2013-01-20 GnuCash IRC logs
00:34:48 *** yawar has quit IRC
00:40:08 *** arrainey has joined #gnucash
00:50:10 *** arrainey has quit IRC
01:00:20 *** kpreid has quit IRC
01:01:02 *** kpreid has joined #gnucash
01:48:30 *** john has joined #gnucash
01:48:30 *** gncbot sets mode: +o john
01:53:27 *** IanL has joined #gnucash
02:18:08 *** john has quit IRC
02:59:21 *** calbasi has joined #gnucash
03:06:41 *** nafg has quit IRC
03:28:31 *** calbasi has left #gnucash
04:37:55 *** Topcat has joined #gnucash
04:47:59 *** Topcat has quit IRC
04:50:29 *** Topcat has joined #gnucash
04:53:20 *** fabior has joined #gnucash
04:58:31 *** Topcat has quit IRC
05:22:54 *** ecocode` has joined #gnucash
07:09:12 *** fell has quit IRC
07:18:42 *** IanL has quit IRC
07:32:08 *** IanL has joined #gnucash
07:36:29 *** Topcat has joined #gnucash
07:46:41 *** Topcat has quit IRC
07:47:04 *** Topcat has joined #gnucash
07:49:18 *** Topcat has quit IRC
07:49:58 *** fell has joined #gnucash
07:49:58 *** gncbot sets mode: +o fell
07:56:55 *** gjanssens has joined #gnucash
07:56:55 *** gncbot sets mode: +o gjanssens
07:57:48 <warlord> good morning gjanssens
07:58:35 <gjanssens> good morning warlord
07:58:35 <gncbot> gjanssens: Sent 10 hours and 11 minutes ago: <warlord> The wiki instructions appear to be slightly wrong (or confusing). I did the git svn init but then I needed to run the git-svn-mirror update, not git-update, in order to pull from svn. Anyways I'm pulling both gnucash and gnucash-docs right now. I know it took jralls 10 days, but I'm hoping it'll be faster locally.
07:58:36 <gncbot> gjanssens: Sent 10 hours and 10 minutes ago: <warlord> Anyways, we can continue on it more in the morning and see about doing the svn -> git/gitolite -> github...
07:59:02 <warlord> the git-svn-mirror "completed"... modulo the fact that there is no "origin" set.
07:59:17 <gjanssens> good
07:59:37 <gjanssens> So where are we now:
07:59:44 <warlord> E.g.:
07:59:45 <warlord> Checking out files: 100% (2510/2510), done.
07:59:45 <warlord> Checked out HEAD:
07:59:45 <warlord> file:///repo/gnucash/trunk r22717
07:59:45 <warlord> Counting objects: 163221, done.
07:59:45 <warlord> Delta compression using up to 2 threads.
07:59:47 <warlord> Compressing objects: 100% (41850/41850), done.
07:59:49 <warlord> Writing objects: 100% (163221/163221), done.
07:59:51 <warlord> Total 163221 (delta 135431), reused 147369 (delta 120718)
07:59:53 <warlord> No origin remote at /root/git-svn-mirror line 272.
08:00:00 <warlord> (and the same for gnucash-docs)
08:00:14 <gjanssens> hmm
08:00:37 <gjanssens> That would already be a question for jralls
08:01:07 <warlord> I wouldn't call it a question per-se. The git-svn checkouts do not have any other 'git origin'.. So nowhere to 'push' to.
08:02:30 <warlord> I suppose I need to point the git-svn repo at the gitolite repo?
08:02:44 <gjanssens> Yes, I would think so
08:05:26 <gjanssens> The wiki tells to point at github, right ?
08:05:39 <warlord> Right.. But I want to point to gitolite.
08:05:51 <warlord> (although the wiki instructions didn't work as-is)
08:05:55 * gjanssens is still opening various windows to follow along
08:06:01 <warlord> Okay.
08:07:12 <gjanssens> What part of the wiki instructions wasn't working ?
08:07:35 <gjanssens> In any case, we do want to have our gitolite repo as origin
08:07:38 <gjanssens> Not github
08:08:08 <warlord> When I did 'git svn init'... I couldn't run git-update.
08:08:43 <gjanssens> When I tried it a couple of days back, I figured git-update would not be necessary
08:08:52 <gjanssens> Which is indeed a bit confusing in the wiki
08:09:30 <gjanssens> I think the part about "following the contributor's set up" was mostly meant to run svn init
08:10:05 <gjanssens> git-update by the way is a separate script to download
08:10:10 <gjanssens> It's not part of git
08:10:19 <warlord> Yes, I know.
08:10:24 <gjanssens> Ok
08:10:31 <warlord> BUt when I did run it against my git-svn init repo, it didn't work.
08:11:10 * gjanssens has installed git-update somewhere in his path, which means he runs git update instead of git-update
08:11:21 <gjanssens> Just a minor convenience
08:11:54 <gjanssens> Where did you clone the git repo from ? Github ?
08:12:17 <warlord> I didnt.
08:12:22 <warlord> I never did a 'git clone'.
08:12:28 <warlord> I just did 'git svn init...'
08:12:34 <gjanssens> Ah, then that's the problem
08:12:38 <warlord> Must I have run a 'git clone' first?
08:12:54 <warlord> Or can I do a 'git clone' over an existing 'git svn init'?
08:12:58 <gjanssens> The committer's set up starts with the non-committer's set up, ie clone the git repo
08:13:07 <gjanssens> clone first, then git init
08:13:39 <warlord> Oh... Damn.
08:13:40 <warlord> Okay...
08:13:50 <gjanssens> But given the origin story, I think we should take one more step first
08:14:03 <warlord> I don't mind starting over..
08:14:14 <gjanssens> If we want the gitolite repo to be the origin for the git-svn repo
08:14:20 <warlord> The origin issue is probably because I didn't
08:14:22 <warlord> 'git clone'
08:14:25 <gjanssens> It's easiest to clone from the gitolit repo
08:14:28 <gjanssens> indeed
08:14:39 <gjanssens> So we should first clone for gitolite
08:14:49 <warlord> Right.... But then how do we get the gitoline repo to push to github? I guess we add a remote?
08:14:54 <gjanssens> from github (which should be the origin for the gitolite repos
08:15:02 <gjanssens> )
08:15:12 <warlord> So how do we set that up?
08:15:23 <gjanssens> Become the git user
08:15:36 <warlord> And does that have to happen first?
08:15:41 <gjanssens> cd into the directory where all the gitolite repos are
08:16:26 <gjanssens> run: git clone --bare git@github.com:Gnucash/gnucash.git
08:16:40 <gjanssens> And the (I believe) gitolite-setup
08:16:47 <gjanssens> All that won't take long
08:17:05 <gjanssens> You can also change origins afterwards
08:17:13 <warlord> But then how do I push w/o authentication? Or how do I set up auth to push to github?
08:17:16 <gjanssens> It just seems easier to let git take care of it
08:17:50 <warlord> Well, gitolite seems to want to create the repo..
08:18:03 <gjanssens> You will have to configure ssh for user git to always connect to github.com using a specific ssh key
08:18:25 <gjanssens> The public key of the pair should be uploaded your github account
08:20:53 <warlord> I wonder if I should create a gnucash-specific github account?
08:22:40 <gjanssens> I don't think so
08:22:59 <gjanssens> git is rather flexible in that regard
08:23:16 <gjanssens> It doesn't really matter who pushes changes
08:23:29 <gjanssens> Each commit has a committer and an author
08:23:49 <gjanssens> Those don't change when pushing commits between repositories
08:24:02 <gjanssens> You only need to have permission to do so
08:24:03 <warlord> Right. But it does mean that anyone with access to the git account could theoretically access anything I can on github.
08:24:41 <gjanssens> That's possible yes
08:24:53 <gjanssens> Likely even
08:25:15 <gjanssens> From that point of view it may be useful to have a gnucash-only account
08:25:20 <warlord> .. which is why it might be "better" to have a special account.
08:25:57 <gjanssens> You can then indeed limit that account to only access gnucash related stuf on github
08:26:18 <gjanssens> Keeping your derekatkins account more free for potentially other repos on github
08:29:05 <warlord> Okay, just created a code-gnucash-org github account.
08:30:44 <gjanssens> Good. I have added it to the GnuCash Developers team on github
08:31:01 <gjanssens> which means this user now has write access to the 4 repos
08:31:30 <gjanssens> Next is to set up an ssh keypair for this user and upload the public key to github
09:04:11 <warlord> Okay
09:09:32 <warlord> Keys generated. Working on adding them to github..
09:10:13 <gjanssens> Ok, that's in the account profile page
09:10:31 <warlord> yeah..
09:12:40 <warlord> ok. done,
09:13:01 <warlord> Next.. How do I add an 'origin' to an existing git repo?
09:13:47 <warlord> Do I just add it manually to the config file?
