[poppler] Moving the poppler test framework to git
krh at bitplanet.net
Sun Nov 12 13:55:53 PST 2006
On 11/10/06, Brad Hards <bradh at frogmouth.net> wrote:
> On Wednesday 08 November 2006 12:39, Jeff Muizelaar wrote:
> > I'd like to move the test framework that lives in the test module of
> > poppler cvs to git. Given that I seem to be the only one that develops
> > it, I don't really expect any objections. However, I thought it'd be
> > good to ask first anyways.
> I would prefer that you leave it in CVS. The qt4 unit tests rely on test files
> that are present in that module.
> If we are talking about making a move, then I'd be happy to see all of poppler
> moved to SVN (or potentially some other tool, but most of the decentralised
> tools presume some kind of central maintainer to pull a releasable tree
> together, and we don't really have anyone who is over all of poppler). If
> you still want an offline capability, then SVN+svk might be a possibility.
This is a common misconception about decentralized version control
systems. The fact that they enable an upstream-pulls style of
development doesn't mean your project have to work that way. As Jeff
points out, for cairo and X.org, we use a model much like we did with
CVS. Everybody who had CVS write access have git write access (it's
the same ssh accounts and groups after all) and now push their changes
into the upstream git tree as they previously committed to CVS.
And even if you work with a centralized repository, a decentralized
version control system still has a lot of benefits over a
client-server bound system such as cvs or svn: anything you can do
on-line you can also do off-line, including commits, diffs, history
browsing, plus most operations are much faster as they don't require
network IO. And it's easier to publish your own repo with your local
changes if you want that.
If anybody is curious, I have a git mirror of poppler here that I've
been syncing for a while:
If there is consensus for moving to git (moving everything, that is -
poppler, poppler-data and the test and performance suites), I'd be all
for it, but I don't see a big reason to move to SVN. However, I have
a feeling there is no such consensus, and staying with CVS isn't a big
deal for me, so I haven't been pushing this.
More information about the poppler