Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-09-24

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

All times shown according to UTC.

Time Nick Message
01:22 bala joined #gluster-dev
01:28 bala joined #gluster-dev
01:40 awheeler joined #gluster-dev
02:05 lalatenduM joined #gluster-dev
02:06 lala joined #gluster-dev
02:19 lala joined #gluster-dev
02:45 kshlm joined #gluster-dev
02:53 vshankar joined #gluster-dev
02:59 kanagaraj joined #gluster-dev
03:22 shubhendu joined #gluster-dev
04:04 itisravi joined #gluster-dev
04:23 bulde joined #gluster-dev
04:27 mohankumar joined #gluster-dev
04:28 bulde joined #gluster-dev
04:30 ppai joined #gluster-dev
04:46 ndarshan joined #gluster-dev
04:46 Shyam joined #gluster-dev
04:49 shubhendu_ joined #gluster-dev
04:49 Shyam left #gluster-dev
04:51 Shyam joined #gluster-dev
04:58 hagarth joined #gluster-dev
05:01 lala joined #gluster-dev
05:10 bala joined #gluster-dev
05:27 raghu joined #gluster-dev
05:30 ndarshan joined #gluster-dev
05:33 aravindavk joined #gluster-dev
05:34 lala__ joined #gluster-dev
05:51 lala joined #gluster-dev
05:55 atalur joined #gluster-dev
05:57 spandit joined #gluster-dev
05:57 bulde joined #gluster-dev
06:06 mohankumar joined #gluster-dev
06:18 rgustafs joined #gluster-dev
06:19 shubhendu_ joined #gluster-dev
06:23 kanagaraj joined #gluster-dev
06:24 ndarshan joined #gluster-dev
06:24 aravindavk joined #gluster-dev
06:27 bala joined #gluster-dev
06:59 mohankumar joined #gluster-dev
07:00 inodb joined #gluster-dev
07:56 shubhendu_ joined #gluster-dev
07:56 aravindavk joined #gluster-dev
07:56 kanagaraj joined #gluster-dev
07:58 ndarshan joined #gluster-dev
07:58 bala joined #gluster-dev
08:35 Shyam joined #gluster-dev
08:47 lala joined #gluster-dev
09:05 Guest9445 joined #gluster-dev
09:48 ppai joined #gluster-dev
10:22 shubhendu_ joined #gluster-dev
10:23 ndarshan joined #gluster-dev
10:26 kanagaraj joined #gluster-dev
10:37 edward2 joined #gluster-dev
10:43 vshankar joined #gluster-dev
10:45 Shyam joined #gluster-dev
10:53 Shyam left #gluster-dev
10:57 aravindavk joined #gluster-dev
11:00 bala joined #gluster-dev
11:08 ppai joined #gluster-dev
12:02 itisravi_ joined #gluster-dev
12:22 hagarth joined #gluster-dev
12:24 kanagaraj joined #gluster-dev
12:43 jclift_ joined #gluster-dev
12:46 avati joined #gluster-dev
12:46 ndarshan joined #gluster-dev
12:47 inodb joined #gluster-dev
13:01 awheeler joined #gluster-dev
13:02 awheeler joined #gluster-dev
13:40 ndarshan joined #gluster-dev
13:43 bala joined #gluster-dev
14:21 __Bryan__ joined #gluster-dev
14:29 portante ndevos, kkeithley: where are we at with libgfpi python wrapper work?  It seems stalled ...
14:31 ndevos portante: the discussion we had with lpabon resulted in the agreement that there would be a .../site-packages/glusterfs/{__init__,gfapi}.py
14:31 portante really?
14:31 portante how come?
14:31 portante is it because of packaging?
14:31 ndevos yes, but then you and kaleb chimed in that there should be only one gluster Python package
14:32 ndevos now, I'm happy either way, but creating the one gluster python package requires the .../site-packages/gluster/__init__.py to be packaged in gusterfs-api
14:33 ndevos if we package that __init__.py in glusterfs-api, other packages with that file will conflict and can not be installed at the same time
14:33 portante so why not provide site-packages/gluster/__init__.py as a common base package shared by both?
14:33 ndevos so, we need an agreement (or something) that glusterfs-api includes that __init__.py, and your packages do not contain that anymore
14:34 ndevos I think that glusterfs-api would be used by your packages in future anyway?
14:34 portante yes, so why don't we just depend on that?
14:34 portante is lpabon around?
14:34 * ndevos looks around, no.
14:35 ndevos depending on glusterfs-api would work, but systems that update could run into conflicts - if one of the new packages is not installed at the same time
14:38 portante there has to be a way to have to packages install into a shared hierarchy using RPM spec files ...
14:38 kkeithley I don't agree that that I said there should be one (sub)package. IIRC I argued that there should only be one gluster namespace.
14:39 ndevos yes, kkeithley, namespace == "python package"
14:39 portante great, so how can we have two packages install into the same namespace?
14:39 portante ndevos: really?
14:39 portante that seems likes a packaging artifact
14:40 kkeithley libgfapi.py in one package, gluster-openstack-swift is another package
14:40 kkeithley all in .../python*/site-packages/gluster.
14:40 ndevos .../site-packages/gluster <- the python package "gluster"?
14:41 ndevos anyway, glusterfs-api would contain .../site-packages/gluster/gfapi.py (isnt that a "python module"?)
14:42 kkeithley I'm arguing for 2 (or more) RPMs that install in .../site-packages/gluster. If that's considered one python package and doesn't violate any python or fedora packaging guidelines then I think that the right thing
14:42 hagarth joined #gluster-dev
14:43 kkeithley I think that's the right thing.
14:43 kkeithley But I'm willing to be convinced otherwise
14:43 kkeithley Are we saying the same thing?
