Robert Collins wrote:
Pierre A. Humblet wrote:
At 11:36 PM 4/1/2003 +0100, Max Bowsher wrote:
I've uploaded a new setup troubleshooting snapshot:
http://www.cygwin.com/setup-snapshots/setup-2.340.2.3-no-set_default_sec.exe
This is simply 2.340.2.3 with the recently-added ntsec code
Can you update that patch to apply to the setup-200303-troubleshooting
branch?
To Robert:
What is the eventual fate of this branch?
Once all the code is reinstated...
Merge back into HEAD and setup-200303.
Cheers,
Rob
Vince Hoffman wrote:
Well I can t get either setup-2.340.2.3-O0-debug.exe or
setup-2.340.2.3-no-set_default_sec.exe to crash :)
using drmingw on the crash from setup-2.340.2.3.exe gives
the attached file. I am running Windows 2000 Pro SP3 on a duron processor.
patched to the latest from windows
Pierre A. Humblet wrote:
I have no idea what in the ntsec code is causing random crashes.
However a couple of week ago I submitted another patch, which
simplifies the logic. Rather than pursuing the current problem
it might be simpler to see if it still happens with the second patch.
Max Bowsher wrote:
I had a question which I never asked:
If the users Primary Group is NOT None/Domain Users, presumably it is set
that way for a reason - should we change it?
That's the conservative attitude I took at first. However I now realize
it doesn't help at all in the case of setup.
Pierre A. Humblet wrote:
Max Bowsher wrote:
I had a question which I never asked:
If the users Primary Group is NOT None/Domain Users, presumably it is set
that way for a reason - should we change it?
That's the conservative attitude I took at first. However I now realize
it doesn't help
Pierre A. Humblet wrote:
I can do that tonight EST but it would be simpler if you send me the original
of the file you want to patch. Otherwise I will need to pull my cvs manual
and figure out how to get it.
follow http://sources.redhat.com/cygwin-apps/setup.html, and grab HEAD.
I've fixed the
Robert Collins wrote:
Pierre A. Humblet wrote:
I can do that tonight EST but it would be simpler if you send me the
original of the file you want to patch. Otherwise I will need to pull my
cvs manual and figure out how to get it.
follow http://sources.redhat.com/cygwin-apps/setup.html, and
The setup-200303 branch tag has been inadvertantly moved, leaving the
original branch untagged and inaccessible, and also giving rise to the
insanely long version number assigned to the current setup snapshot.
I've inspected the RCS files, and can restore the setup-200303 to it's
correct location
Max Bowsher wrote:
The setup-200303 branch tag has been inadvertantly moved, leaving the
original branch untagged and inaccessible, and also giving rise to the
insanely long version number assigned to the current setup snapshot.
I've inspected the RCS files, and can restore the setup-200303 to
On Wed, 2 Apr 2003, Max Bowsher wrote:
The setup-200303 branch tag has been inadvertantly moved, leaving the
original branch untagged and inaccessible, and also giving rise to the
insanely long version number assigned to the current setup snapshot.
I've inspected the RCS files, and can
Robert Collins wrote:
Max Bowsher wrote:
The setup-200303 branch tag has been inadvertantly moved, leaving the
original branch untagged and inaccessible, and also giving rise to the
insanely long version number assigned to the current setup snapshot.
I've inspected the RCS files, and can
Max Bowsher wrote:
I request cvsadmin membership so I can continue the cleanup from this stage,
I can't grant this (unless it's a cvs repository specific thing)..
Can you work with a local copy of the modules and sync it up afterwards?
Rob
On Thu, Apr 03, 2003 at 09:01:53AM +1000, Robert Collins wrote:
Max Bowsher wrote:
I request cvsadmin membership so I can continue the cleanup from this
stage,
I can't grant this (unless it's a cvs repository specific thing)..
Can you work with a local copy of the modules and sync it up
Pierre A. Humblet wrote:
I've fixed the bug in both HEAD and the release branch. We'll save your
second patch for the next release.
OK. Hasn't that bug been in there forever?
I'm not sure. Whats really weird is the way it's suddleny cropped up.
Rob
Robert Collins wrote:
Max Bowsher wrote:
I request cvsadmin membership so I can continue the cleanup from this
stage,
I can't grant this (unless it's a cvs repository specific thing)..
But I suspect Chris can. I assume it shouldn't be a problem, as it doesn't
give me any additional access,
Christopher Faylor wrote:
On Thu, Apr 03, 2003 at 09:01:53AM +1000, Robert Collins wrote:
Max Bowsher wrote:
I request cvsadmin membership so I can continue the cleanup from this
stage,
I can't grant this (unless it's a cvs repository specific thing)..
Can you work with a local copy of
Robert Collins wrote:
Max Bowsher wrote:
Go for it. I didn't realise CVS was so broken.
Forgot to say in my last email:
Do you recall the command you ran that caused this? Presumably something
like cvs (r)tag -F. Or was it a script that went wrong? I see some
instances of cvs rtag -F in
Max Bowsher wrote:
Oh yes, the 2 page assertion failiure.
I can't imagine why it doesn't work for you. It works for me, testing on a
copy of the repository into my home dir *on sources.redhat.com*. You can't
get any closer to reproducing the situation than that, without actually
touching the
Robert Collins wrote:
No, I used:
$ cvs -z4 tag -Fb setup-200303
in a setup-200303-troubleshooting working dir.
Which intuitively says ...? This should not have updated the cvs
repository. It would have been commits at a later date that would have
updated the repository. The ``tag'' updates
On Thu, Apr 03, 2003 at 12:21:35AM +0100, Max Bowsher wrote:
Robert Collins wrote:
Max Bowsher wrote:
I request cvsadmin membership so I can continue the cleanup from this
stage,
I can't grant this (unless it's a cvs repository specific thing)..
But I suspect Chris can. I assume it shouldn't
Earnie Boyd wrote:
Robert Collins wrote:
No, I used:
$ cvs -z4 tag -Fb setup-200303
in a setup-200303-troubleshooting working dir.
Which intuitively says ...? This should not have updated the cvs
repository. It would have been commits at a later date that would have
updated the repository.
22 matches
Mail list logo