| ../irclogs/#mantishelp.2010-02-08.log | ||
| --- scribe started --- | 00:00 | |
| CIA-24 | Mantisbt: hickseydr master-1.2.x * r5f7cef9d27b1 /core/ (lang_api.php user_pref_api.php): Fix #11394: Lost password email not sent when user language is invalid | 01:11 |
|---|---|---|
| CIA-24 | Mantisbt: hickseydr * rc8e30df5d714 /core/ (lang_api.php user_pref_api.php): Fix #11394: Lost password email not sent when user language is invalid | 01:11 |
| CIA-24 | Mantisbt: hickseydr master-1.2.x * r8f92d886cede /issues_rss.php: Fix #8539: Special characters not parsed correctly in RSS output | 02:02 |
| CIA-24 | Mantisbt: hickseydr * raa058c53f537 /issues_rss.php: Fix #8539: Special characters not parsed correctly in RSS output | 02:02 |
| CIA-24 | Mantisbt: hickseydr * rcca25660163b / (3 files in 2 dirs): Issue #10059: Select 'duplicate' resolution automatically | 03:01 |
| davidinc | dhx_m: hello | 06:48 |
| dhx_m | davidinc: hi | 06:48 |
| dhx_m | ...? | 06:50 |
| davidinc | nice 2 c u! | 06:52 |
| dhx_m | have we talked before? :) | 06:54 |
| dhx_m | turing test: provide the name of a planet in the solar system :p | 07:07 |
| Heady| | . | 09:01 |
| dhx_m | hi | 09:04 |
| dhx_m | ohhh something exciting: | 11:07 |
| CIA-24 | Mantisbt: hickseydr * r045a89705c3c / (10 files in 5 dirs): Issue #10730: Implement new crypto_api | 11:07 |
| paul__ | dhx_m: maintanence mode??? | 11:27 |
| dhx_m | paul__: sounded better than "MANTIS_INSTALLER" :) | 11:27 |
| dhx_m | ie. it loads the MantisDB core in a special maintenance mode for upgrading, installation, etc | 11:28 |
| dhx_m | ie. don't connect to the database, don't load plugins, don't check configuration, etc | 11:28 |
| * paul__ isn't sure about this idea :) | 11:30 | |
| dhx_m | neither am I :p | 11:31 |
| dhx_m | I don't really like the idea of the installer and other admin/ scripts loading the Mantis core at all | 11:31 |
| dhx_m | IMO they should have their own "core" | 11:31 |
| dhx_m | where "core" is "minimum operating environment" | 11:31 |
| dhx_m | ie. connect to the database, perform startup checks, etc | 11:32 |
| paul__ | I meant cyrpto changes ;/ | 11:32 |
| paul__ | I thought mt_srand got fixed to take a seed | 11:32 |
| paul__ | in php 5.2.x | 11:32 |
| dhx_m | yes but it's still not cryptographically secure | 11:32 |
| dhx_m | it is now seeded on each new user request to the server | 11:33 |
| dhx_m | (preventing problems with keep-alive connections) | 11:33 |
| paul__ | (preventing problems with keep-alive connections)? mm ? | 11:43 |
| dhx_m | I mean, newer versions of PHP reseed the MT PRNG upon each request made by the user | 11:45 |
| dhx_m | on keep-alive connections | 11:45 |
| dhx_m | ... I think? | 11:45 |
| paul__ | I'm just wondering if we're actually fixing a problem or not ;/ | 11:50 |
| dhx_m | 49 #ifdef PHP_WIN32 | 11:58 |
| dhx_m | 50 #define GENERATE_SEED() (((long) (time(0) * GetCurrentProcessId())) ^ ((long) (1000000.0 * php_combined_lcg(TSRMLS_C)))) | 11:58 |
| dhx_m | 51 #else | 11:58 |
| dhx_m | 52 #define GENERATE_SEED() (((long) (time(0) * getpid())) ^ ((long) (1000000.0 * php_combined_lcg(TSRMLS_C)))) | 11:58 |
| dhx_m | 53 #endif | 11:58 |
| dhx_m | from http://svn.php.net/viewvc/php/php-src/trunk/ext/standard/php_rand.h?view=markup | 11:58 |
| dhx_m | The time is fairly guessable by the attacker | 11:59 |
| dhx_m | the PID can be the same when PHP is run in a threaded mode | 11:59 |
| kirillka | hi all | 12:59 |
| kirillka | arouns? | 12:59 |
| kirillka | around? | 12:59 |
| kirillka | ping | 13:00 |
| ozgurkuru | hello | 13:05 |
| ozgurkuru | I have problem with utf8 chars | 13:06 |
| kirillka | ozgurkuru: what's problem? | 13:09 |
| kirillka | dhx_m: ping | 13:09 |
| kirillka | paul__: ping | 13:09 |
| dhx_m | killefiz: hi :) | 13:09 |
| dhx_m | kirillka: hi | 13:09 |
| dhx_m | wrong person :) | 13:09 |
| kirillka | dhx_m: :) | 13:09 |
| ozgurkuru | When I use şş ext I see ?? | 13:09 |
| ozgurkuru | I have problem with Turkish letters | 13:09 |
| kirillka | dhx_m: Daryn, how I can delete remote branch? | 13:09 |
| ozgurkuru | ş,ğ,İ,ı | 13:09 |
| kirillka | ozgurkuru: where problem do you have? in db, in html, letter or rss? | 13:10 |
| kirillka | dhx_m: David, sorry | 13:10 |
| dhx_m | kirillka: git push remotereponame :remotebranchname | 13:10 |
| dhx_m | note the colon in front of the remote branch name | 13:11 |
| ozgurkuru | kirillka, in db | 13:11 |
| dhx_m | that tells git to delete the branch instead of pushing it | 13:11 |
| kirillka | I push on server wrong branch | 13:11 |
| kirillka | I want delete this branch | 13:11 |
| dhx_m | yep just use the colon prefix to your branch name | 13:12 |
| kirillka | dhx_m: sorry, I don't understand | 13:12 |
| dhx_m | let's say your remote repository is named 'mforge' | 13:13 |
| dhx_m | so you'd normally push branches to it using: | 13:14 |
| dhx_m | git push mforge somebranch | 13:14 |
| dhx_m | now, to delete 'somebranch' from the remote mforge repository, you use: | 13:14 |
| dhx_m | git push mforge :somebranch | 13:14 |
| kirillka | dhx_m: Oh. I understand.. First part is empty. Thanks | 13:16 |
| dhx_m | :) | 13:16 |
| kirillka | dhx_m: One more thanks - I delete all wrong branchs | 13:20 |
| dhx_m | ok | 13:20 |
| kirillka | ozgurkuru: What db collation? | 13:21 |
| kirillka | ozgurkuru: is it utf8_general_ci? | 13:21 |
| kirillka | or utf8_turkish_ci? | 13:22 |
| ozgurkuru | kirillka, utf8_bin | 13:22 |
| ozgurkuru | but I find problem columns are latin1 now I'changing them | 13:23 |
| kirillka | ozgurkuru: you must change all to utf8, but I can wrong - I think you must convert to utf8_general_ci | 13:24 |
| kirillka | not utf8_bin | 13:24 |
| kirillka | dhx_m: I right? or wrong? | 13:24 |
| ozgurkuru | kirillka, oke Not important I use php script for this its easy.. | 13:25 |
| dhx_m | I think that's correct | 13:37 |
| ozgurkuru | kirillka, I solve that problem | 14:33 |
| ozgurkuru | thanks :D | 14:33 |
| kirillka | ozgurkuru: welcome | 14:34 |
| ozgurkuru | I wrote php script for that | 14:34 |
| ozgurkuru | I'm publish it in a short time | 14:34 |
| kirillka | daryn: mo | 14:40 |
| daryn | hello | 14:43 |
| paul__ | dhx_m: whats php_combined_lcg ? | 14:45 |
| dhx_m | paul__: http://en.wikipedia.org/wiki/Linear_congruential_generator | 14:46 |
| paul__ | http://svn.php.net/viewvc/php/php-src/trunk/ext/standard/lcg.c?revision=293253&view=markup | 14:55 |
| paul__ | mm | 14:56 |
| * paul__ asks php guys | 14:56 | |
| dhx_m | the current time is predictable | 15:02 |
| dhx_m | as is the thread ID/process ID | 15:02 |
| dhx_m | Mersenne Twister uses an LCG to create the initial seed | 15:03 |
| CIA-24 | Mantisbt: hickseydr * reb562360554d / (11 files in 3 dirs): Issue #10730: Use crypto_api for generating nonces and improve hashing | 15:04 |
| nuclear_eclipse | dhx_m: if we're going to go far enough to start adding salts or whatnot, why don't we just go whole hog and implement bcrypt? :P | 15:04 |
| nuclear_eclipse | there's a PHP extension for it, and it's the best way to store passwords in the db atm | 15:05 |
| dhx_m | nuclear_eclipse: well the Hash extension has been part of PHP since 5.1 :) | 15:05 |
| paul__ | dhx_m: regarding your last commit | 15:05 |
| dhx_m | nuclear_eclipse: I'm using it now (the Whirlpool algorithm in particular) for handling hashes | 15:05 |
| paul__ | dhx_m: what happens if someone has the rssfeed key stored in their client? | 15:06 |
| dhx_m | paul__: they'll need to update that after my change | 15:06 |
| dhx_m | paul__: ie. they just need to click the RSS icon and copy the URL into their client again | 15:06 |
| dhx_m | the advantage of these changes is users only need to keep one private key/secret: $g_crypto_master_salt | 15:07 |
| dhx_m | whereas most users before would have been unaware of the need to set those salts themselves | 15:07 |
| dhx_m | and now they can't even use Mantis without setting $g_crypto_master_salt :p | 15:08 |
| dhx_m | I think the next thing I need to do is let admins set their password at install time | 15:08 |
| dhx_m | rather than use 'root' as a default | 15:08 |
| dhx_m | I'll leave it for now though | 15:08 |
| paul__ | i've got a patch for that | 15:24 |
| paul__ | dhx_m: remember I added db_do_nothing as a schema update into trunk a few months back? ;p | 15:25 |
| dhx_m | paul__: heh yep :) | 15:27 |
| dhx_m | so... let's release 1.2.0? :) | 15:27 |
| paul__ | soon | 15:28 |
| paul__ | i'm thinking about this rss change atm | 15:28 |
| dhx_m | is it ready yet? is it ready yet? is it ready yet? :p | 15:28 |
| paul__ | how do we authenticate rss feeds atm? | 15:28 |
| dhx_m | via a key derived for each user account | 15:29 |
| dhx_m | so you pass &rss_key=something in the URL to the RSS feed | 15:29 |
| dhx_m | as well as your username | 15:29 |
| dhx_m | and it checks if you have them right | 15:29 |
| paul__ | mm | 15:30 |
| paul__ | I think we should change this :) | 15:30 |
| dhx_m | my change makes that more secure via using a 384bit key instead of a 128bit key (that is derived with a weak salt) | 15:30 |
| paul__ | in moodle | 15:30 |
| dhx_m | more to the point, we were using md5 before and that is anything but secure thesedays | 15:30 |
| paul__ | I think they go with a approach of | 15:31 |
| paul__ | for webservices | 15:31 |
| paul__ | allowing people to associate a logon with an account | 15:31 |
| paul__ | so i'm partly wondering if we should allow people to do this for webserviceread, webservicewrite, rssfeeds | 15:31 |
| paul__ | that way | 15:32 |
| paul__ | it would be possible to add functionality via aplugin or whatever to see who is accessing the rss feed and log | 15:32 |
| dhx_m | see bug 11219 | 15:32 |
| dhx_m | http://www.mantisbt.org/bugs/view.php?id=11219 | 15:32 |
| dhx_m | you could consider the current RSS keys as being automatically generated (weakly mind you) | 15:33 |
| dhx_m | with one key per user account whether they want it or not | 15:34 |
| paul__ | I see oauth as being a bit irrelevant as well thats no doubt one of many implementations | 15:34 |