14:43 ndevos well, of course multiple RPMs can install .py files under .../site-packages/gluster - we wont maintain gfapi.py and swift/g4s/.. in the same upstream project
14:44 kkeithley The debate centered on which RPM is going to install .../site-packages/gluster/__init__.py
14:44 ndevos but the one thing that needs a decision, is which RPM contains .../site-packages/gluster/__init__.py
14:44 kkeithley Otherwise I believe we're all in violent agreement? Am I wrong?
14:44 ndevos and, as a result of that, how are updates handled
14:45 * ndevos agrees with that
14:45 portante can't there be logic that says: if site-packages/gluster/__init__.py exist don't overwright it?
14:45 portante or you make a third RPM package that both gfapi and g4s depend on that provides the gluster hierarchy root
14:46 ndevos well, you could, but it would be extremely ugly, and probably not really conform the fedora guidelines
14:46 ndevos I dont like adding more packages
14:46 kkeithley But glusterfs-openstack-swift will (eventually) need libgfapi.py. Why can't we just have glusterfs-openstack-swift Require: glusterfs-api, which is where libgfapi.py will come from?
14:47 portante that works for me
14:47 kkeithley IOW why wouldn't libgfapi.py be in the glusterfs-api RPM?
14:47 ndevos glusterfs-api can provide 'python-gluster', contain both gfapi.py and the __init__.py and glusterfs-openstack-swift requires python-gluster
14:49 kkeithley If 'python-gluster' is a "virtual" python package (i.e. it provides .../site-packages/gluster) that satisfies some python and/or RPM semantics then I think we're all in agreement
14:49 ndevos in order to get updates running smoothly, we can add explicit 'Conflicts: glusterfs-openstack-swift <= $version_with_init_py' in glusterfs-api, glusterfs-openstack-swift just adds the 'Requires: python-gluster'
14:50 ndevos that would prevent installing a new glusterfs-api alongside the glusterfs-openstack-swift where both contain the __init__.py
14:51 kkeithley so far there isn't a glusterfs-openstack-swift RPM, let alone one with the __init__.py. At least not AFAIK
14:52 kkeithley Not in Fedora or EPEL. Is there one in RHS?
14:52 ndevos well, if there is something that has .../site-packages/gluster/ it should have the __init__.py file too
14:54 kkeithley I'm confused then. If glusterfs-openstack-swift has both Requires: glusterfs-api (with .../site-packages/gluster/__init__.py) and the rest of its files in .../site-packages/gluster/*, does that not satisfy that?
14:56 ndevos huh? glusterfs-openstack-swift has a requires on files under .../site-packages/gluster/* it provides itself?
14:57 ndevos if there are no packages released that contain .../site-packages/gluster/* files, then there is not need to worry about an upgrade path
14:58 kkeithley glusterfs-api has .../lib64/libgfapi.so, .../site-packages/gluster/libgfapi.py, and site-packages/gluster/__init__.py.  glusterfs-openstack-swift has .../site-packages/gluster/swift/*
14:58 kkeithley glusterfs-openstack-swift has a Requires: glusterfs-api.
14:58 ndevos that looks good then, all is set
14:58 kkeithley Would that satisfy everyone?
14:58 ndevos everyone here yes, but someone needs to convince lpabon :)
14:58 portante that seems to work from my, now to convince lpabon ...
14:59 portante lpabon!
14:59 * ndevos yells "lpabon" and people look strange at me
15:00 kkeithley glusterfs-openstack-swift has a Requires: glusterfs-api, and/or Requires python-gluster and glusterfs-api has a Provides: python-gluster.
15:00 kkeithley Why doesn't lpabon like this?
15:00 kkeithley (I know, ask him.)
15:00 portante ;)
15:01 ndevos kkeithley: see one of the last comments on http://review.gluster.org/5835
15:02 ndevos I think that patch should do it, just get it merged!
15:02 * ndevos will be back in 20 minutes or so....
15:07 kkeithley At the risk of stating the obvious those comments just boil down to single namespace (site-packages/gluster) versus two namespaces (site-packages/glusterfs and site/packages/gluster)
15:08 kkeithley Unless more votes for two namespaces materialize from somewhere then it seems like we have a decision. The rest is just packaging details.
15:09 portante agreed
15:26 ndevos yes, one site-packages/gluster is surely nicer than two
15:57 mohankumar joined #gluster-dev
16:17 [o__o] left #gluster-dev
16:19 [o__o] joined #gluster-dev
16:21 [o__o] left #gluster-dev
16:23 [o__o] joined #gluster-dev
16:24 [o__o] left #gluster-dev
16:26 [o__o] joined #gluster-dev
16:47 __Bryan__ joined #gluster-dev
16:49 __Bryan__ joined #gluster-dev
17:12 lala joined #gluster-dev
17:15 lalatenduM joined #gluster-dev
19:20 johnmark xavih: heya. let me know if you're around
20:57 badone joined #gluster-dev
21:42 foster avati: i don't think discard by itself is sufficient for size management, it expects a range of sectors
21:43 foster (i.e., we'd have to do rmw based on what we think the sectors are that straddle a range)
22:41 badone joined #gluster-dev
23:47 mjrosenb joined #gluster-dev
23:54 semiosis joined #gluster-dev
23:54 johnmark joined #gluster-dev
23:54 badone joined #gluster-dev
23:54 xavih joined #gluster-dev
23:54 samkottler joined #gluster-dev

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