09:14:04 <gjanssens> That will certainly work
09:14:05 <warlord> or is there more Ihave to do?
09:14:13 <gjanssens> No
09:14:29 <gjanssens> There probably is a git command that does the same, but I don't know it
09:14:41 * warlord is trying to understand what 'gitolite mirror' does.
09:15:42 <gjanssens> As far as I understood it installs a trigger to push changes to a remot git repo
09:16:00 <gjanssens> The caveat being that the remote repo should also be managed by gitolite
09:16:09 <warlord> Oh. Hmm
09:16:15 <warlord> Okay.. Maybe not quite what I want.
09:16:50 <warlord> I suppose once the gitolite repo exists I can just use git ... file:///path/to/repo ?
09:17:00 <gjanssens> I think we may want to install our own trigger that simply runs a git push --all origin
09:17:10 <gjanssens> Yes
09:17:24 <gjanssens> Er, maybe not
09:17:37 <gjanssens> It depends for what
09:17:41 <gjanssens> For gitolite to function, the repo should only be modified via gitolite
09:17:42 <gjanssens> The ssh setup ensures this to be the case
09:17:43 <warlord> Right.. but I don't know how to do the git-svn -> gitolite push
09:17:58 <gjanssens> By setting the proper "origin" in the git-svn repo
09:18:37 <gjanssens> You can probably use a construct like "git@localhost/path/to/repo"
09:19:24 <gjanssens> And then make sure the user that will run git-svn-mirror is configured with the proper ssh keys
09:19:37 <gjanssens> So that user's got write permission in the gitolite repos
09:20:12 <gjanssens> git-svn-mirror update always ends with a push to origin
09:20:15 <warlord> Hmm.. true...
09:20:57 <warlord> This means I can't do what I was planning to do, which is to have the svn post-commit hook just run the 'git-svn-mirror' script..
09:21:22 <gjanssens> svn post-commit is run by the wrong user ?
09:21:29 <warlord> I need an intermediary .. so I can set up ssh +a gitolite user
09:21:37 <warlord> svn post-commit is run by the svn user.
09:21:40 *** benoitg has joined #gnucash
09:22:03 <warlord> (and by 'svn' user, I mean each svn commiter has his/her own account, and the commit hook is run by them)
09:22:22 <warlord> so when you svn commit, the hook gets run by you.
09:22:25 <gjanssens> Ah right
09:22:28 <warlord> when I commit, it's run by me.
09:22:29 <warlord> etc.
09:23:13 <warlord> To change users I either need sudo, or I need something like cron .
09:24:32 <gjanssens> Or perhaps use the same trick gitolite uses
09:24:36 <warlord> What I could do is have the svn post-commit 'touch /tmp/git-svn', and then have the cron job test for that file.
09:24:51 <warlord> I'm not sure how to do that in this case...
09:25:03 <warlord> Note that there *is* a 'gitolite push' command.
09:25:24 <gjanssens> Thinking out loud here
09:25:54 <gjanssens> set up a special ssh key in the local ssh config file for user git
09:26:18 <gjanssens> Instead of forcing gitolite as command to run, forc git-svn-mirror as command
09:28:01 <gjanssens> Then in the svn post-commit hook, add an ssh -l <special-user> git@localhost bla
09:28:38 <warlord> But that means each svn user needs access to that private key
09:28:39 <gjanssens> That would of course still mean that this particular keypair must be available to all svn users somehow
09:28:54 <warlord> I think the cron job route is still easiest..
09:29:26 <gjanssens> But the cron job will cause synchronisation delays
09:29:46 <gjanssens> Which caused me head-aches when John initially set up his server
09:30:07 <gjanssens> How do you trigger his jeeves server ?
09:30:16 <gjanssens> Some kind of port-knock ?
09:30:45 <warlord> Yep
09:30:52 <gjanssens> Can that be used here ?
09:31:36 <gjanssens> Everything remains localhost, so you can even shield it from the outside world in your firewall
09:31:42 <warlord> Well I was thinking that the cron job would run every minute.. So at most there would be a 60-second delay
09:32:25 <gjanssens> We can certainly start like that to get going
09:32:48 <gjanssens> If it turns out to be annoying, we can still think of an alternative trigger
09:33:07 <warlord> Right.
09:33:18 <gjanssens> Did you already clone into gitolite ?
09:33:35 <warlord> not yet. I need to set up that remote "origin"
09:33:45 <gjanssens> No, git does that for you
09:34:29 <gjanssens> The repo you clone from will automatically be used as origin
09:35:46 <warlord> But remember, I didn't clone.
