Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-03-25

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:02 xavih_ joined #gluster-dev
00:03 awheeler joined #gluster-dev
00:03 _Bryan_ joined #gluster-dev
00:03 awheeler_ joined #gluster-dev
00:14 _Bryan_ joined #gluster-dev
00:16 sac_ joined #gluster-dev
00:17 yinyin joined #gluster-dev
00:20 kshlm|AF1 joined #gluster-dev
00:28 lkoranda joined #gluster-dev
00:41 JoeJulian joined #gluster-dev
00:42 52AAAPGAV joined #gluster-dev
00:44 awheeler_ joined #gluster-dev
00:44 awheeler joined #gluster-dev
00:48 nixpanic_ joined #gluster-dev
00:49 nixpanic_ joined #gluster-dev
00:52 jdarcy joined #gluster-dev
01:00 yinyin joined #gluster-dev
01:05 jules_ joined #gluster-dev
01:26 xavih joined #gluster-dev
02:13 jules_ joined #gluster-dev
02:33 aravindavk joined #gluster-dev
02:46 aravindavk joined #gluster-dev
02:47 inodb_ joined #gluster-dev
02:47 a2 joined #gluster-dev
02:57 joaquim__ joined #gluster-dev
03:13 aravindavk joined #gluster-dev
03:23 bala joined #gluster-dev
03:44 jclift joined #gluster-dev
03:44 bharata joined #gluster-dev
04:19 sgowda joined #gluster-dev
04:21 aravindavk joined #gluster-dev
04:23 sripathi joined #gluster-dev
04:27 vshankar joined #gluster-dev
04:34 yinyin joined #gluster-dev
04:48 lalatenduM joined #gluster-dev
04:54 yinyin joined #gluster-dev
05:28 yinyin joined #gluster-dev
05:33 rastar joined #gluster-dev
05:34 mohankumar joined #gluster-dev
05:41 sac__ joined #gluster-dev
05:46 bulde joined #gluster-dev
05:48 raghu joined #gluster-dev
05:55 H__ joined #gluster-dev
06:12 bala joined #gluster-dev
06:24 deepakcs joined #gluster-dev
06:41 rgustafs joined #gluster-dev
07:39 yinyin joined #gluster-dev
07:41 yinyin joined #gluster-dev
07:47 xavih joined #gluster-dev
08:27 sripathi joined #gluster-dev
08:34 bulde1 joined #gluster-dev
08:39 xavih_ joined #gluster-dev
08:40 lala_ joined #gluster-dev
08:44 raghu` joined #gluster-dev
08:44 sac__ joined #gluster-dev
08:46 bulde joined #gluster-dev
08:53 bharata joined #gluster-dev
09:03 rastar1 joined #gluster-dev
09:03 sripathi joined #gluster-dev
09:16 sgowda joined #gluster-dev
09:18 jules_ joined #gluster-dev
09:36 bulde joined #gluster-dev
09:39 sgowda joined #gluster-dev
09:42 sripathi joined #gluster-dev
10:01 sripathi joined #gluster-dev
10:08 nixpanic joined #gluster-dev
10:08 nixpanic joined #gluster-dev
10:14 yinyin joined #gluster-dev
10:26 mohankumar joined #gluster-dev
10:30 bharata joined #gluster-dev
10:30 sgowda joined #gluster-dev
10:34 rastar joined #gluster-dev
11:04 jdarcy joined #gluster-dev
11:16 bulde joined #gluster-dev
11:35 bulde joined #gluster-dev
11:39 sripathi joined #gluster-dev
11:57 mohankumar joined #gluster-dev
12:04 jdarcy joined #gluster-dev
12:13 bulde joined #gluster-dev
12:16 jdarcy joined #gluster-dev
12:24 mohankumar sgowda: hagarth: i posted BD thin xlator patches to Gluster Gerrit
12:24 mohankumar joined #gluster-dev
12:58 kkeithley It'd be really nice if we could solve the problem gerrit has with adding reviewers, e.g. I start to type jclift (Justin Clift) or ndevos (Niels de Vos), and up pops jclift@redhat.com or ndevos@redhat.com, and those are the names/accounts by which gerrit reports that they've reviewed patches.
12:59 kkeithley But when I try to add them, gerrit claims there is no such user!