| dhx_m | I guess the idea behind OAuth and more complex authentication is that you can create multiple logins for yourself, each with potentially lower permissions | 15:34 |
| dhx_m | so you can make single use accounts for yourself that are deleted after one login, and only have read-only RSS access to a particular project, for instance | 15:35 |
| dhx_m | OAuth is a layer on top of the required fundamental access control layer in Mantis though | 15:35 |
| dhx_m | and it seems very complex to me :( | 15:35 |
| paul__ | http://docs.moodle.org/en/Development:External_services_security | 15:36 |
| paul__ | look at the 'simple db' table and the 2 screenshots at bottom | 15:37 |
| dhx_m | yep I had the same idea in mind | 15:38 |
| dhx_m | but it'll be somewhat harder for us because those are toy examples | 15:39 |
| dhx_m | whereas with Mantis you may want much finer control over what external services are allowed to do with your account | 15:39 |
| nuclear_eclipse | dhx_m: I was still waiting on paul's yet-to-be-seen fixes to columns and ldap.... | 15:40 |
| nuclear_eclipse | paul__: stop getting distracted! | 15:40 |
| * paul__ hasn't been feeling well | 15:40 | |
| dhx_m | nuclear_eclipse: he's never going to commit it :p | 15:40 |
| paul__ | tbh, i've just not been in mood for doing stuff recently | 15:40 |
| paul__ | :( | 15:40 |
| paul__ | not really been in mood for doing some work stuff either | 15:41 |
| dhx_m | can those things wait until 1.2.1? | 15:41 |
| paul__ | the point of sending an email saying 1.2 something in next week | 15:41 |
| paul__ | was to motivate myself ;p | 15:41 |
| dhx_m | :) | 15:41 |
| paul__ | need to go find someone at work | 15:41 |
| paul__ | brb | 15:41 |
| dhx_m | anything I can help with? | 15:41 |
| dhx_m | ok | 15:41 |
| nuclear_eclipse | dhx_m: the reason I'm waiting is because he's looking at making API changes, and a) I hate making large API changes in a point release, and b) one of the changes I want in 1.2 he told me to hold off on because it would break his ability to merge onto latest... =\ | 15:43 |
| dhx_m | ah hmmm... API changes I'd agree with the need to wait | 15:43 |
| dhx_m | or move them to 1.3.x | 15:43 |
| webgambit | I'm trying to tie svn and mantis together, but they're on different servers. Anyone got a minute to lend a hand? | 16:33 |
| nuclear_eclipse | webgambit: what version of mantis are you using? | 16:36 |
| webgambit | it was the nightly build from last friday. | 16:37 |
| webgambit | what i've got is the svn is on a different server from mantis, so this isn't a mantis issue in and of itself. | 16:37 |
| nuclear_eclipse | webgambit: look at the source-integration plugins that I created for 1.2 | 16:38 |
| webgambit | ok. | 16:38 |
| nuclear_eclipse | they aren't yet compatible with 1.3 though | 16:38 |
| nuclear_eclipse | so I hope your nightly build is of 1.2 :P | 16:38 |
| nuclear_eclipse | http://git.mantisforge.org/w/source-integration.git | 16:39 |
| webgambit | i'll have to check. :) | 16:39 |
| webgambit | all I'm trying to do though is get the svn hook post_commit to make the call across servers to let mantis know something happened. | 16:39 |
| webgambit | nuclear_eclipse: I've downloaded and installed your source control plugins, but I'm thinking they're a bit overkill for what I was wanting to do. | 17:10 |
| mantisbt_80569 | hello | 23:42 |
Generated by irclog2html.py