09:35:51 <warlord> I just did 'git svn init'
09:36:02 <gjanssens> You're thinking the wrong repo
09:36:22 <gjanssens> The git-svn repo should not be under gitolite control
09:36:27 <gjanssens> It can't be even
09:36:29 <warlord> It's not.
09:36:40 <gjanssens> Ok
09:36:41 <warlord> I need to push from the git-svn repo into the git remo
09:36:43 <warlord> repo
09:37:03 <warlord> that's what I'm working with right now..
09:37:25 <warlord> svn commit needs to hook into the git-svn repo, and then push that into gitolite.
09:37:35 <warlord> .. and then that gets pushed to github
09:37:47 <warlord> so I'm working on the svn -> git-svn -> gitolite right now.
09:37:55 <gjanssens> There's no problem setting it up in reverse order, which is actually easier
09:38:07 <warlord> ??
09:38:08 <gjanssens> First clone github into gitolite
09:38:21 <gjanssens> That will define the proper origin for gitolite
09:38:44 <warlord> Okay, how do I do that? there's already a gitolite bare repo..
09:38:45 <gjanssens> Then clone the gitolite repo to wherever you want to manage the git-svn stuff
09:39:13 <gjanssens> a gitolite bare repo for gnucash ?
09:39:17 <warlord> Yes
09:39:23 <warlord> gitolite created a bare git repo
09:39:43 <gjanssens> Ok, a couple of steps back
09:39:59 <gjanssens> What command did you use to create that repo ?
09:40:14 <warlord> I edited conf/gitoline.conf and ran 'git push'
09:40:20 <warlord> in the gitolite-admin repo
09:40:53 <gjanssens> Ah, I would have created it differently, see:
09:40:56 <gjanssens> http://sitaramc.github.com/gitolite/rare.html#existing
09:40:57 *** benoitg has quit IRC
09:42:06 <gjanssens> We're not starting with new repos, but trying to import existing ones
09:42:29 *** jmd has joined #gnucash
09:43:20 <warlord> 1 sec..
09:46:10 <gjanssens> But I think we can also continue on the empty repos
09:46:22 <gjanssens> What we're looking for is this:
09:46:28 <gjanssens> http://git-scm.com/book/ch2-5.html
09:47:37 <warlord> I already removed the bare repo..
09:47:51 <warlord> So ... I git clone from github..
09:47:57 <gjanssens> ok
09:48:13 <gjanssens> make sure to use --bare in the clone command
09:49:44 <warlord> Do I want that? Or do I want --mirror?
09:51:16 <gjanssens> Perhaps
09:51:30 <gjanssens> It's not clear from the man page in which direction the mirror would work
09:51:50 <gjanssens> At least not for me
09:52:24 * gjanssens is looking for some more info on the web
09:52:37 <warlord> hmm.. it might be a one-way mirror.. not necessarily a two-way mirror..
09:52:48 <warlord> Set up a mirror of the source repository. This implies --bare.
09:52:48 <warlord> Compared to --bare, --mirror not only maps local branches of the
09:52:48 <warlord> source to local branches of the target, it maps all refs (including
09:52:48 <warlord> remote-tracking branches, notes etc.) and sets up a refspec
09:52:49 <warlord> configuration such that all these refs are overwritten by a git
09:52:51 <warlord> remote update in the target repository.
09:53:18 <gjanssens> From http://stackoverflow.com/questions/3959924/whats-the-difference-between-git-clone-mirror-and-git-clone-bare
09:53:28 <gjanssens> I understand the mirror would be in the wrong direction
09:53:44 <gjanssens> The gitolite one would be "slave" of the github one
09:53:55 <gjanssens> Exactly the opposite of what we want
09:53:57 <warlord> ok, so just use --bare
09:54:02 <gjanssens> Yep
09:55:26 <warlord> okay, doing that..
09:55:47 <warlord> Do we only have 'gnucash' and 'gnucash-docs' repos on github?
09:56:30 <gjanssens> No, there's also gnucash-htdocs
09:56:36 <warlord> Ah, not just 'htdocs'?
09:56:37 <gjanssens> And the auxiliary git-helper-scripts repo
09:56:38 <gjanssens> No, see: https://github.com/Gnucash
09:56:42 <warlord> Should we keep that name, or rename it back to 'htdocs' on the gitolite server?
09:56:46 <gjanssens> I'd keep the name
09:56:54 <gjanssens> On gitolite on gnucash.org, the name is overkill
09:56:57 <warlord> right..
09:56:58 <gjanssens> But on github, it's one repo among thousands
09:56:59 <warlord> that's my point.
09:57:03 <gjanssens> Naming the repo just htdocs there is less evident
09:57:04 <warlord> We can keep it gnucash-htdocs on github, but why not have it be 'htdocs' on code.gnucash.org?
09:57:08 <gjanssens> Because that would be less consistent for people interested in both locations
09:57:12 <gjanssens> New contributors that first discover gnucash-htdocs on github
09:57:19 <gjanssens> I prefer the same name in both locations
09:57:20 <gjanssens> We also have gnucash-docs and not docs
09:57:23 <warlord> true, but it's always been 'gnucash-docs' in svn.
09:57:26 <warlord> this would be a change in the repo name from svn -> git
09:57:39 <warlord> I'll note that the 'git-helper-scripts' doesn't have 'gnucash' in the name..
09:57:42 <warlord> And on Github they are all under the GnuCash moniker.
09:57:49 <gjanssens> Agreed on git-helper-scripts, which is imo an unfortunate choice by John
09:57:51 <gjanssens> Regarding the Gnucash moniker, that's only on the central page
09:57:55 <gjanssens> Each clone on github is under the name of the user that cloned
09:57:57 <warlord> True...
09:58:01 <warlord> But do we really expect other github users to fork the repos on github?
09:58:01 <gjanssens> So you will also find a gjanssens/gnucash jrall/gnucash and so on
09:58:06 <gjanssens> The Gnucash repo is forked 16 times already on github
09:58:07 <gjanssens> gnucash-docs 3 times
09:58:09 <gjanssens> So yes, people do clone on github
09:58:10 <gjanssens> In fact, git is all about cloning
09:58:16 <gjanssens> (cloning or forking is the same for me here)
09:58:19 <warlord> How many times is htdocs forked? ;)
09:58:19 <gjanssens> Once :)
09:58:27 <warlord> right..
09:58:28 <gjanssens> But to retrace a bit: as I said I like consistency in the names
09:58:28 <warlord> Wow, this clone from github is taking a long time.
09:58:31 <warlord> Sure. I can always rename it later.
09:58:33 <gjanssens> Having gnucash in the names helps identify them as gnucash related wherever you are on github
09:58:35 <warlord> Sure, but this isn't *on* githuib.
09:58:37 <warlord> this is on code.gnucash.org
09:58:38 <warlord> I'
09:58:39 <warlord> m not prposing renaming them on github
09:58:47 <gjanssens> Yes, my answer there was I prefer consistency all over the place
09:58:48 <gjanssens> And yes, that's not consistent with svn
09:58:51 <warlord> Right...
09:58:53 <warlord> Me, I prefer brevity
09:58:56 <gjanssens> But svn will be out of the picture in the relatively short term
09:59:19 <warlord> sure..
10:00:25 <gjanssens> So if we want to change, now would be a good moment
10:00:25 <warlord> fair enough..
10:00:25 <warlord> I'll keep it consistent with github.. we can always adjust it.
10:00:26 <gjanssens> tx
10:00:28 <warlord> (before we go live)
10:00:41 <gjanssens> ok
10:00:50 <warlord> Anyways, the clones are happening now.. Let me set up a special git-svn user.
10:01:04 <gjanssens> Github is indeed very slow
10:01:32 * gjanssens was trying to look up the fork tree for gnucash-htdocs
10:01:43 <gjanssens> still waiting for it...
10:02:19 <gjanssens> Came up immediatly once I reloaded the page, oh well...
10:04:47 <warlord> Weird.
10:05:02 <warlord> Right now my system is "resolving deltas"
10:07:31 <gjanssens> Part of the cloning proces ?
10:08:09 <warlord> Yep.
10:08:17 <warlord> My system is currently under heavy load..
10:10:31 <warlord> Okay, so it sounds like my next steps are going to be to use the git-svn user to do a git clone, then git svn init, and then set up the svn post-commit hook to get the git-svn user to run the git-svn-mirror script..
10:10:45 <warlord> (git clone from gitolite)
10:11:51 <gjanssens> I thought you decided to use cron instead of the svn post-commit hook ?
10:12:10 <gjanssens> git clone from gitolite in any case is correct
10:12:50 <gjanssens> But if you can have the git-svn user run the git-svn-mirror script in post-commit, I'd definitely prefer that option
10:13:19 <warlord> Well, I don't know if I want to run the mirror every minute.. I was thinking having it check-to-run every minute.
