[petsc-dev] get rid of https://bitbucket.org/petsc/petsc-dev NOW
Barry Smith
bsmith at mcs.anl.gov
Thu Sep 4 11:01:41 CDT 2014
On Sep 4, 2014, at 10:10 AM, Satish Balay <balay at mcs.anl.gov> wrote:
>
> On Thu, 4 Sep 2014, Barry Smith wrote:
>
>>
>> On Sep 4, 2014, at 9:12 AM, Satish Balay <balay at mcs.anl.gov> wrote:
>>
>>> how about 'petsc-dev' mailing list?
>>
>> :-) Though that could be slightly confusing it is just the petsc developers mailing list and distinct from “petsc-dev” which means some mystical repository somewhere (what branch?) that does need to be eradicated from our vocabulary.
>
> Incomplete statements will always such ambiguity. There is no harm in
> asking for clarification.
>
> petsc-dev It primarily refers to 'master branch' [or taball off
> it]. If one is using it differently - then there is an error.
>
> If eliminating petsc-dev [wrt repo/tarball] we also need to change
> http://www.mcs.anl.gov/petsc/developers/index.html
Good point, I’ve eliminated all the unneeded gratuitous uses of petsc-dev there.
We really should change the nightly tar ball name to petsc-master.tar.gz from petsc-dev.tar.gz.
We should also change all the web pages from petsc-dev to petsc-master but I won’t lobby for that since would be a major upheaval to google.
Barry
>
> What about maint? How does one interpret it? Does it refer to 'maint'
> branch aka release or petsc-maint e-mail? How about the bad corelation
> 'petsc-dev at mcs' for petsc-develoment discussion petsc-maint at mcs -
> petsc release discussions only?
>
> Or perhaps rename petsc-dev@ as petsc-development@ and petsc-maint@ as petsc-maintainance@
>
> This maint/master/next terminology is great for gitworkflow users -
> not petsc users [who don't know or care about gitworkflows(8) terminology]
>
> Or we should always say 'master *branch*' or 'maint *branch*' [and a
> followup e-mail: master = development-stable next=development-testing
> etc..] Look at all the email replies from petsc developers that don't
> mention it [and we expect our users to be complete..]
>
>> Note the email we just got:
>>
>> I'm trying to build petsc-dev and it's failing on configuring Zoltan. I've attached the configure.log
>>
>> Thanks,
>> John
>>
>>
>> I have not clue what branch John is talking about.
>
> petsc-dev refers to master branch. But one doesn't have to rely on that when we have logs..
>
> $ grep VERSION_GIT configure.log
> #define PETSC_VERSION_GIT "unknown"
> Defined "VERSION_GIT" to ""v3.5.1-194-gde9c8a0""
> #ifndef PETSC_VERSION_GIT
> #define PETSC_VERSION_GIT "v3.5.1-194-gde9c8a0"
>
> $ git branch -r --contains de9c8a0
> origin/HEAD -> origin/master
> origin/knepley/feature-fe-fv-hybrid
> origin/master
> origin/tisaac/petsc-next-upstream
> origin/tisaac/thplex-on-master
> origin/tisaac/warn-next-upstream
>
> Satish
More information about the petsc-dev
mailing list