On 2016-11-14 20:05-0500 Hazen Babcock wrote:

> On 09/23/2016 06:29 AM, Tom Schoonjans wrote:
>> Moving the git repository to Github itself would be trivial. Just
>> use https://github.com/new/import, but I do recommend using the PLplot
>> organization as owner. Just add yourself to it first and get sufficient
>> privileges. Hazen I think you would be able to help out here since you
>> seem to be in charge of this organization.
>
> I have transferred my personal PLplot github mirror to the PLplot 
> organization (Hezekiah and are currently the only members):
> https://github.com/PLplot/PLplot
>
> Now I would like to add some things into master that I think would be of some 
> benefit to the project, but are primarily there specifically for Github.
>
> (1) A README.md file. This would have a short project description as well as 
> links to the SF site, the documentation, etc. as well as some badges (see (2) 
> below).
>
> (2) Some .yml files, for example, one to provide "continuous" testing using 
> Travis-CI (.travis.yml).
>

Hi Hazen:

I don't mind you adding individual github-related files like you have
indicated above, but I do have some overall concerns with use of
github by PLplot developers. For example, as a major contributor to
PLplot, I really like our "no merge commit" workflow rules since it
leads to a clear history mere humans can understand on all master and
topic branches.  I am also satisfied with the service provided by our
SF git server and our configuration of that server so that it enforces
our workflow rules on the server side. Therefore, I currently plan to
stick with that server as our definitive git repository for now.

My primary concern is use of github would allow PLplot developers to
be sloppy about following our workflow rules leading to convoluted git
history that is impossible for humans to understand. However, if you
can reassure me that the github version also _enforces_ our workflow
rules like the SF server does (or you plan to configure the github git
server for PLplot that way soon) that would clearly answer this
concern. That github workflow enforcement would make it trivial to
push github master branch updates to our definitive SF git server
following exactly the workflow documented in README.developers that
any PLplot developer with a local git repository follows now.  But if
you don't have that github enforcement, then the github master branch
history will inevitably get filled with merge commits making it
impossible to push that branch to the SF server.

In sum, it would be a positive thing to establish a github git server
for PLplot that enforces our workflow rules on the server side since
that insures workflow discipline for all those using that repository
and would be a step towards the possibility that we might move our
definitive git repository to github. (Or at least it gives us that
option in the future if we ever get dissatisfied with the SF git
service).  But without that workflow enforcement, I believe a github
repository for PLplot would not be very practical for many of us to
use because of the "merge commit" and bad history reasons stated above.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

------------------------------------------------------------------------------
_______________________________________________
Plplot-general mailing list
Plplot-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-general

Reply via email to