[IPython-user] Problem doing svn co behind proxy
Sat Aug 11 14:01:58 CDT 2007
On 8/11/07, David Cournapeau <firstname.lastname@example.org> wrote:
> Fernando Perez wrote:
> > On 8/10/07, Ville M. Vainio <email@example.com> wrote:
> >> One more reason to switch to mercurial some day ;-)
> > I really love hg, but that switch requires some infrastructure support
> > to preserve our Trac setup and SVN history, and none of that is easily
> > available at this moment, nor do I have the time to invest in it.
> I don't know much hg, and certainly do not want to push anything, but
> just mention that bzr has support for svn, and it is quite good. I have
> never tried with ipython svn, but it works for numpy and scipy svn,
> which are arguably bigger, right ? bzr also serves directly in http. I
> don't know how feasable it would be, but maybe putting some ready only
> bzr repository synchronised with the svn would be possible ?
> I have some problems with proxy (my uni network has no DNS, everything
> is done through the proxy: great way to expose proxy bugs in softwares),
> and none with bzr.
Well, I have to admit that recenly I found bzr maddeningly slow. I
used it with Gael V. while playing with some Envisage-related code,
and I couldn't believe how slow it was for simpmle things. hg, on the
other hand, is nice and zippy. I know bzr does some things better
than hg (it's stupid that hg can't serve out of a plain http
directory, for example), but the slowness would drive me ballistic in
day to day work. I'm just too impatient to wait for a tool to make
tea and bake bread every time I ask it for a simple task. hg was
written from day one to cope with something like the linux kernel, so
speed was a *design constraint* instead of the afterthought it is for
So if we ever find the time to switch to a distributed VCS, it will be
either hg, or bzr only if they've fixed these problems.
I honestly just don't have *any* time to invest on this. But if
anyone sends me a shell/python script that does a sync of svn and
exposes it publicly over http, I'm perfectly happy to call it in the
nightly cron job that makes the ipython1 API docs and the SVN exports.
More information about the IPython-user