RE: http access to source code

["Drew Adams" <drew.adams@oracle.com>]

> > This sorry sorry-thank-you page has been thanking us for 
> > our patience for a _very_ long time now.
> 
> Have you tried to access that page a few times in a row.  

Yes, indeed, I have - naturally.  That's usually what I do, for the reasons you
mention below.

> I've seen this message as well, but if I retry right away,
> it usually works 

Sometimes it does; I agree.  Too often it does not.

To me, this HTTP access is woefully inadequate.  Others do not seem to complain,
however.  Dunno whether there are only a few people who try HTTP access, with
nearly everyone using BZR instead.  I use HTTP to download specific source
files, so for me this is a sorry regression wrt the state of things prior to the
move to BZR. (I'm *not* making any general statement about BZR vs CVS or GIT or
whatever.)

> (my mental model to explain it: the Emacs repository is so large
> that it takes too long to swap it in, so the webserver timeouts
> and gives an error even though the operation actually works fine,
> so if you try it again right away, it can work because some of
> the data is already swapped in and
> the whole process can then finish before the timeout).

Yes, I've imagined the same thing.  And I do try refreshing (repeatedly).  And
sometimes I do get through.  But not often enough to consider that workaround a
remedy.

More anecdotal support for the model you suggest is that on the days that you
ultimately can get through, it definitely is not a good idea to open two browser
windows to different parts of the repository. ;-)