Monday, 2010-11-01

*** Quits: roentgen (~arthur@miranda/user/roentgen) (Remote host closed the connection)01:49
*** Joins: kirillka (~Miranda@195.242.142.17)02:16
*** Quits: Ragnor (~Ragnor@dslb-092-072-246-212.pools.arcor-ip.net) (Disconnected by services)04:06
*** Joins: Ragnor (~Ragnor@dslb-188-100-038-131.pools.arcor-ip.net)04:07
*** Joins: rolfkleef (~rolf@urtica.xs4all.nl)04:07
*** Quits: dhx1 (~anonymous@c122-107-170-247.eburwd5.vic.optusnet.com.au) (Remote host closed the connection)04:30
*** Quits: siebrand (~beis@sm.xs4all.nl) ()04:34
*** Joins: dhx1 (~anonymous@c122-107-170-247.eburwd5.vic.optusnet.com.au)04:52
*** Joins: cobexer (~cobexer@188-22-111-92.adsl.highway.telekom.at)05:58
*** Quits: cobexer (~cobexer@188-22-111-92.adsl.highway.telekom.at) (Write error: Connection reset by peer)06:07
*** Joins: cobexer (~cobexer@188-22-111-92.adsl.highway.telekom.at)06:08
*** Joins: davidinc (~davidinc@213.55.100.134)06:26
*** Joins: istvanb (d917e473@gateway/web/freenode/ip.217.23.228.115)06:45
istvanbhello!06:46
istvanbI can not use the "assigned to" filter properly if the "all project" is selected. Then I see only the users with admin account.06:47
istvanbis this something mantis specific?06:47
dhx1istvanb: hi06:48
istvanbhi06:48
dhx1istvanb: I noticed you were trying to ask me a question the other day06:49
istvanbyeah :)06:49
istvanbI think I forgot about that... probably related one of your issues06:49
istvanbgeez06:49
istvanblet me thing06:49
istvanbthink06:49
dhx1custom fields06:49
istvanbyeah06:50
istvanbI have asked the question on the forums06:50
dhx1the problem you just described is because you're browsing in the "All projects" context and I imagine it is only returning a list of users who are able to be assigned to an issue in the global context06:50
istvanbI just tried to use cusom fields06:50
dhx1it really needs to find users who can be assigned to an issue in *any* context06:51
dhx1or more precisely, any child project of the root "ALL PROJECTS"06:51
istvanbok... so lets back to the "all projects" issue first06:51
*** Joins: bartik2003 (~bartik200@92-60-59-178.rev.bonet.sk)06:51
istvanbI understand what you say, however it looks really inconvenient for me (user perspective) because if the given user has 10 projects then I can not get a quick overview of his workload06:52
dhx1agreed06:52
dhx1is there a bug for this issue?06:52
istvanbhttp://www.mantisbt.org/bugs/view.php?id=457006:53
istvanbclosed long long time ago06:53
istvanbit looks like it was fixed once06:54
bartik2003at least in 1.1.6 it works.06:55
istvanbcan anybody confirms that this is an issue with 1.2.3 ?06:56
istvanbI dont want to ring the bells before its clear that it is not related to my computer only06:56
*** Joins: cobexer_ (~cobexer@88-117-55-224.adsl.highway.telekom.at)06:56
*** cobexer_ is now known as \cobexer|away06:56
*** Quits: cobexer (~cobexer@188-22-111-92.adsl.highway.telekom.at) (Ping timeout: 240 seconds)06:57
*** Quits: istvanb (d917e473@gateway/web/freenode/ip.217.23.228.115) (Ping timeout: 265 seconds)07:01
bartik2003Just a quick look into the code but project_get_all_user_rows seems to return the right list in 1.2.307:02
*** Joins: istvanb (d917e473@gateway/web/freenode/ip.217.23.228.115)07:03
istvanbbrb07:03
dhx1thanks bartik200307:07
bartik2003how can I submin minor changes ?07:08
bartik2003you still have a weird file_get_extension07:08
bartik2003:-)07:08
bartik2003my version contains only this line: return pathinfo($p_filename, PATHINFO_EXTENSION);07:09
dhx1the best approach is to file a new bug report with an attached patch in git format (git format-patch)07:09
bartik2003it is not really a bug07:09
bartik2003And my presumption is that pathinfo is utf8 safe...07:10
bartik2003I see you have added tests, thats nice.07:10
dhx1http://php.net/manual/en/function.pathinfo.php07:11
dhx1I was trying to phase out "file_get_extension" completely07:11
dhx1MantisBT 1.2.x already uses fileinfo (magicdb MIME detection) to determine file types instead of the naive file extension approach07:12
istvanbguys I am sorry I had to be away...07:13
istvanbso can anybody confirms that the filtering thing is an issue with 1.2.3?07:13
bartik2003Hmm. Would this detected a csv file ? I see only text/plain :-)07:14
bartik2003Another thing I have noticed. the gpc_ is working with _POST _GET etc. you coud use _REQUEST http://php.net/manual/en/reserved.variables.request.php07:15
dhx1bartik2003: it isn't that smart AFAIK, but it can determine binary file formats more accurately07:15
bartik2003dhx1: that is ok. I just meant that sometimes the extension can come handy.07:15
dhx1bartik2003: perhaps file extensions still have some use as overrides for cases where MIME detection isn't possible (CSV files, etc)07:15
bartik2003dhx1: to be honest I was now almost year and a half tinkering with mantis. Somehaow I became to like it :-)07:18
istvanbhello hello hello hello :)07:18
istvanbdont forget about the filtering thing :)07:18
bartik2003dhx1: Have you ever thought about moving the languages into the database ?07:20
dhx1bartik2003: I think it tends to be easy to customise (whether via configuration options or via hacking the code)07:20
dhx1bartik2003: not at this stage, what benefits would it bring?07:20
dhx1It'd actually be pretty easy to implement07:21
istvanbbrb07:21
dhx1(in a naive form with 1 query per string)07:21
dhx1istvanb: can you file a bug report describing the issue?07:21
bartik2003Well the whole thing about moving all or as much components to the dabase would give you the possibility to move onlz the database and have functioning install. Like backup copies, test systems...07:21
dhx1I agree07:22
bartik2003Somehow I'm a bit influenced by SAP...07:22
dhx1but to move languages (and configuration) entirely into the database I think we'd need to restructure most of the code07:22
istvanbok07:23
dhx1well actually maybe not...07:23
bartik2003Another thing I have hit lately is initiating long running processes like importing large csv files.07:23
dhx1it depends on the efficiency of loading 1000's of language strings from the database indiscriminately or whether it is better to select only strings which are required (using huge SELECT queries)07:24
dhx1both approaches seem very ugly/slow to me07:24
bartik2003Thought about the db caching of SQL statements ?07:25
bartik2003http://www.databasejournal.com/features/mysql/article.php/3110171/MySQLs-Query-Cache.htm07:25
bartik2003mantis has a caching mechanism of course but one on the DB level might help too I just don't have that big of a db to test with...07:27
dhx1yep we'd need to test to see whether it's still efficient to load language data from the database with SQL server caching07:27
dhx1as the queries required could contain 100+ conditions (WHERE translation_key='username' OR translation_key='password'....)07:28
bartik2003sorry when I'm too overwhelming. I just don't get that much time if any to chat about mantis.07:29
bartik2003why that much conditions ? to be honest I have seen onlz the lang_api.php code where I did my changes (project dependent translation)07:30
dhx1we've already got a huge list of other priorities to sort through, so I don't see languages moving to the database any time soon (cost vs benefit isn't strong enough)07:31
