Sunday, 2010-06-27

« Saturday, 2010-06-26 Index Monday, 2010-06-28 »
../irclogs/#mantishelp.2010-06-27.log
--- scribe started ---00:00
ToffeePopsHi, we recently updated to mantis 1.2.1... I've just noticed that private issues no longer appear on the main 'view issues' screen with a padlock next to them.00:22
ToffeePopsThey do appear with a padlock under 'my view', but on the main list it's very useful to see that padlock icon - is there any way to turn that on?00:22
paulrdhx_m: moo10:40
dhx_mpaulr: howdy10:40
paulrwow :P10:40
dhx_mpaulr: hah :)10:40
paulrit works!10:40
dhx_mindeed :)10:40
dhx_mI'm currently without a working web server at the moment I'm afraid (no testing possible for now)10:41
paulrhehe10:41
dhx_min the middle of upgrading to PHP 5.3.3 with FPM :)10:41
paulrfpm = cgi right?10:41
dhx_myep they've added their own fast CGI manager10:41
paulrnod10:42
paulrthat's an RC?10:42
dhx_myep10:42
dhx_mI was using PHP 5.2 before so it's finally time to bother upgrading :)10:43
dhx_m(5.3 is not yet supported by Gentoo)10:43
dhx_mthe only thing I'm unsure about is the state of PHP accelerators10:45
dhx_mthey all seem to be dying projects10:46
dhx_m(or at least very slow)10:46
paulrthey having a debate whether to put APC in trunk10:53
paulrso do you have time again now for mantis?10:56
dhx_myep10:59
dhx_min the middle of ripping all JavaScript out of MantisBT and converting it to jQuery in an external file10:59
paulrcan I distract you briefly10:59
dhx_mthat way we can get CSP (content security policy) working :)11:00
dhx_msure11:00
paulryou broke pgsql/mssql :P11:01
paulrI think we should get away from adodb for schema updates11:01
dhx_mpgsql was working fine for me...?11:02
paulrfrom an existing build?11:06
paulrwith data in?11:06
paulrbut yes11:09
paulryou basically break upgdates11:09
paulrerm upgrades11:09
paulralbeit, I see a bug in my old code11:10
dhx_mhmmm which patchset broke it then?11:10
paulryou dropped my datadict additions remember? ;p11:11
paulrthereby, breaking what they fixed ;p11:11
paulranyway, given we only suppor what db's11:11
paulrmssql, pgsql, db2, mysqli?11:11
dhx_mhttp://git.mantisbt.org/?p=mantisbt.git;a=commitdiff;h=b8b21142be9e05232afada2e601f6c84ad0cd396#patch3211:15
paulryes11:15
paulrspot the bug in it11:15
paulrbut yes11:15
dhx_mif ( $existing[strtoupper($colname)]->type == 'bool' && $t == 'INTEGER') {11:15
dhx_mthat part?11:15
paulrSING (redirect_delay::BOOL)::INT ';11:15
paulrno ;p11:15
paulrthat should be colname ;p11:15
paulrbut basically11:15
paulrour upgdate fails in postgres11:15
dhx_mthe code is so crap I may as well be reading a hex dump11:16
dhx_m:p11:16
paulrif someone has data in redirect_dely column11:16
paulrwhich was a int/bool and converts11:16
paulras postgres can't convert bool<>int directly without a using clause11:16
paulrfor example ;p11:16
paulr-                       if (preg_match('/^([^ ]+) .*DEFAULT (\'[^\']+\'|\"[^\"]+\"|[^ ]+)/',$v,$matches)) {11:16
paulr+                       if (preg_match('/^([^ ]+) .*DEFAULT ([^ ]+)/',$v,$matches)) {11:16
paulrnot quite sure what that is11:16
dhx_mwell the problem has been that we don't know what has and hasn't been pushed upstream11:17
paulrthe problem is with adodb11:17
paulrupstream is so inconsistent11:17
paulryou relaly need to diff what they do ;p11:17
dhx_mso can we scrap it? :p11:18
paulrI've reported a 'bug' to upstream before that they've broken it completely by adding an 'a' randomnly to a php include file11:18
paulrso they've had to do a new release ;p11:18