10:13:42 <warlord> I was thinking more that the svn-user would signal to the git-svn user to run the script at the next minute.
10:14:01 <gjanssens> Oh, right.
10:14:02 <gjanssens> Ok
10:15:08 <warlord> Okay, I think I know what has to happen here. It seems "strange" to do it in this direction, but okay.
10:33:20 *** KaiForce has joined #gnucash
10:33:23 *** IanL has quit IRC
10:33:25 <gjanssens> Strange ?
10:33:25 <gjanssens> BTW the repos cloned for gitolite can now be added to gitolite already (not sure if you did that already)
10:34:07 *** fabior has quit IRC
10:34:23 <warlord> I cannot do that until they finish cloning. I'm still working on cloning 'gnucash'
10:34:30 <warlord> Still have 3 more after that.
10:34:51 <gjanssens> They're still cloning for gitolite !?
10:36:09 <warlord> Yeah. It's taking a LONG time..
10:36:09 <gjanssens> Wow
10:36:09 <warlord> I think the vm server is overloaded right now.
10:36:09 <warlord> top - 10:35:32 up 80 days, 20:15, 2 users, load average: 6.16, 6.92, 6.90
10:36:09 <warlord> (that's from the vm server)
10:36:09 *** pnema has joined #gnucash
10:36:09 <gjanssens> Is the vm on the same server as the windows build server ?
10:36:09 <warlord> I think it's due to my preclear_disk process running in the background.
10:36:12 <warlord> Yes
10:36:34 <gjanssens> Ok, that would make sense
10:36:41 <warlord> The build server and code are botha vm guests on the same vm host server
10:37:00 <gjanssens> What is preclear_disk actually ? Never heard of it, but sounds useful
10:37:42 <warlord> It's a script from the unRAID process that basically puts a disk through a bunch of read/write tasks to make sure the disk is "okay". I run that and I run 'badblocks' before I add a disk to my array.
10:38:12 <warlord> The whole process takes about a week of burn-in.
10:38:26 <gjanssens> Interesting. I'll remember that for the next time I have to replace a disk
10:39:03 <warlord> :)
10:41:36 <warlord> I think M might want me back soon...
10:41:51 <gjanssens> Yes, kind of figured that...
10:44:21 *** jmd has quit IRC
10:44:22 <gjanssens> How do we continue from here ?
10:44:22 <gjanssens> You know what to set up now and how ?
10:44:22 <gjanssens> I mean once the repos are cloned
10:44:22 <warlord> I think so, yes. I need to:
10:44:22 <warlord> a) create the git-svn user, generate an ssh key, and add that key to gitolite with RW access to the repos
10:44:22 <gjanssens> ok
10:44:47 <warlord> b) from the git-svn user, git clone from gitolite, and then git svn init in those repos
10:45:01 <gjanssens> yep
10:46:53 <warlord> c) setup the script to run git-svn-mirror as the git-svn user
10:46:53 <gjanssens> using ssh, not file://
10:46:53 <gjanssens> ok
10:46:53 <warlord> right -- the git clone is from git://git@localhost, the svn can be file:///
10:46:53 <gjanssens> indeed
10:47:04 <warlord> Then the only missing piece is the push to github from gitolite
10:47:04 <gjanssens> For which we need to add a hook in gitolite
10:47:05 <warlord> right...
10:48:46 <warlord> question -- will the git commits from my instance and the git commits from jralls' instance interfere?
10:48:46 <warlord> or will git-svn create equivalent commits that are "the same"?
10:48:46 <gjanssens> Yes, they probably will
10:48:46 <gjanssens> They are the same (git uses a hash calculated on the commit to determine equality)
10:48:46 <warlord> okay...
10:48:46 <gjanssens> but when John pushes first to github
10:48:46 <gjanssens> git on gitolite will refuse to push
10:48:51 <gjanssens> simply because the github repo has more recent activity
10:49:02 <gjanssens> It won't compare equality beforehand
10:49:17 *** jmd has joined #gnucash
10:49:36 <gjanssens> So once everything on code works fine, John will have to disable his push to github
10:49:53 <warlord> Right. We'll need to coordinate with john to migrate the push to github.. which means I'll need to perform a pull to sync up everything first, I think..
10:50:01 <gjanssens> more or less at the same time we enable it on code
10:50:32 <warlord> Yeah. I'll need to somehow get gitolite to pull from github, then get the git-svn to pull from gitolite..
10:51:00 <gjanssens> Yes, but the pull will result in a no-op, because you already received all commits via the svn detour
10:51:01 <warlord> Oh, that's true..
10:51:07 <warlord> I guess the push would only fail for the second attempt; john and I should have equivalent repos
10:51:27 <gjanssens> Yes
10:51:30 <warlord> okay..
10:51:53 <gjanssens> Regarding the push hook: we can manage the hooks in the gitolite-admin repo
10:52:04 <warlord> Right...
10:52:17 <warlord> And we can play with that later.
10:55:00 <gjanssens> Yes, that's not the first concern right now
10:55:00 <warlord> right.
10:55:00 <gjanssens> If we get the whole chain up until gitolite working fine, we're already well underway
10:55:00 <gjanssens> Adding the final hook will only be a small addition
10:55:00 <warlord> Yep. Okay, I think I know what needs to happen... even if I feel it's 'backwards' ;)
10:55:00 <gjanssens> Heheh, as long as it works :)
10:57:04 <warlord> True.. It's just backwards to how I would think to convert, starting with svn
10:58:01 <gjanssens> Yes, the twist is in git svn init, which should be run on a real up-to-date repo, not an empty one
11:03:54 <warlord> Right..
11:03:54 <warlord> Okay..
11:03:55 <warlord> That's what seemed "wrong" to me yesterday ;)
11:03:55 <warlord> On code:
11:03:55 <warlord> top - 10:58:47 up 80 days, 19:48, 2 users, load average: 229.44, 215.66, 191.8
11:03:55 <gjanssens> code is inside the vm, or the host ?
11:07:05 <warlord> it's a vm guest.
11:07:05 <gjanssens> Working hard I'd say...
11:07:05 <warlord> Yeah, probably because the vm host isn't responding to I/O fast enough
11:07:05 <warlord> ... because of the preclear
12:04:15 <gjanssens> yes. I gotNOTICE AUTH :*** Looking up your hostname...
12:06:10 *** gncbot has joined #gnucash
12:10:17 *** arrainey has joined #gnucash
12:15:20 *** arrainey has quit IRC
12:16:00 *** arrainey has joined #gnucash
12:21:03 *** arrainey has quit IRC
12:22:42 *** arrainey has joined #gnucash
12:22:42 *** Topcat has joined #gnucash
12:23:30 *** Topcat has quit IRC
12:34:05 <gjanssens> op gncbot
12:45:00 <gjanssens> @op gncbot
12:45:00 *** warlord sets mode: +o gncbot
12:45:00 <warlord> gjanssens: use: /mode +o gncbot
12:45:00 * gjanssens just figured it out
12:45:00 <gjanssens> But you beat me to it
12:45:00 <gjanssens> tx :)
12:45:00 <warlord> :)
12:45:00 <warlord> poor server...
12:45:00 <warlord> it's so tired..
12:45:39 *** gncbot has joined #gnucash
12:50:23 *** casimono has joined #gnucash
12:51:05 *** pnema has quit IRC
12:51:31 *** pnema has joined #gnucash
12:52:54 <casimono> can it be the wiki server down? I cannot open the wiki page ...
12:54:59 <warlord> casimono: yeah, the server is way overloaded right now and barely responding to requests.
12:55:48 <warlord> It may take another hour or two before it clears itself up. Part of the issue is also the vmhost being overloaded at the moment.
12:55:59 <warlord> But part of is it a migration being run on the gnucash server itself.
12:56:05 <casimono> ok. will try later again. thx for quick info
12:56:31 <warlord> you're welcome. sorry.
12:56:40 *** warlord sets mode: +o gncbot
14:32:14 *** casimono has quit IRC
14:32:15 *** casimono has joined #gnucash
14:32:15 *** casimono has quit IRC
14:32:15 *** casimono has joined #gnucash
14:32:48 *** gncbot has joined #gnucash
14:32:48 <gjanssens> So there must be some competition on the server for either cpu or io
14:33:08 <gjanssens> Given the raid work, I do think it's io
14:33:14 <warlord> yeah. the clone just finished.
14:33:20 <gjanssens> Good
14:33:21 *** warlord sets mode: +o gncbot
14:33:40 <gjanssens> Best to let it rest until after the raid work
14:33:44 <warlord> And now the load is coming down.
14:33:47 <warlord> Yep, that's my plan.
14:41:31 <gjanssens> Have a nice WE
14:41:31 <warlord> I know what needs to happen now, so I'll work on it slowly.
14:41:31 <warlord> Thanks, you too. High to K from us!
14:41:31 <gjanssens> Perfect
14:41:31 <gjanssens> Hi back to M !
14:44:00 *** arrainey has quit IRC
14:44:41 *** arrainey has joined #gnucash
14:54:53 *** arrainey has quit IRC
14:55:36 *** arrainey has joined #gnucash
15:21:27 *** arrainey1 has joined #gnucash
15:23:38 *** john has joined #gnucash
15:23:38 *** gncbot sets mode: +o john
15:25:52 *** arrainey has quit IRC
15:31:43 *** arrainey1 has quit IRC
15:32:22 *** arrainey has joined #gnucash
15:37:26 *** arrainey has quit IRC
15:38:00 *** arrainey has joined #gnucash
15:43:01 *** arrainey has quit IRC
15:43:47 *** arrainey has joined #gnucash
15:48:50 *** arrainey has quit IRC
15:49:36 *** arrainey has joined #gnucash
15:51:42 *** jmd has quit IRC
15:53:24 *** gjanssens has quit IRC
15:59:41 *** arrainey has quit IRC
16:00:16 *** arrainey has joined #gnucash
16:05:19 *** arrainey has quit IRC
16:05:54 *** arrainey has joined #gnucash
16:09:34 *** yawar has joined #gnucash
16:24:03 *** arrainey has quit IRC
16:24:36 *** arrainey has joined #gnucash
16:29:40 *** arrainey has quit IRC
16:30:16 *** arrainey has joined #gnucash
16:30:39 *** aqua___ has joined #gnucash
16:45:28 *** arrainey has quit IRC
16:46:02 *** arrainey has joined #gnucash
16:50:35 *** john has quit IRC
16:56:06 *** arrainey has quit IRC
16:56:42 *** arrainey has joined #gnucash
17:01:42 *** Jimraeh1 has left #gnucash
17:11:13 *** arrainey has quit IRC
17:11:47 *** arrainey has joined #gnucash
17:13:25 *** fabior has quit IRC
17:22:23 <yawar> anyone know why the custom report runner window accessed from Reports > Custom Reports is titled `GnuCash'?
17:23:23 *** aqua___ has quit IRC
17:41:47 *** pnema has joined #gnucash
17:42:31 *** arrainey has quit IRC
17:44:43 *** Jimraeh1 has joined #gnucash
17:47:55 *** arrainey has joined #gnucash
18:00:01 <yawar> hi folks. i'm working on documentation bug 633590 (https://bugzilla.gnome.org/show_bug.cgi?id=633590) and am publishing my progress to a personal github branch (https://github.com/yawaramin/gnucash-docs/commits/bug-633590-trunk)--would welcome feedback
18:00:43 <yawar> argh. github link is at https://github.com/yawaramin/gnucash-docs/commits/bug-633590-trunk
18:01:46 *** arrainey has quit IRC
18:42:12 *** yawar has quit IRC
18:45:20 *** notam2000 has joined #gnucash
18:58:20 *** pnema has quit IRC
19:20:31 *** fell_ has quit IRC
19:25:09 *** fell has joined #gnucash
19:25:09 *** gncbot sets mode: +o fell
19:44:29 *** fell has quit IRC
19:55:10 *** pnema has joined #gnucash
21:03:26 *** pnema has quit IRC
21:05:15 *** pnema has joined #gnucash
21:32:22 *** arrainey has joined #gnucash
21:42:32 *** arrainey has quit IRC
21:43:09 *** arrainey has joined #gnucash
21:48:22 *** arrainey has quit IRC
21:49:01 *** arrainey has joined #gnucash
21:54:08 *** arrainey has quit IRC
21:54:43 *** arrainey has joined #gnucash
21:59:47 *** arrainey has quit IRC
22:00:33 *** arrainey has joined #gnucash
22:05:38 *** arrainey has quit IRC
22:06:23 *** arrainey has joined #gnucash
22:21:37 *** arrainey has quit IRC
22:22:22 *** arrainey has joined #gnucash
22:27:51 *** nafg has joined #gnucash
22:38:41 *** arrainey has quit IRC
22:39:20 *** arrainey has joined #gnucash
22:39:51 *** notam2000 has quit IRC
22:54:57 *** arrainey has quit IRC
22:55:29 *** arrainey has joined #gnucash
23:10:42 *** arrainey has quit IRC
23:11:24 *** arrainey has joined #gnucash
23:28:19 *** pnema has quit IRC
23:47:07 *** arrainey has quit IRC
23:47:41 *** arrainey has joined #gnucash