dhx1well the alternative is to retrieve 5000 translation strings (or whatever it could turn out to be) from the database on each page load07:31
dhx1which is a few megabytes at a time (let's say 3MB)07:31
dhx1I know it can be cached, but it still seems like a lot of overhead for little gain07:32
bartik2003Is this huge list accessible ?07:33
dhx1bartik2003: the #1 priority is really a UI redesign of MantisBT (which requires separation of logic and UI code, replacement of key parts of the UI API, etc)07:34
dhx1then we have other issues such as custom field functionality, user groups, authentication and access control, etc07:34
bartik2003Don't get me wrong i'm not looking to request this. I see this rather as my hobby implementation :-)07:34
dhx1if you'd like to do some work towards this issue, please feel free to do so07:35
bartik2003I did use custom fields extension from http://deboutv.free.fr/mantis/07:35
dhx1but as with all large undertakings, it is best to communicate on the mantisbt-dev mailing list to make sure you get some advice/feedback first :)07:35
dhx1those plugins are for MantisBT 1.1.x only though07:35
dhx1MantisBT 1.2.x has it's own new plugin system07:36
bartik2003dhx1: of course I't would be shame to not get a advice and then realize somebody else did it better :-)07:36
dhx1bartik2003: well more to the point, to make sure that the theory is tested before many hours are wasted on something that will never work :)07:37
dhx1for instance, sending off 100-200 queries per page load would be unacceptable07:37
dhx1and perhaps requesting 3MB of language strings from the database is also horrible in terms of performance (I'm almost certain it would be)07:37
bartik2003Yes I know about the 1.2.x plugin szstem will have to move there. But when I started 1.1.6 was the newest and I have done some customizing like avatars, csv importing...07:38
bartik2003dhx: but given that in most cases you are using onlz 1/2 languages having those 3MB cached (for the whole language) might be advantageous.07:40
dhx1bartik2003: the problem is likely going to be PHP's performance with accessing the array of data returned by the database query functions07:41
dhx1bartik2003: however it could turn out to be high performance... it's just a matter of running tests to see whether it is a viable option07:42
istvanbdhx1: yeah, no probs07:42
bartik2003dhx1: ever thought about using an existing framework ?07:43
dhx1bartik2003: yep there has been a bit of discussion about using Zend in MantisBT (especially for authentication)07:43
bartik2003wouldn't that imply licensing (using zend)07:45
bartik2003ah I see BSD and Apache licenses.07:45
dhx1http://framework.zend.com/license07:46
dhx1modified BSD license07:46
bartik2003well I wasn't aiming that high :-) I was looking at something like cake, radicore seems nice but has to be licensed and a bung of other things... but radicore is worth reading http://www.tonymarston.net/php-mysql/data-dictionary.html for example07:50
istvanbI have filed an issue: http://www.mantisbt.org/bugs/view.php?id=1250107:52
dhx1istvanb: thanks07:52
istvanbI know you guys are busy with the secret php code of the nuclear silo, but can I ask my question about the custom fields?07:53
dhx1bartik2003: yeah there are a lot frameworks out there (one of the common complaints about PHP programmers is their tendency to create n+1000 new frameworks :p)07:54
nuclear_eclipsethe worst part is seeing frameworks for your frameworks07:55
dhx1nuclear_eclipse: haha07:55
nuclear_eclipse"here's a framework framework to create frameworks for your framework"07:55
istvanbI am just wondering if I have a custom enum, and the Required on report is off as well as the display on report then I guess the enum should be stored with its default value. Is this something what you guys plan?07:56
