After having received tacit approval from the Board to implement infrastructure for libspatialindex, I would like to explore what variant of distributed version control system SAC would like to support. Because it is such a low-key project, I think a DVCS would be a good fit for libspatialindex. There's no committee, a couple of rather inactive main contributors, and frequent one-off patches that tend to show up to fix this or that or add new features here and there.
I think that the DVCS we choose must support Trac. I don't wish at all to revisit our bug tracking software. My investigation shows both Mercurial and git Trac plugins support our existing (0.10.3.1) version of Trac. Versions of these plugins also exist to support newer versions of Trac (0.11+), so we will have some upgrade room.
Are there any other contenders besides hg and git?
As with subversion, I'm willing to be the point guy on this, but I want us to only provide one. For folks with experience with DVCS's, which fits with how OSGeo operates, and which should we advertise to our member projects (or project wards like libspatialindex) as supporting?
After having received tacit approval from the Board to implement infrastructure for libspatialindex, I would like to explore what variant of distributed version control system SAC would like to support.
Good idea!
Are there any other contenders besides hg and git?
I was investigating DVCS myself some time ago,
together with SOCI (soci.sf.net) team and the conclusion was that
these two are the only mature, efficient, robust and well-supported
tools available.
There is plenty ofcomparisons and reviews of DVCS' on Web and blogs
and most of the ones I've seen are focused on Mercurial and git.
It seems to be indicative.
After having received tacit approval from the Board to implement
infrastructure for libspatialindex, I would like to explore what
variant of distributed version control system SAC would like to support.
Good idea!
Are there any other contenders besides hg and git?
I was investigating DVCS myself some time ago,
together with SOCI (soci.sf.net) team and the conclusion was that
these two are the only mature, efficient, robust and well-supported
tools available.
There is plenty ofcomparisons and reviews of DVCS' on Web and blogs
and most of the ones I've seen are focused on Mercurial and git.
It seems to be indicative.
Best regards,
Thanks, that is a good link to get a gist of what drives the decision.
I would suggest to go for Mercurial. This is not motivated by any
technical understanding but the fact that with GeoToolkit we have a
OSGeo friendly project that is already using it.
Best regards,
Arnulf.
- --
Arnulf Christl
Spatial Systems Architect
WhereGroup www.wheregroup.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
On Jul 28, 2009, at 10:15 AM, Arnulf Christl wrote:
Thanks, that is a good link to get a gist of what drives the decision.
I would suggest to go for Mercurial. This is not motivated by any
technical understanding but the fact that with GeoToolkit we have a
OSGeo friendly project that is already using it.
I admit to leaning toward Mecurial as well, if only for my major Python bias as well as my past experience with it on bitbucket. If there isn't much more discussion on this topic, I will probably start working toward implementing Mecurial.
On Jul 28, 2009, at 10:15 AM, Arnulf Christl wrote:
Thanks, that is a good link to get a gist of what drives the decision.
I would suggest to go for Mercurial. This is not motivated by any
technical understanding but the fact that with GeoToolkit we have a
OSGeo friendly project that is already using it.
I admit to leaning toward Mecurial as well, if only for my major Python bias as well as my past experience with it on bitbucket. If there isn't much more discussion on this topic, I will probably start working toward implementing Mecurial.
There is one major cons of using Git - its developers completely ignores existence of Windows platform, so the installation drags whole
bundle of MinGW stuff.
On Tue, Jul 28, 2009 at 06:00:00PM +0100, Mateusz Loskot wrote:
There is one major cons of using Git - its developers completely ignores
existence of Windows platform, so the installation drags whole
bundle of MinGW stuff.
Some FlightGear Scenery people are happily using MSysGit:
Martin.
--
Unix _IS_ user friendly - it's just selective about who its friends are !
--------------------------------------------------------------------------
On Tue, Jul 28, 2009 at 06:00:00PM +0100, Mateusz Loskot wrote:
There is one major cons of using Git - its developers completely ignores
existence of Windows platform, so the installation drags whole
bundle of MinGW stuff.
Some FlightGear Scenery people are happily using MSysGit: