February 2012 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] petsc-dev on bitbucket
by Satish Balay 09 Feb '12

09 Feb '12

On Thu, 9 Feb 2012, Jed Brown wrote: > On Thu, Feb 9, 2012 at 20:59, Satish Balay <balay(a)mcs.anl.gov> wrote: > > > If we have to create multiple accounts to better organize repos - so > > be it.. > > > > I take it the concern is that we end up with a long list of release > repositories that would look like clutter (although they'll be sorted so > they wouldn't be near the top). Actually the number of repos is not a concern. But a reasonable organization is. The current *groups* we are are: releases tutorials externalpackages externalpackages-old Perhpas we might benifit with a few more - as times goes by. > Looking at the history, we really only ever have one release repository > that is accepting patches. After the 3.2 release, there are zero patches > applied to petsc-3.1. So maybe it would make sense to just have a single > "petsc-release" repository. You can update to tags if you want to revisit > something old. > Note that this would make our installation tutorials more > future-proof, always clone petsc-release if you want the release. sure you get some and loose some [features] :). Users get to update from release to release with the same clone [but then their externalpacakges will mess up their install ]. Yeah you have to tag just before pulling everything from 'dev' to release - and be able to create the equivalent 'old release' clones. Still I think the current organization is decent - and not in need of overhaul. If this gets importantat we can do it later aswell. Satish

1 0

Re: [petsc-dev] MatMatMult gives different results
by Matthew Knepley 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 9:10 PM, Satish Balay <balay(a)mcs.anl.gov> wrote: > On Thu, 9 Feb 2012, Jed Brown wrote: > > > It's all workable, > > but I'm not seeing a clear advantage to keeping the primary at > > petsc.cs.iit.edu, > > currently the 2 reasons that are offered agains petsc.cs.iit.edu are: > - ssh breaks - need https [to level comaprision - we can probably set > this up for folks who want this] > > - reliability concerns. > > > (provided Sean volunteers to set up mirroring in that direction). > > Personally - I'd prefer not to deal with that. If its decided that > bitbuckets is the thing ot use - we should disable petsc.cs.iit.edu, > > As you've noticed - mirroring has already caused confusion. And I > suspect this will keep coming up - much more often than the frequence > of a downed machine. The master should be on Bitbucket. We could keep a repo at IIT that is updated every time as a backup only if we need it, but for nothing else. Matt > > Satish > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Jed Brown 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 21:39, Matthew Knepley <knepley(a)gmail.com> wrote: > This organization is semantic, and Completely outside the version control > structure. I want something to tell me "this repo is about simulating > rockets" > like Kit, the voice in Michael's car. Tags are fine for this. So is a > hierarchy. > So is silly XML metadata. > Bitbucket's normal repository summary has space for this bit of text. It doesn't specifically have tags or automatic suggestions of repositories that might be related (but are not forks). I think the summary is good enough.

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Sean Farley 09 Feb '12

09 Feb '12

> > I want something to tell me "this repo is about simulating rockets" > like Kit, the voice in Michael's car. > Matt, I wasn't alive at the beginning of Knight Rider and even I know it's KITT for Knight Industries Two-Thousand.

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Matthew Knepley 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 8:59 PM, Satish Balay <balay(a)mcs.anl.gov> wrote: > On Thu, 9 Feb 2012, Sean Farley wrote: > > > > > > > Hg named branches are kind of screwy and bookmarks (which are less > screwy) > > > are still an "extension" > > > > > > > This is true. Named branches are directly inherited from older versioning > > systems (cvs, svn, etc.). After 2.0 Matt Mackall was convinced that > > bookmarks so be completely analgous to git's branching: > > > > $ hg branch foo > > marked working directory as branch foo > > (branches are permanent and global, did you want a bookmark?) > > > > Mercurial 2.1 includes fixes that move and update bookmarks automatically > > (and also allow pulling divergent bookmarks: foo@1, foo@2) > > > > so it makes sense to have separate clones to use for release management. > > > But these releases get merged back, so just tagging them would place > the > > > tarball in the right place. > > > > > > > Righto. > > > > > > > But PETSc release tarballs include generated documentation, so using > > > bitbucket's auto-generated tag-tarballs is not enough for releases. > > > > > > > Ah, that's right. I think this could be fixed by using the tag info in > the > > scripts that generate the documentation. > > Sure alternatives are possible. I guess the primary question is: is > are you suggesting this alternate workflow - just becuase its possible > - or because its better? This conversation started off with solving a > different problem - and then turned into changing workflow. > > We have a decent workflow currently - and the whole change of workflow > [eventhough its possible] -I don't think buying us much. And I suspect > it has additional complexity which some of us don't want to deal with. > I do not understand how this changes workflow. It does not change mine at all. My workflow relies on being able to push and pull from repositories. We can still do this. We do gain reliability, and backups. Please explain what we cannot do now. Matt > If we have to create multiple accounts to better organize repos - so > be it.. > > Satish > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Matthew Knepley 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 8:26 PM, Satish Balay <balay(a)mcs.anl.gov> wrote: > On Wed, 8 Feb 2012, Sean Farley wrote: > > > > > > > Hell, if you *really* want to, just create the account: > petsc-release(s) > > > then the URL would be > > > > > > http://bitbucket.org/petsc-release/petsc-3.1 > > > > > > > Actually, it's even easier than that: > > > > https://bitbucket.org/petsc/petsc-dev/downloads > > > > which provides downloads for all taged changesets. > > tags are no good. we implrement branches in different clones. > I am not sure what you mean by this. Let me be explicit. This organization is semantic, and Completely outside the version control structure. I want something to tell me "this repo is about simulating rockets" like Kit, the voice in Michael's car. Tags are fine for this. So is a hierarchy. So is silly XML metadata. Matt > You could argue that we should throw away branches in clones have all > clones in a single branch - and change our workflow. > > But I think this will be too confusing to most of us [yeah you could > change your bash prompt to always indicate wich branch you are on - > wich is equivalent to 'cd different clone' - but not all of us are > that sophisticated] > > Satish > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] MatMatMult gives different results
by Satish Balay 09 Feb '12

09 Feb '12

On Thu, 9 Feb 2012, Jed Brown wrote: > It's all workable, > but I'm not seeing a clear advantage to keeping the primary at > petsc.cs.iit.edu, currently the 2 reasons that are offered agains petsc.cs.iit.edu are: - ssh breaks - need https [to level comaprision - we can probably set this up for folks who want this] - reliability concerns. > (provided Sean volunteers to set up mirroring in that direction). Personally - I'd prefer not to deal with that. If its decided that bitbuckets is the thing ot use - we should disable petsc.cs.iit.edu, As you've noticed - mirroring has already caused confusion. And I suspect this will keep coming up - much more often than the frequence of a downed machine. Satish

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Jed Brown 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 20:59, Satish Balay <balay(a)mcs.anl.gov> wrote: > If we have to create multiple accounts to better organize repos - so > be it.. > I take it the concern is that we end up with a long list of release repositories that would look like clutter (although they'll be sorted so they wouldn't be near the top). Looking at the history, we really only ever have one release repository that is accepting patches. After the 3.2 release, there are zero patches applied to petsc-3.1. So maybe it would make sense to just have a single "petsc-release" repository. You can update to tags if you want to revisit something old. Note that this would make our installation tutorials more future-proof, always clone petsc-release if you want the release.

1 0

Re: [petsc-dev] petsc-dev on bitbucket
by Satish Balay 09 Feb '12

09 Feb '12

On Thu, 9 Feb 2012, Jed Brown wrote: > On Wed, Feb 8, 2012 at 11:31, Barry Smith <bsmith(a)mcs.anl.gov> wrote: > > > Yes. If you can't have that then it is a pretty crappy system. Sort of > > like IBM 360 where you only had a root directory with a truckload of files > > in it. In excusable in 2012 (or even 1972). > > > Let's all stop using google and switch to 1998 Yahoo. It's so much easier > to find things by navigating a hierarchy. all these comparisions don't make anysense to me. perhaps I'm dense. Whats the 'bitbucket' whay to easily get the info represented by this single link: http://petsc.cs.iit.edu/petsc/ Here I can get 2 baisc type of info - the global info regarding all the repos - and some grouping info - where I can identyfy the related repo collections. Don't care if its implemented with hierirachal dirs or something else. Satish

1 0

Re: [petsc-dev] MatMatMult gives different results
by Jed Brown 09 Feb '12

09 Feb '12

On Thu, Feb 9, 2012 at 20:48, Satish Balay <balay(a)mcs.anl.gov> wrote: > > > http://confluence.atlassian.com/display/BITBUCKET/Setting+Up+the+Bitbucket+… > > unless I'm mistaken - it looks like https - not ssh. > This is push notification of the commit. Your machine responds to this notification by issuing a pull. This is the most reasonable thing because it's rare that someone wants to give out remote access. > > I guess the alternative - is to push to your personal bitbucket clone > - and ask someone else to pull/sync to the master. [Isn't this one of > the things you and Sean were advocating as advantages of bitbucket? If > so - I think it should be equally useful for syncing with petsc-dev - > when hosted at petsc.cs.iit.edu] Sure, pushing to our own bitbucket repo and issuing a "pull request" is a fine model, but it's not what we have been working with (and it makes merges someone else's responsibility, to some extent). It's all workable, but I'm not seeing a clear advantage to keeping the primary at petsc.cs.iit.edu, (provided Sean volunteers to set up mirroring in that direction).

1 0