12:59 vshankar joined #gluster-dev
14:08 portante joined #gluster-dev
14:16 wushudoin joined #gluster-dev
14:32 puebele joined #gluster-dev
15:06 jdarcy joined #gluster-dev
15:10 jdarcy joined #gluster-dev
15:30 kshlm|AFK joined #gluster-dev
15:30 kshlm|AFK joined #gluster-dev
15:45 jclift joined #gluster-dev
16:32 kr4d10 joined #gluster-dev
16:32 kr4d10 left #gluster-dev
16:37 kkeithley have I ever said how much I hate autoconf, automake, and libtool?
16:53 jclift kkeithley: Probably.  I'm with you on that one too.
16:53 jclift "They suck" is an understatement.
16:53 jclift Mostly when trying to diagnose WTF something isn't working.
16:53 jclift <insert rant here> :)
16:54 jclift kkeithley: If you need good assistance with something, try pinging Eric Blake (Red Hat guy) from the libvirt team.
16:54 kkeithley btw, I tried to add you as a reviewer on change 4720, but gerrit doesn't seem to know who jclift@redhat.com is
16:54 jclift kkeithley: Use the username of "jc" instead.
16:55 kkeithley I seem to have figured it out, but it took way too long
16:55 kkeithley ah
16:55 jclift kkeithley: It'll pop up a list of users, but don't be tempted to choose one.  Just do "jc" and enter type thing.
16:55 jclift kkeithley: It's a left over artifact of the Fedora OpenID system changing endpoint.  It's left more than one userid for me in the system.
16:56 jclift "jc" works. :)
16:57 jclift kkeithley: What was wrong with the pkg-config check on OSX?
16:58 jclift kkeithley: Asking because I've tested here on OSX 10.7 without probs.
16:58 kkeithley where do you get pkg-config from?
16:58 jclift kkeithley: So, I'd like to figure out what/why/etc.
16:58 jclift Homebrew.
16:58 jclift kkeithley: http://mxcl.github.com/homebrew/
16:59 jclift kkeithley: It's one of the packaging systems for OSX.  *Extremely* popular with the OSS crowd.
16:59 jclift kkeithley: We have libvirt and other things in there already, etc.
17:02 jclift kkeithley: As an FYI, there are two other "legacy" packaging systems for OSX.  "MacPorts" and "Fink".  MacPorts still has a large userbase, but it's a pita to get changes in, and it has a seriously bad reputation for breakage.  Homebrew is very easy to use, good rep, etc.
17:02 kkeithley Seems like I ran into the same problem on Solaris too. I even had pkg-config from their open-source set of packages, but it was broken. Seems like if we can eliminate a potential problem point that would be a good thing.
17:02 jclift kkeithley: Yeah, no objection here.
17:02 * jclift will look at the patch now
17:02 kkeithley I've heard of MacPorts, don't remember Fink per se.
17:03 jclift Very legacy. :)
17:06 rastar joined #gluster-dev
17:10 kkeithley grrr, I updated XCode and it has deleted my /Developer, including /Developer/usr/bin/{autoconf,automake,glibtool}
17:12 jbrooks joined #gluster-dev
17:16 kkeithley Now I have to try to reconstruct where I got those bits from
17:19 kkeithley jclift: yesterday my build was falling down in .../xlators/mount/fuse. Are you saying your builds on the Mac were successful?  Maybe I don't have macfuse?
17:19 jclift kkeithley: Install homebrew. :)
17:20 jclift kkeithley: There's a "formula" (homebrew terminology for package) of "fuse4x".
17:20 kkeithley Well, or I could just install fedora on my Mac too as far as that goes. I'm trying to be as vanilla as possible.
17:21 jclift kkeithley: So, after installing Homebrew, you just need to type "brew install fuse4x"
17:21 jclift Ahhh, yeah.
17:21 jclift kkeithley: I use VMware Fusion a lot too.
17:21 jclift kkeithley: With the updated XCode, have you downloaded the matching command line tools for it?
17:22 jclift kkeithley: Guessing that was the prob with it nuking auto* ?
17:22 kkeithley I did
17:22 jclift Damn
17:22 jclift I'm still running XCode 4.6.  Apparently there's a 4.6.1 I should upgrade too.
17:22 jclift Maybe later on today.
17:24 kkeithley Right, I just updated to 4.6.1 yesterday, including the cmd line tools. But I didn't fire up XCode itself. When I did that this morning to check the version it asked me to confirm deleting old stuff. I guess that's when it blew away my /Developer/usr/bin/auto*
17:27 jclift You've fixed it though?
17:27 kkeithley No
17:27 * jclift wonders if re-installing the command line tools would work
17:28 kkeithley AppStore says it's already installed, doesn't let me install again.
17:28 jclift Er... where are you getting the command line tools from?
17:28 kkeithley and already deleted the dmgs
17:28 jclift 1 sec
17:29 kkeithley developer.apple.com sent me to the AppStore for both XCode and the XCode Command Line Tools
17:30 jclift kkeithley: It shouldn't have. :(
17:30 jclift kkeithley: I've just tried here, and the .dmgs are downloading now: https://developer.apple.com/download​s/index.action?name=for%20Xcode%20-#
17:31 jclift kkeithley: No redirecting to AppStore (for me anyway).
17:31 jclift kkeithley: Which version of OSX are you using?
17:32 kkeithley Snow Leopard, 10.8.3
17:32 jclift 10.8.3 is Mountain Lion
17:32 jclift But, no worries.
17:32 kkeithley Don't go by me though, I'm still jetlagged after coming back from India
17:33 kkeithley It's 10.8.3. Yeah, Mountain Lion
17:36 kkeithley yeah, I'm getting Cmd Line Tools from developer.apple.com. Not sure what I remember wrt to being redirected to the AppStore
17:38 kkeithley hmmm. root password on gqaib boxes has changed, so I can't tell you which IB cards are in them. Check with Ujjwala, she's the owner of those boxes.
17:46 lpabon joined #gluster-dev
17:51 jclift kkeithley: That change for the nuking of pkg-config looks good visually.
17:52 jclift kkeithley: Just about to test it across Fedora, CentOS, etc.  Give it an hour I guess.
17:52 * jclift gets coffee
17:53 kkeithley Yeah, I've done Fedora, and the green check means that gerrit/jenkins have done the regression on build.gluster.org, which is a CentOS 6.3 box
17:53 kkeithley but go for it.
17:55 blues-man joined #gluster-dev
18:27 jclift kkeithley: On OSX, I've been using "--disable-fuse-client".  So, haven't had to deal with fuse issues yet.
18:27 jclift One thing at a time, etc. :D
18:29 jclift kkeithley: Ahhh, now it's coming back to me.  http://fpaste.org/KGM4/
18:29 jclift kkeithley: Already BZ'd that.  And found likely good solution.
18:30 jclift kkeithley: openat() mkdirat() and fstatat() have replacement functions in Gnulib, and the licensing looks compatible with glusterd.
18:30 jclift kkeithley: I briefly tried integrating them um... weekend 2 weeks ago, but didn't have much luck.
18:31 * jclift is going to ask Eric Blake (guy mentioned above for autotools) for assistance.
18:31 jclift kkeithley: After all, he's one of the authors for the replacement functions. :D
18:31 johnmark jclift: heh... nice :)
18:31 johnmark is this for the glusterfs client?
18:31 jclift johnmark: Just glusterd so far.
18:31 johnmark ah, ok
18:32 jclift johnmark: Well, we don't seem to have a --without-gluster-daemon compilation option, so making it work is mandatory.
18:32 jclift (though that's an idea... :D)
18:37 johnmark heh heh
18:37 jclift johnmark: Hmmmm, I'm not sure that would work.  I don't remember if we need glusterd running on the client side too.
18:37 johnmark that would be interesting
18:37 * jclift knows glusterd manages glusterfsd for some thing.
18:37 jclift thins
18:37 johnmark jclift: yeah, the 2.x was without glusterd
18:37 johnmark I thought it had to be running on the client
18:37 jclift things
18:37 johnmark but maybe it's only on the server
18:37 * johnmark needs to check
18:38 jclift Yeah, not sure.  I just started writing up the BZ for --without-glusterd, then realised it might not be possible.
18:38 jclift Prob better if we get the Gnulib stuff done so it all works anyway.
18:38 * jclift gets back to testing stuff
18:54 jclift kkeithley: To get Gluster compiling all the way through on OSX, use --disable-fuse-client, and these two patches:
18:54 jclift https://github.com/justinclift/glusterfs/com​mit/0c13c6eae08481e5e69b1342accbbe2e19e6d13e
18:54 jclift https://github.com/justinclift/glusterfs/com​mit/bb47ffcace484b231586a2e3fc9a41ebec017df1
18:55 jclift kkeithley: Note, those are just dodgy "make it work" patches.  Functional for compiling, but could be completely wrong approach for stuff to actually work. :D
18:56 kkeithley Okay. Let me see where --disable-fuse-client gets me and then I'll look at your two patches.
18:57 jclift kkeithley: Np.  With --disable-fuse-client you'll probably then hit the missing openat() mkdirat() and linkat() stuff.  When you're up to there, look at the above patches. :D
18:57 * jclift gets back to testing stuff
18:59 jclift lpabon: For the dev workflow stuff page, wouldn't it be easier to just point people at the "how to compile rpms" at the appropriate point?  i.e instead of duplicating content?
19:00 lpabon jclift: good point, i'll clean that up
19:02 kkeithley hmmm, since the Mac libtool doesn't hide the compilation, including all the warnings, I'm looking at all the warning: 'SSL_* is deprecated (declared at /usr/include/openssl/ss.h.:$lineno warnings.
19:04 kkeithley I wonder if we'd see (if we could actually see them) these warnings on Linux.
19:04 jclift kkeithley: Probably.  Different warning suppression switches or something maybe.
19:04 * jclift ignored that stuff :D
19:09 kkeithley only deprecated on MacOS X version 10.7 and later, seems to be Apple specific. False alarm.
19:09 kkeithley and I just hit the openat(), mkdirat(), linkat() stuff
19:12 jclift kkeithley: Workaround: https://github.com/justinclift/glusterfs/com​mit/0c13c6eae08481e5e69b1342accbbe2e19e6d13e
19:13 jclift kkeithley: This is likely the better approach though: http://www.gnu.org/software/gnu​lib/MODULES.html#module=openat
19:14 kkeithley yup. I wasn't really intending to work on MacOS X build on company time apart from running my fix up the gerrit flagpole.
19:16 kkeithley Since I think it fixes Solaris as well, and somebody was "het up" a bit the other day about 3.3 and 3.4 not building on Solaris any more, I figured it'd be a good thing to get fixed.
19:17 lpabon jclift: Check out the dev workflow now, let me know what you think
19:18 jclift kkeithley: Yeah, it's an after hours thing for me too.
19:19 jclift lpabon: Give me a few mins.  Fixing a small prob that's turned up in the RPM compiling steps.
19:19 lpabon jclift: np
19:27 jclift lpabon: Seems worded decently. :)
19:27 jclift lpabon: Out of curiosity, why only Fedora systems being directed there?
19:28 * jclift is wondering if there's something wrong for CentOS/RHEL ones.
19:31 lpabon awesome
19:31 lpabon haven't had the time to setup a debian/ubuntu system yet
19:31 lpabon or RHEL/CentOS
19:31 lpabon doing that tomorrow :-).. still getting all my VMs setup
19:31 lpabon Maybe I can change it to RHEL/CentOS/Fedora instead of just Fedora
19:31 lpabon that'd prob work
19:31 lpabon what do you guys think?
19:31 jclift lpabon: I know for sure the compilation for rpms instructions work on F16-18 & both RHEL 6.4 + CentOS 6.4.
19:31 jclift lpabon: Tested every one of those systems individually, just to be 100% certain.
19:32 jclift lpabon: But, the rest of the dev workflow... you'd have to check. :D
19:35 lpabon joined #gluster-dev
19:38 kkeithley If you're talking about RPMs created by .../extras/LinuxRPM/ (and on the Fedora build system, a.k.a. koji) they build on fedora 15-20 and RHEL 5-7, including the CentOS clones.
19:39 jclift kkeithley: Cool.
19:39 kkeithley E.g. see here https://koji.fedoraproject.org/​koji/packageinfo?packageID=5443
19:40 kkeithley well, f15 fell off the end of the world when f18 shipped.
19:40 jclift Sure
19:41 jclift kkeithley: With gerrit, I'm very confident your change works on everything.  Can I just +2 it?
19:41 jclift Not sure if that's bad form, etc.
19:42 kkeithley I should refile it with a BZ number before you do that. I wanted to see what people thought before I did that though
19:43 kkeithley And I know it works on MacOS X. I should fire up my OpenSolaris box and confirm that it works there, although I'm pretty confident it will.
19:43 jclift Cool, go for it.  Also might want to update the commit text from "Snow Leopard" to "Mountain Lion"
19:43 jclift :)
19:43 kkeithley yup, I already fixed that. ;-)
19:44 kkeithley Although the stampede of people saying "yes, do that" is, ah, underwhelming.
19:44 kkeithley I'm just going to recycle your BZ rather than open a new one
19:45 jclift :)
19:46 jclift At some point I'm going to have to create a "gluster status" command that outputs peer and volume status and anything else useful.
19:46 * jclift is getting sick of typing long commands
20:16 jdarcy joined #gluster-dev
20:21 jdarcy_ joined #gluster-dev
21:12 wushudoin joined #gluster-dev
21:13 wushudoin joined #gluster-dev
23:35 yinyin joined #gluster-dev

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary