| # | Time | Nick | Message |
|---|
| # | 01:06:14 | bshum | @roulette |
| # | 01:06:14 | pinesol_green | *BANG* Hey, who put a blank in here?! |
| # | 01:06:15 | pinesol_green reloads and spins the chambers. |
| # | 02:20:53 | bshum has quit IRC |
| # | 02:49:16 | bshum has joined #evergreen |
| # | 02:49:53 | bshum has quit IRC |
| # | 02:49:53 | bshum has joined #evergreen |
| # | 03:01:19 | AaronZ-PLS has joined #evergreen |
| # | 03:04:37 | AaronZ-PLS has quit IRC |
| # | 03:17:29 | mrpeters-isl has quit IRC |
| # | 07:37:56 | Aizya has joined #evergreen |
| # | 07:53:39 | Aizya has left #evergreen |
| # | 09:18:18 | natschil has joined #evergreen |
| # | 09:29:47 | natschil has quit IRC |
| # | 09:43:26 | natschil has joined #evergreen |
| # | 14:12:48 | bjwebb has joined #evergreen |
| # | 14:12:48 | bjwebb has joined #evergreen |
| # | 14:58:31 | dbs has joined #evergreen |
| # | 14:58:31 | dbs has joined #evergreen |
| # | 15:00:31 | dbs | installing git-core on testing.evergreen-ils.org / UPEI build slave so I can update the CI server |
| # | 15:01:20 | dbs | gmcharlt: I think we can point the bzr mirror on Launchpad to git as a source repo. it's pretty much required for i18n purposes |
| # | 15:01:41 | gmcharlt | dbs: ok |
| # | 15:02:48 | tsbere | gmcharlt: I had an idea for a repo or two. Gonna build the conf file and push to a secondary branch on the admin repo for you to review. |
| # | 15:03:02 | gmcharlt | okey-doke |
| # | 15:03:24 | dbs | tsbere++ |
| # | 15:03:26 | dbs | gmcharlt++ |
| # | 15:03:30 | dbs | awesome work guys |
| # | 15:07:18 | tsbere | gmcharlt: Pushed. |
| # | 15:09:15 | gmcharlt | tsbere: saw it; go for it |
| # | 15:09:59 | tsbere | Wasn't sure if you had other thoughts for expanding upon it. |
| # | 15:10:23 | tsbere | gmcharlt: Also, was thinking about pushing a single master branch to each (possibly by giving us access to do so) explaining what the "working" repos are. Thoughts on that? |
| # | 15:14:01 | tsbere knows that a lot of things dislike a repo not having a master (or other branch referenced by HEAD) |
| # | 15:17:54 | tsbere | Ok, I like the "don't bug me with stuff I already saw" feature of the emailing. |
| # | 15:19:09 | phasefx | we probably need to give artunit (poke poke) and friends a good poke; they may be oblivious to us wanting to change ILS-Contrib to git |
| # | 15:21:44 | lisppaste | tsbere pasted "Possible working repos "master" README" at http://paste.lisp.org/display/122019 |
| # | 15:22:08 | tsbere | gmcharlt: Any thoughts on that possible README? (or dbs, or phasefx, or anyone else around) |
| # | 15:23:16 | dbs | phasefx: if ILS-Contrib stops working for SVN, then the few of us actively using it will learn plenty fast :) |
| # | 15:23:29 | phasefx | dbs: :D |
| # | 15:26:04 | phasefx | tsbere: maybe give an example of how to push such a branch? |
| # | 15:29:02 | dbs | Hmm. So why would "git push origin dbs/git_for_ci_server:master" tell me "fatal: The remote end hung up unexpectedly"? |
| # | 15:29:11 | phasefx | so I'd do something like this? git remote add working-eg git@git.evergreen-ils.org:working/Evergreen.git git co -b 'user/phasefx' work work git push working-eg 'user/phasefx' ? |
| # | 15:29:18 | dbs | (attempting to push to OpenSRF master) |
| # | 15:30:04 | tsbere | dbs: Would need "git remote -v" output for origin to tell ya? |
| # | 15:30:04 | phasefx 's brain melts |
| # | 15:30:38 | tsbere | phasefx: branch names would be user/USER/feature type deal. Working on an updated README with examples, will annotate in a min ;) |
| # | 15:31:08 | lisppaste | dbs pasted "fatal: The remote end hung up unexpectedly (pushing to OpenSRF master)" at http://paste.lisp.org/display/122020 |
| # | 15:31:33 | phasefx | not used to thinking of branch names with slashes |
| # | 15:31:43 | tsbere | dbs: You cloned the git url. You need to push to ssh urls. You can fix that with git remote commands. |
| # | 15:32:28 | dbs | tsbere: ah, whee |
| # | 15:32:34 | tsbere | dbs: git remote set-url --push origin git@git.evergreen-ils.org/OpenSRF.git |
| # | 15:32:44 | tsbere | No, wait, that / needs to be a : |
| # | 15:32:55 | tsbere normally uses ssh:// full, and thus uses the / out of habit |
| # | 15:33:29 | dbs | success! yay |
| # | 15:33:32 | gmcharlt | tsbere: yes, let's give those trees masters, synced to the master branch of the corresponding main repos |
| # | 15:33:33 | phasefx | oh hey, you guys have trac using the git plugin! tsbere++ gmcharlt++ |
| # | 15:34:11 | gmcharlt | dbs: speaking of ILS-contrib, do you coniferites want ILS-contrib/conifer to become a separate repos? |
| # | 15:34:12 | dbs has a little more work to do on buildbot but is getting there |
| # | 15:34:23 | tsbere | gmcharlt: That.....is harder. Doable, but harder. |
| # | 15:35:02 | lisppaste | tsbere annotated #122019 "Longer working README" at http://paste.lisp.org/display/122019#1 |
| # | 15:35:09 | tsbere | phasefx: ^^ |
| # | 15:36:13 | phasefx | tsbere: rock. would those trees getting masters change anything? |
| # | 15:36:29 | dbs | gmcharlt: I think that makes sense |
| # | 15:36:47 | tsbere | phasefx: Well, I was originally thinking about putting that README as the only file in each's master branch. I guess instead that gets to go to the wiki or something. |
| # | 15:37:15 | tsbere | phasefx: Beyond that, not really. |
| # | 15:37:28 | tsbere | (beyond the setup work on my end, anyway) |
| # | 15:40:27 | phasefx | just pushed something to working, got |
| # | 15:40:28 | phasefx | remote: *** hooks.mailinglist is not set so no email will be sent |
| # | 15:40:29 | phasefx | remote: *** for refs/heads/user/phasefx/test update 0000000000000000000000000000000000000000->cce32dd86098980d1fc87523141f0adf958f9ce5 |
| # | 15:40:32 | phasefx | remote: sendmail: fatal: git(1002): No recipient addresses found in message header |
| # | 15:40:52 | tsbere | Yea....it does that. I should find a way to not have it yell when it isn't sending mail. |
| # | 15:41:03 | phasefx | it seemed to be pushing a lot.. I guess a master synced with the main one would require less stuff sent over the wire? |
| # | 15:41:13 | tsbere | yes, it would. |
| # | 15:41:18 | tsbere | But I haven't gotten there yet ;) |
| # | 15:41:30 | phasefx will try deleting it now :) |
| # | 15:43:03 | phasefx | so, if I just did git push working-eg user/phasefx/test without the : on the last argument, it'll just assume I mean the branch I currently have checked out? |
| # | 15:43:46 | tsbere | No, it will think you mean your local branch of user/phasefx/test (I think) |
| # | 15:44:08 | phasefx | ah |
| # | 15:44:33 | dbs | oh dear |
| # | 15:44:49 | phasefx | so far I've always named my local branches identically to remote branches |
| # | 15:44:55 | dbs | I think the buildbot's git-poller started polling at revision 1 |
| # | 15:46:59 | tsbere | phasefx: I think I fixed the "throwing errors when it shouldn't" issue |
| # | 15:47:40 | phasefx | sweet, thanks. not seeing the word "fatal" will prevent lots of fear/worry I bet |
| # | 15:47:58 | tsbere | phaefx: Feel free to test. ;) |
| # | 15:48:01 | tsbere | er |
| # | 15:48:03 | tsbere | phasefx |
| # | 15:48:09 | dbs | @later tell csharp mmm, can you get 'git-core' installed on the GPLS buildslaves, por favor? |
| # | 15:48:09 | pinesol_green | dbs: The operation succeeded. |
| # | 15:50:09 | phasefx | tsbere: yeap, confirmed |
| # | 15:50:49 | tsbere just hopes he didn't break the other mailing stuff in the process, but wants to wait until he has one of the "auto-mirror" scripts ready before testing happens |
| # | 15:53:23 | tsbere | ok... |
| # | 15:53:41 | tsbere | Anyone have anything to push to one of the two main repos, preferably the master branch thereof? |
| # | 15:54:20 | tsbere is tempted, BTW, to use this trick to have git.evergreen-ils.org push changes to git.mvlcstaff.org <_< |
| # | 15:56:09 | phasefx | tsbere: alright, pushed something to master |
| # | 15:57:37 | tsbere notes that something failed :( |
| # | 15:57:59 | tsbere | Good news: I didn't break the emailing. Bad news: The working repo wasn't updated. |
| # | 15:59:58 | phasefx | tsbere: oh, doh! I pushed to the main thing |
| # | 16:00:19 | phasefx | is that what you wanted? |
| # | 16:00:46 | tsbere | phasefx: I want it so that when someone pushes to Evergreen the master branch on working/Evergreen gets auto-updated. That didn't happen. |
| # | 16:00:55 | phasefx | ah, got you |
| # | 16:11:18 | tsbere might have fixed it |
| # | 16:11:38 | tsbere is tempted to rig this up in working/testing so he can push to testing and see if it works |
| # | 16:12:22 | tsbere | oooh, look, working was updated! |
| # | 16:13:14 | phasefx | tsbere++ |
| # | 16:15:02 | tsbere | gmcharlt: Do we want working repos for anything else right now, while I am in here playing with hooks anyway? |
| # | 16:20:42 | tsbere | phasefx: Added the working repos stuff to dev:git on the dokuwiki. Seems as good a place as any to put it. |
| # | 16:21:11 | phasefx | tsbere: cool deal |
| # | 16:25:14 | tsbere assumes that pushing to the working repos will be much less painful now that they have a copy of master.....unless you are working off of an older rel_whatever branch |
| # | 16:28:53 | tsbere | and emails and updates continue to happen. :D |
| # | 16:29:37 | phasefx thinks commit hashes are ugly :) |
| # | 16:30:45 | phasefx | wonder if we could have the shortlog format for a commit be the subject header, or if that's even desirable |
| # | 16:30:59 | tsbere | Given that a given update can contain more than one commit? |
| # | 16:31:09 | phasefx | good point |
| # | 16:31:15 | tsbere | It sends one email per push, not one per commit. Thankfully. |
| # | 16:31:16 | dbs resigns to buildbot building every freaking revision since revision #1 (whatever the hash is) for each branch |
| # | 16:32:29 | phasefx | smokebot |
| # | 16:32:33 | tsbere | phasefx: Unfortunetly, the email subject format is hardcoded into the script, and I don't want to maintain a copy just to change it. |
| # | 16:32:59 | phasefx | tsbere: no worries; I'll adapt to the new world order |
| # | 16:34:53 | dbs goes to see if he can at least turn off the "[open-ils-commit]" subject prefix - I think that's a per-user maillist option? |
| # | 16:35:36 | tsbere | dbs: Tis a mailing list wide option. |
| # | 16:36:05 | tsbere | and until contrib is no longer on SVN I think it needs to stay around. :( |
| # | 16:36:10 | phasefx looks for an RSS feed to subscribe to |
| # | 16:36:50 | tsbere | phasefx: http://git.evergreen-ils.org/?p=Evergreen.git;a=summary contains an RSS header for easy subscribing, I believe |
| # | 16:37:17 | tsbere | Issue: I think it only does one branch. |
| # | 16:37:36 | gmcharlt | tsbere: working repo for SIPserver, perhaps |
| # | 16:38:25 | tsbere | gmcharlt: Which isn't around yet. If we rig up SIPServer, do we want to ask atz to retire the github version as the "main"? And do we want the full history version to be the first version we push? |
| # | 16:38:41 | dbs | +1 to full history version |
| # | 16:38:47 | gmcharlt | +1 to full history version |
| # | 16:38:58 | gmcharlt | +1 to talking to atz first |
| # | 16:39:39 | tsbere | I assumed any repo we set up would have the full history included. The talking to atz first, however, was also assumed. |
| # | 16:40:12 | phasefx | tsbere: looks like trac might be another option, but it's so slow it's probably pegging a CPU everytime I look at it |
| # | 16:40:40 | tsbere | Also, the way I have this rigged up, I can make the server auto-push to pretty much anywhere on commit, provided you install git@git.evergreen-ils.org's pubkey. |
| # | 16:40:54 | rangi | you could run your own gitorious .. it is agpl3 :) |
| # | 16:42:00 | phasefx | I like the gitweb rss so much better than the email, just not as instantaneous |
| # | 16:42:02 | dbs | rangi: that's why I use gitorious |
| # | 16:42:51 | rangi | dbs: ditto, im just looking for a reason to run up a copy on a server |
| # | 16:51:58 | tsbere | phasefx: Looks like trac has issues with, say, more than 500 commits on git... |
| # | 16:57:03 | gmcharlt | dbs: is contribs/Conifer.git fine as a repo corresponding to ILS-Contrib/conifer? |
| # | 16:58:18 | dbs | gmcharlt: works for me! |
| # | 16:58:41 | gmcharlt | dbs: ok, I've got a git-svn clone in the works that I'll push to init the repo |
| # | 16:59:01 | gmcharlt | I'll give you access, and you can lean on your compatriots in Canadian crime to send us their SSH keys |
| # | 17:00:46 | gmcharlt | dbs: what do you want for the owner and description text of that repo? |
| # | 17:01:03 | dbs | gmcharlt: oh, the demands! |
| # | 17:01:25 | tsbere | gmcharlt: You could skip the owner (repo = "description") for now to simplify the question ;) |
| # | 17:01:41 | dbs | Owner = "Conifer consortium of Ontario academic and special libraries (http://projectconifer.ca)"? |
| # | 17:01:56 | gmcharlt | tsbere: I prefer to make dbs agonize ;) |
| # | 17:02:09 | gmcharlt | dbs: thanks |
| # | 17:02:16 | dbs | Description = "Customizations and scripts pertinent to Project Conifer (http://projectconifer.ca)"? |
| # | 17:02:25 | tsbere | dbs/gmcharlt: How will gitweb handle that long an owner string? |
| # | 17:02:36 | dbs | tsbere: I have no idea |
| # | 17:02:51 | dbs | alternate owner = "dbs" |
| # | 17:02:59 | gmcharlt | tsbere: truncates on the repo listing, shows full on the repo's summary, IIRC |
| # | 17:03:03 | gmcharlt | we'll find out in a moment |
| # | 17:03:09 | dbs has no idea how gitweb works |
| # | 17:03:19 | tsbere | How about just owner = http://projectconifer.ca for a short one? ;) |
| # | 17:05:19 | tsbere | Hmmm. Truncates it fairly quickly, doesn't it. |
| # | 17:07:37 | gmcharlt | dbs: ok, contrib/Conifer.git seed, let me know how it looks |
| # | 17:34:53 | dbs | re CI server - I have commented out all of the schedulers while the GitPoller does its bizarre one-revision-at-a-time thing, and will restart the schedulers once that is done |
| # | 17:42:40 | tsbere | dbs: Is this going to be a problem when you add a new branch for it to look at? |
| # | 17:43:04 | dbs | tsbere: it shouldn't be, but I guess we'll find out when that happens |
| # | 17:43:34 | dbs | it's an easy enough workaround, in any case - and indications are that this problem is fixed in buildbot master |
| # | 17:46:34 | dbs | goody, only 18,000 evergreen commits to go |
| # | 21:52:42 | dbs has quit IRC |
April 10-13