support utf820:39
paul_evaluted where we use them that requires utf820:39
paul_and evaluated compatibility functions without requiring php extensions for those uses20:39
paul_to come up with some possible solutions20:39
nuclear_eclipsepaul_: I still you're taking the far too20:40
nuclear_eclipseoops20:40
paul_nod, I agree20:40
nuclear_eclipseI think you're taking the route that costs a lot more time and effort20:40
paul_for the most part, we probably want to use utf8_strlen or whatever20:40
paul_however20:40
paul_there are some functions we dont want/need to use it in (is_Blank an obvious one)20:40
nuclear_eclipsepaul_: well, like I said, is_blank shouldn't be using *any* form of string function...20:41
paul_there are some uses where we need to evaluate if we want to use utf8_strlen (i'm thinking functions where we upload files that could contain utf8 characters and insert them into a blob)20:41
nuclear_eclipseotherwise, I disagree with your plan to evaluate every single use of a string function to look for utf8 needs...20:41
paul_nuclear_eclipse: i'd agree with you on is_blank, however, strlen seemed to give better results when profiling (strangly)20:42
nuclear_eclipseif anything, that may just be because strlen is part of the C implementation20:42
nuclear_eclipsewhat did you profile against?20:43
paul_empty maybe20:43
* paul_ can't recall20:43
nuclear_eclipseand I still think that performance does not trump code consistency/maintainability20:43
nuclear_eclipseif empty() doesn't work right, then why are you testing against it?20:44
paul_empty && foo != 020:44
paul_is what i tried iirc20:44
paul_or more20:44
paul_empty && foo[0] == 020:44
paul_I can't even see hwhere they define strlen20:50
nuclear_eclipsepaul_: it depends on whether mbstring is enabled, if so, in mbstring/core.php, otherwise in native/strlen.php20:54
paul_i mean in php source ;p20:54
nuclear_eclipseoh20:55
paul_but anyway20:58
paul_right now i just need to know if i'm pushing bugobject changes from a month ago :P20:59
paul_incidently john20:59
paul_can we use:20:59
paul_.double a {20:59
paul_border-bottom: 3px double;20:59
paul_line-height: 1.7em;20:59
paul_}20:59
paul_to display the double underline in changelog/roadmap20:59
paul_aka css?20:59
nuclear_eclipsenot really21:00
paul_why not? :)21:01
paul_i mean, wouldn't that allow users to theme/style it21:01
nuclear_eclipsea big benefit of the <pre> output from that page is that it can go directly into a text file and look identical ;)21:01
nuclear_eclipseeg, mantisbt/docs/ChangeLog21:01
paul_true21:01
nuclear_eclipseanywho, paul_ / vb123 : here's a branch with the phputf8 library implemented, for both substr and str_pad: http://git.mantisforge.org/w/mantisbt/jreese.git?a=shortlog;h=refs/heads/phputf821:02
nuclear_eclipsewe could just push this right now, and then modify for performance later; it is an LGPL library, so we can modify it without issue21:03
paul_I'd rather the old branch21:03
nuclear_eclipseI really don't understand why; it's basically identical to the old branch, just with a better implementation and a lot of tests21:04
nuclear_eclipseseriously, the only difference between this and the old branch is where the functionss come from21:04
paul_I concluded a month ago that some of the function implementations aren't great21:05
nuclear_eclipseso we can "fix" them later21:05
paul_as I looked at that library back in feb/march21:05
paul_and concluded *not* to use it21:05
nuclear_eclipsewe're not *stuck* with the implementations paul21:05
nuclear_eclipsebut it gets us 100% for free, and under a license allowing us to change whatever the hell we feel like changing...21:06
nuclear_eclipseI really don't know what it is that you have against using a library that we didn't create...21:06
nuclear_eclipseanywho, time to head home21:11
* paul_ smacks vb123 21:32
* paul_ smacks vb123 harder21:32
paul_nuclear_eclipse: when are we releasing 1.1 release?21:43

Generated by irclog2html.py