bartik2003we'll never used anz of those frameworks not that I didn't try but perhaps it's just that for my small projects a framework is an overkill...07:56
dhx1I'm trying to work out why someone would use PHP instead of C++ if you're going to get serious with a website07:56
dhx1*web app07:56
nuclear_eclipsedhx1: PHP scales really well for a dynamic language, and is a lot easier and faster to write and deploy with than C++07:57
dhx1well I guess you don't need to compile it after each code change07:57
nuclear_eclipsentm that there are very few "good" HTTP interaction models for C++07:58
bartik2003scripting languages have that nice thing that they can generate themselves on the fly :-D07:58
nuclear_eclipseexactly07:58
dhx1but then you have downsides: bad debugging abilities, poor performance, etc07:58
dhx1what's wrong with "make"? :p07:58
bartik2003I was looking lately at this (in conjuction with cpp) http://www.ultimatepp.org/07:59
nuclear_eclipseyou can't exactly have a C++ program dynamically route URLs like "/<class>/<action>" without somehow linking in all those classes/actions ahead of time, which kind of defeahs the purpose of dynamic routing07:59
nuclear_eclipsedhx1: with xdebug + eclipse, PHP is actually really nice to debug, a lot better to debug than C++ in fact08:00
dhx1I assume you're referring to plugins with your comment about routing?08:01
nuclear_eclipsenot necessarily08:01
nuclear_eclipsehttp://codeigniter.com/user_guide/general/routing.html08:01
nuclear_eclipseI use CI for the application I'm currently working on for my employer, and you can set up routing patterns that map to class/method names and after defining the patterns, creating "new routes" is as simple as dropping a new file/class into the controllers directory08:03
dhx1I'm not sure I see the problem... wouldn't you just write a "router" within the core of the C++ web app that allows URLs to be registered against callbacks?08:03
dhx1so any request in the format \/bugs\/.+ (regex) would be routed to a callback that is designed to handle requests to view a bug?08:04
nuclear_eclipsethe difference is that in C++ you would have to have a well-defined method for linking/registering all the routes/classes/methods that you have, as opposed to being able to just blindly load/call methods on the fly08:04
dhx1dynamic library loading?08:05
nuclear_eclipsethe point is that in PHP, it can just dynamically search and execute classes without ever having to know about them ahead of time08:05
nuclear_eclipsein C++, you have to define all of your applications interacitons ahead of time before you can start using them08:05
dhx1yep08:05
*** Quits: djSupport (~djsupport@188-221-240-190.zone12.bethere.co.uk) (Read error: Connection reset by peer)08:06
dhx1bug you could also load a plugin dynamic link library and have it register a callback upon loading?08:06
nuclear_eclipseyes, but it's the act of having to register a set of callbacks that is defining the actions ahead of time -- you have to code that registration for every class/route ahead of time08:07
nuclear_eclipsein something like CI, *all you have to do* is create a new method on your class, and that's it08:07
nuclear_eclipseor create a new class altogether08:07
nuclear_eclipsethere's no need to write code to register your classes/methods with routes08:08
dhx1ah I see08:08
dhx1it's automagic08:08
nuclear_eclipseand if you really want the performance of C++, just do what Facebook does08:09
nuclear_eclipsehttp://github.com/facebook/hiphop-php/wiki08:10
dhx1hmm08:12
dhx1you don't get type safety with that though08:12
nuclear_eclipsebut the reality is that straight PHP+APC is almost always more than you will ever need08:12
dhx1yep08:12
dhx1for MantisBT I would have thought the concern is more with database design08:13
nuclear_eclipseexactly, Mantis is extremely inefficient, which is more of a problem than a slow interpreter08:13
dhx1I'm not sure if you've tested the performance of queries with 100+ WHERE conditions vs. returning 3000+ rows of data and parsing with PHP?08:15
dhx1I was discussing this before with bartik2003 in relation to storing translations in the database08:15
*** Quits: istvanb (d917e473@gateway/web/freenode/ip.217.23.228.115) (Quit: Page closed)08:15
nuclear_eclipsedhx1: it's a matter of where you want the processing load08:17
dhx1probably at the SQL server end where it's running as compiled machine code... complete with query caching, etc08:17
nuclear_eclipseI've never tested it, but I would imagine that you'd be better off leaving the filtering to the database, to reduce memory and transfer overhead, even if you do remaining sorting in the codebase08:18
dhx1I guess it'd be highly dependant on the efficiency of each SQL server08:18
nuclear_eclipsentm the database is (usually) already extremely optimized for the filtering tasks, as opposed to having to write your own filtering code08:18
dhx1yep08:19
nuclear_eclipseanyways, I gotta go to work08:19
nuclear_eclipseI'll leave you with this: http://devzone.zend.com/article/2803-Introducing-xdebug08:19
dhx1ok ttyl :)08:19
nuclear_eclipseblows C++ out of the water as far as I'm concerned08:20
nuclear_eclipsecheers08:20
dhx1C++ = gdb though :)08:20
dhx1oops now I remember, it has a remote ability to wait for a debugger (gdb) to attach to the PHP process08:21
dhx1ignore that previous comment08:22
bartik2003sorry I was dancing ddr winx club with my daugters08:27
bartik2003so the next step is configure a big cache for mysql :-)08:27
*** Joins: daryn (~daryn@h158.249.190.173.static.ip.windstream.net)09:11
*** Quits: davidinc (~davidinc@213.55.100.134) (Ping timeout: 250 seconds)10:33
*** Quits: bartik2003 (~bartik200@92-60-59-178.rev.bonet.sk) (Ping timeout: 255 seconds)10:39
*** \cobexer|away is now known as cobexer_10:54
*** cobexer_ is now known as cobexer10:54
*** Joins: roentgen (~arthur@miranda/user/roentgen)10:59
*** Quits: kirillka (~Miranda@195.242.142.17) (Quit: kirillka)11:01
*** Quits: daryn (~daryn@h158.249.190.173.static.ip.windstream.net) (Remote host closed the connection)11:38
*** Joins: daryn (~daryn@h158.249.190.173.static.ip.windstream.net)11:39
*** Joins: siebrand (~beis@sm.xs4all.nl)11:42
CIA-27Mantisbt: s.mazeland master-1.2.x * re1a519c16444 / (3 files in 2 dirs): Fix #10887: 'Remove User' button to remove user from project is confusing.11:50
CIA-27Mantisbt: s.mazeland master-1.2.x * r0757e95cb7df /core/print_api.php: Fix #12500: Missing localization in extended project browser.11:50
CIA-27Mantisbt: s.mazeland master-1.2.x * ra256ece0d679 / (63 files in 7 dirs): * Replace &nbsp; with HTML5 compliant &#160;11:50
CIA-27Mantisbt: s.mazeland master-1.2.x * r325c6a1daa14 /lang/strings_english.txt: Fix #9036: Message confirm_custom_field_unlinking has confusing text.11:50
*** Quits: cobexer (~cobexer@88-117-55-224.adsl.highway.telekom.at) (Remote host closed the connection)12:04
*** Quits: roentgen (~arthur@miranda/user/roentgen) (Ping timeout: 276 seconds)12:08
*** Joins: cobexer (~cobexer@88-117-55-224.adsl.highway.telekom.at)12:10
*** Joins: roentgen (~arthur@miranda/user/roentgen)12:24
*** Joins: Github (~Github@sh1-ext.rs.github.com)12:30
Githubmantisbt: master-1.2.x Siebrand Mazeland * 0757e95 (1 files in 1 dirs): Fix #12500: Missing localization in extended project browser.12:30
