This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: sim checkout broken


On Thu, 27 Jun 2013, Tom Tromey wrote:
> Tom> Yeah, good point.  I didn't try that.
> Tom> The reason it works for me is this line from .cvsrc:
>
> Tom> checkout -P
>
> Well, that works for the initial checkout, but now I can't make it work
> at all for subsequent checkouts.  I don't get it, since I know this was
> working yesterday.

It probably stopped working when you read that I'd tried that
too! 8]  On the bright side, it looks like you've repeated my
observation.  Or maybe not really brighter but at least a
change in illumination.

I copied the entire src repo out of sourceware (~5GiB, a
445025730 tarball, so I could experiment with different
solutions - and luckily no-one checked in something
significant while I copied).

After a necessary adjustment (changing LockDir in
CVSROOT/config) checking out locally, with -d /path/to/cvsfiles
works (i.e. complains but exits 0).  Weirdness.  It's not the
pserver method because I get the same behavior using ssh; i.e.
-d :ext:hp@sourc...

For the record, the ~/.cvsrc trying ssh was:
diff -N -p -u
cvs -z3 -q
checkout -P
update -Pd

Luckily I get the same exit(1) behavior using the ssh method
with my copy of the repo, so I'll try some semi-random ideas.
(And I welcome any hunch.)

> I think moving the file out of 'common' should help.

Maybe.  I'm not so sure but I'll try it locally.

By the way, I re-noticed the CVSROOT/history file, with about
1532526399 bytes of useless information (AFAIK).  It even has a
history.old 2147487356 bytes long.  Time for a history.old2
early next year?  Ok, I'm digressing.  Can't we just switch to
git? :D

brgds, H-P


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]