dhx_mI'm in the mood for chopping off our dependencies (Projax waved goodbye a few days ago at long last :p)11:18
paulryes11:18
paulranyway, i'll work out what to do re adodb11:18
dhx_mlol11:18
paulrstill like the moodle db layer idea11:19
paulrit's either that or11:19
paulrsplit adodb and datadict11:19
paulrand just do our own11:19
paulrbut more right now11:20
paulrwant to fix up mantisforge scripts a bit11:20
dhx_mthe test framework?11:21
paulrnot necessarily11:21
paulrmore things like the git hashes11:21
dhx_mah yeah11:22
paulrhow can i do a by-char diff ;/11:29
dhx_m?11:30
paulrtrying to compare one long line11:30
dhx_mnot too sure11:31
paulrbasically, want to get rid of javascript/min11:32
dhx_mplease do :)11:33
paulrwe may as well just do that as part of release aka: http://code.google.com/p/closure-compiler/downloads/list11:33
paulrit's a version of that that jquery pass files through anyway11:33
paulrso until we hit problems11:34
paulrmay as well just trust google :)11:34
dhx_mdo we even need /min anymore?11:34
paulr'yes'11:34
paulrto compile our jscript files ;p11:34
dhx_mI'm not sure why there is a need to do that if we're going to be using 100% jQuery soon11:35
paulrwell for example the calendar11:35
dhx_mour JavaScript files are small and won't be requested by the client too often (E-TAG headers, etc)11:35
dhx_mthe calendar is now jQuery11:35
paulrmay as well just build them11:35
paulrlike jquery do with the same compiler11:36
paulrmm11:36
paulrI thought that jquery calendar didn't do time?11:36
paulronly date?11:36
dhx_mwhere do we use a calendar with time?11:36
paulrif you have timetracking on11:36
dhx_mcheck with daryn as he replaced the calendar :)11:36
paulrwhen was this?11:37
paulronly reason I know this was11:37
paulrat one point I looked to see if we could replace it11:38
paulrand concluded that I thought the jqeury calendar wouldn't let us set a time11:38
dhx_mhttp://git.mantisbt.org/?p=mantisbt.git;a=commit;h=1819bbdf8c2d629798fa48537f9bb167e8d3300511:40
dhx_mah he didn't convert it... just added jQuery UI for converting in the future11:40
paulrright11:40
paulrI'm pretty sure I concluded before11:40
paulrour current calendar needs a time option11:40
paulrand that jquery's just has a datepicker for dates only11:41
dhx_myep11:42
dhx_munless it's changed recently I think that's still the problem11:42
paulranyway11:44
paulrI want to try to make the 'min' directory just get built11:44
paulrwhich should be a simple a case of a script11:44
paulralso11:45
paulrpeople need to engage brain more :)11:45
dhx_mwhich compressor will you use?11:45
paulrsame one as jquery use atm11:46
paulri.e. google's one11:46
paulralbeit whatever latest version equates to11:46
dhx_msounds good to me :)11:47
dhx_mno need to recompress jQuery though11:47
paulrno but it's easier11:48
ToffeePopsHi, we recently updated to mantis 1.2.1... I've just noticed that private issues no longer appear on the main 'view issues' screen with a padlock next to them.11:51
ToffeePopsThey do appear with a padlock under 'my view', but on the main list it's very useful to see that padlock icon - is there any way to turn that on?11:51
dhx_mhang on, I'll try and test it when I have my web server back up11:52
paulrdhx_m: we need a script to build those hashes12:02
paulryou did it semi-manually before right?12:02
dhx_myep12:02
dhx_mdid you get the commands I used last time?12:02
paulryes12:03
paulrbut remind me the output12:03
dhx_mgood (because I might have lost them) :p12:03
paulrdunno if I kept them though ;p12:03
dhx_mI'll have a look in a moment12:13
paulrI think I worked out we can do it via php12:13