Githubmantisbt: master-1.2.x Siebrand Mazeland * a256ece (63 files in 7 dirs): * Replace &nbsp; with HTML5 compliant &#160; ...12:30
Githubmantisbt: master-1.2.x Siebrand Mazeland * e1a519c (3 files in 2 dirs): Fix #10887: 'Remove User' button to remove user from project is confusing. ...12:30
Githubmantisbt: master-1.2.x Siebrand Mazeland * 325c6a1 (1 files in 1 dirs): Fix #9036: Message confirm_custom_field_unlinking has confusing text. ...12:30
Githubmantisbt: master-1.2.x commits a9b22d8...325c6a1 - http://bit.ly/cRqbdg12:30
*** Parts: Github (~Github@sh1-ext.rs.github.com)12:30
*** cobexer is now known as \cobexer|away12:45
*** Quits: micahg (~micah@ubuntu/member/micahg) (Ping timeout: 264 seconds)12:55
*** Quits: \cobexer|away (~cobexer@88-117-55-224.adsl.highway.telekom.at) (Read error: Operation timed out)13:02
*** Joins: \cobexer|away (~cobexer@88-117-55-224.adsl.highway.telekom.at)13:05
*** Joins: moto-moi (~hylke@2001:888:13e4:0:21f:e2ff:fe0c:ce28)13:44
*** Quits: Rixie (~Rixie@0x4dd7390e.adsl.cybercity.dk) (Quit: Rixie)13:44
*** Quits: \cobexer|away (~cobexer@88-117-55-224.adsl.highway.telekom.at) (Remote host closed the connection)13:58
*** Joins: micahg (~micah@ubuntu/member/micahg)14:22
*** Joins: djSupport (~djsupport@188-221-240-190.zone12.bethere.co.uk)14:28
*** Joins: djSupport_ (~djsupport@188-221-240-190.zone12.bethere.co.uk)15:21
*** Quits: djSupport (~djsupport@188-221-240-190.zone12.bethere.co.uk) (Ping timeout: 245 seconds)15:23
*** Quits: mellen (~thansen@x1-6-00-22-02-00-0c-40.k253.webspeed.dk) (Ping timeout: 276 seconds)16:00
*** Joins: paulr (~IceChat09@2001:470:9310:aaaa:38c9:32c9:84bf:2013)16:29
paulrmoo16:29
*** Joins: mellen (~thansen@x1-6-00-22-02-00-0c-40.k253.webspeed.dk)16:56
*** Quits: daryn (~daryn@h158.249.190.173.static.ip.windstream.net) (Quit: Ex-Chat)17:58
*** Quits: roentgen (~arthur@miranda/user/roentgen) (Remote host closed the connection)18:09
*** Quits: moto-moi (~hylke@2001:888:13e4:0:21f:e2ff:fe0c:ce28) (Ping timeout: 276 seconds)18:55
*** Quits: djSupport_ (~djsupport@188-221-240-190.zone12.bethere.co.uk) (Read error: Connection reset by peer)19:21
*** Joins: djSupport (~djsupport@188-221-240-190.zone12.bethere.co.uk)19:21
*** Quits: rolfkleef (~rolf@urtica.xs4all.nl) (Ping timeout: 276 seconds)19:39
*** Quits: micahg (~micah@ubuntu/member/micahg) (Ping timeout: 255 seconds)19:53
*** Quits: paulr (~IceChat09@2001:470:9310:aaaa:38c9:32c9:84bf:2013) (Quit: Man who run behind car get exhausted)19:59
*** Quits: scribe9343423 (~scribe934@static.96.23.63.178.clients.your-server.de) (Remote host closed the connection)20:00
*** Joins: scribe9343423 (~scribe934@static.96.23.63.178.clients.your-server.de)20:00
*** Quits: djSupport (~djsupport@188-221-240-190.zone12.bethere.co.uk) (Read error: Connection reset by peer)20:41
*** Joins: daryn (~daryn@h222.220.31.71.dynamic.ip.windstream.net)21:10
*** Quits: daryn (~daryn@h222.220.31.71.dynamic.ip.windstream.net) (Ping timeout: 252 seconds)21:41
*** Joins: daryn (~daryn@h165.80.88.75.dynamic.ip.windstream.net)21:44
*** Quits: daryn (~daryn@h165.80.88.75.dynamic.ip.windstream.net) (Ping timeout: 250 seconds)21:49
*** Joins: daryn (~daryn@h165.80.88.75.dynamic.ip.windstream.net)21:53
*** Quits: daryn (~daryn@h165.80.88.75.dynamic.ip.windstream.net) (Ping timeout: 255 seconds)22:03
*** Joins: micahg (~micah@ubuntu/member/micahg)22:13
*** Joins: daryn (~daryn@h203.144.28.71.dynamic.ip.windstream.net)22:16
*** Quits: giallu (~giallu@fedora/giallu) (Ping timeout: 245 seconds)22:36
*** Quits: daryn (~daryn@h203.144.28.71.dynamic.ip.windstream.net) (Ping timeout: 264 seconds)23:31

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!