This is the mail archive of the cygwin mailing list for the Cygwin 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: "git clone" over ssh fails intermittently w/snapshot 20120123


Christopher Faylor wrote:

>> On Mon, Jan 30, 2012 at 08:05:46AM -0800, Kevin Layer wrote:
>> >Christopher Faylor wrote:
>> >>> Is this new behavior just seen with this snapshot or is it something you
>> >>> noticed in 1.7.9?  If you've just seen it in the latest snapshot it
>> >>> would be useful (as I've previously pleaded) to know in which snapshot
>> >>> the bad behavior first showed up.
>> >
>> >Never seen until the snapshot was installed.
>> 
>> That isn't precise enough.  Are you saying that you saw the problem in
>> the 2012-01-23 snapshot but not in the 2012-01-22 snapshot?  I'll bet
>> that you aren't.  If you think this is a problem only manifested in
>> snapshots then we need to know which snapshot first had the
>> problem.

No.  I didn't see it in 1.7.9, but do see it in the referenced
snapshot.

>> >More info:
>> >
>> >Before the failure, I had removed and reinstalled cygwin, then
>> >installed the snapshot.  The machine on which git works was upgraded
>> >little at a time.
>> 
>> OTOH, "little at a time" might mean that you stepped through
>> snapshots.

No, I meant that the machine on which I've never seen the failure,
even with the same 1/23 snapshot, had the packages on it upgraded once
a month (roughly).  I didn't remove everything and reinstall once a
month, for example.

On the machine that I have seen the errors, I removed everything,
because I was in a weird state with my msysgit testing, so I thought
it best to start from scratch.  I removed msysgit and gnuwin32, as
well.

>> However, since the only change to 2012-01-23 was to tweak fifos and, as
>> far as I know, git clone doesn't use a fifo, it's hard to see why there
>> would be a regression in 2012-01-23.

If you can think of some other tests I could perform, I'd be happy to
do it.  The machine with the failures is not a critical development
machine, so it's pretty open to what I can do.


--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple


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