dhx_mgrr Linux 2.6.35-rc3 has a networking bug that is annoying me12:15
paulrheh12:25
paulrwell12:26
paulrit's messy12:26
paulrbut works :P12:26
paulrwindow.Sizzle = Sizzle;12:27
paulrhmm :P12:27
paulrdev/jquery.js:3686: WARNING - unreachable code12:27
dhx_manother compressor that breaks code? :)12:33
paulrheh12:33
paulrdunno :)12:33
paulrdoubt it12:33
paulrMINJAR = java -jar ${BUILD_DIR}/google-compiler-20091218.jar12:34
paulr@@${MINJAR} --js ${JQ} --warning_level QUIET >> ${JQ_MIN}12:35
paulrright so they a) hide warnings12:35
paulrb) run a 20091218 version which is out of date now12:35
dhx_mheh12:35
paulr   compiler-20091217.zip Zip file for closure-compiler Dec 17 3.7 MB 26912:36
paulr   compiler-20091217.tar.gz Tarball for closure-compiler Dec 17 3.7 MB 20312:36
paulrmm12:36
paulrgoogle don't list a 20091218 release though ;/12:36
paulrwe could get same version from http://github.com/jquery/jquery/tree/master/build/12:38
dhx_mdo they have a repository for Closure?12:39
paulrgoogle?12:39
paulrduno :)12:39
dhx_myep12:39
paulranyway, we can get the 20091218 compile from jquery's12:40
dhx_mhttp://closure-compiler.googlecode.com/files/compiler-latest.zip12:40
paulror just use current one and if we hit problems change12:40
paulrnod12:40
dhx_mfiles in that .jar were touched on June 16th12:40
paulrnod12:40
paulrthats' what i'm running atm12:40
dhx_mperhaps the script could download the latest Closure Compiler?12:40
paulrmm12:41
paulrthat might be more annoying12:41
paulrbut yea12:41
paulreither way12:41
paulrwe can build 'javascript/min'12:41
paulrfor release tarballs12:41
paulr(gives us an additional way to tell difference between release/svn users ;p12:42
paulrso next thing12:42
paulrhashes12:42
paulrfrom git :P12:42
dhx_mchecking12:45
dhx_m#git rev-list --all | git diff-tree --stdin -r -c -M -C | awk 'BEGIN{commit=""}{if(index($0,":")==0){commit=$1;}else{print "\t" "'"'"'" $4 "'"'"'" " => " "'"'"'" commit "'"'"'" ","}}' > integrity.php12:47
dhx_m#git tag -l release* | git cat-file --batch | grep "^object " | cut -d ' ' -f 2 | git cat-file --batch | grep "^tree " | cut -d ' ' -f 212:47
dhx_m#git ls-tree -r {tree-object-hash} | grep " blob " | awk '{print "\t\t" "'"'"'" $4 "'"'"'" " => " "'"'"'" $3 "'"'"'" ","}'12:47
dhx_msomething like those12:47
dhx_malso have this one:12:47
dhx_m#git rev-list --all | git diff-tree --stdin -r | awk 'BEGIN{commit=""}{if(index($0,":")==0){commit=$1;}else{print $6,$4,commit}}' | sort -s -k1,1 | awk 'BEGIN{file="."}{if(index($0,file)!=1){file=$1;print "\t);\n\t" "'"'"'" $1 "'"'"'" " => array(";}if($2!="0000000000000000000000000000000000000000"){print "\t\t" "'"'"'" $2 "'"'"'" " => " "'"'"'" $3 "'"'"'" ",";}}'12:47
dhx_mit's been so long since I wrote those that I don't know what does what :p12:47
paulrGetting objects/e5/0944ce028886db539855a0d7d0acbf5dc3a94b12:53
paulrwhats that commit12:53
dhx_m?12:53
paulrGetting objects/e5/0944ce028886db539855a0d7d0acbf5dc3a94b12:55
paulrbool(false)12:55
paulrf6dea35cf03319fe1f15ada084eaec417cc64cb112:56
paulrhttp://git.mantisbt.org/?p=mantisbt.git;a=commit;h=f6dea35cf03319fe1f15ada084eaec417cc64cb112:56
paulrhmm12:56
paulr403 Forbidden - Unknown commit object12:57
dhx_mwhat command did you run?12:57
dhx_mgit rev-list --all | git diff-tree --stdin -r -c -M -C | awk 'BEGIN{commit=""}{if(index($0,":")==0){commit=$1;}else{print "\t" "'"'"'" $4 "'"'"'" " => " "'"'"'" commit "'"'"'" ","}}'12:57
dhx_m?12:57
paulri'm running a php version12:58
paulrGetting branch [a6a8ed17748a91e019239beff77869c94198a232] refs/tags/release-1.2.1^{}12:58
paulrGetting objects/a6/a8ed17748a91e019239beff77869c94198a23212:58
paulrGetting branch [cd27a2d26c9b57c403c03ffd5a8d5b3f73594cbc] refs/tags/renamed-libraries-1.2.x12:58
paulrGetting objects/e5/0944ce028886db539855a0d7d0acbf5dc3a94b12:58
paulrf6dea35cf03319fe1f15ada084eaec417cc64cb1,12:58
dhx_mPHP git library?12:58
paulryes12:58
dhx_mugh that'll be slow?12:58
paulrseems to be fairly quick12:58
dhx_mfor listing all commit hashes ever?12:59
dhx_msorry... blob IDs :)12:59
paulrwhere was your output files?13:01
paulrso remind me13:05
paulrrelease blobs is files by brnahcn/tag ?13:05
dhx_mgit tracks files by an "object hash"13:05
dhx_mevery time a file changes in a commit, git creates a new object hash to represent the new file13:06
dhx_mrelease_blobs = keeps the object hashes for each file in the past MantisBT releases (we can choose how many releases to cater for)13:07
dhx_mcommit_blobs = keeps object hashes from *every* commit in the git tree13:08
dhx_mgoing back to the start of the repository13:09
dhx_m~4MB file13:09
paulri'm hoping I can generate it in php13:09
dhx_mit's fast to lookup though :)13:09
paulrthen we can just build array etc easily13:09
dhx_myep13:09
dhx_mI think for official releases we can just use release_blobs with the last two years worth of releases13:10
dhx_mthat way people can easily see if they're using old versions of MantisBT files that are no longer needed13:10
dhx_mif the object hash isn't in release_blobs then all they know is they've got an unknown file that doesn't trace back to any MantisBT commits13:11
dhx_m(any MantisBT releases in the last X releases)13:11
paulrwe tag releases right?13:20
nuclear_eclipseyep13:20
nuclear_eclipsewe have release tags going all the way back to 0.X.X releases13:20
nuclear_eclipse(I think)13:21
nuclear_eclipseok, nvm, just back to 1.0.013:21
paulrwe pgp sign releases?13:22
nuclear_eclipseI have been13:22
paulrcommirs13:22
nuclear_eclipsehttp://git.mantisbt.org/?p=mantisbt.git;a=tag;h=1eaa1d6ff23c107811dffb2a6f386e8dd57f036513:22
paulrseems quite a short sig13:23
nuclear_eclipsewell, it's just signing the git tag data13:24
paulris it possible to delete tags permanently from git?13:24
nuclear_eclipseyes13:24
paulraka do we need renamed-libraries-1.2.x13:25
nuclear_eclipseno13:25
nuclear_eclipseI've been wanting to drop that for a while13:25
nuclear_eclipse(deleted)13:27
dhx_m:)13:28
nuclear_eclipseyou just need to delete the tag locally now, `git tag -d renamed-libraries-1.2.x`13:28
dhx_mdo we need the 'rpxnow' repo?13:28
paulryes13:29
paulrwe should keep that for now at least imo13:29
dhx_mwhat is it?13:29
paulrwas victor's attempt at adding stuff13:29
paulrwhat is 2982c250eca8503408f1a1d8371382aadd515c7313:29
nuclear_eclipse1.0.013:30
paulrmm13:39
paulrwould I also have to delete the tags from mantisforge seperately?13:39
nuclear_eclipseprobably13:46
paulrwe dont have a http:// url for git on git.mantisbt.org do we?13:46
nuclear_eclipseit's just a matter of `git push <remote> :renamed-libraries-1.2.x`13:46
paulrwhat I could probably do is 'rm' the dir and let it repull13:47
paulras long as that wont break a fork13:47
paulr5691113:55
paulrdhx_m: was you using a package?21:10
|HaruAFK|mh, is there a shorter way than this?:21:30
|HaruAFK|$projects = explode(",", plugin_config_get('project_id',array()));21:30
|HaruAFK|if ($projects != null) {21:30
|HaruAFK|foreach ($projects as $key => $project) {21:30
|HaruAFK|$projects[$key] = (int)$project;21:30
|HaruAFK|}21:30
|HaruAFK|}21:30
|HaruAFK|print_project_option_list($projects);21:30
« Saturday, 2010-06-26 Index Monday, 2010-06-28 »

Generated by irclog2html.py