Mail Archives: cygwin/2010/04/08/04:15:10
Am 08.04.2010, 08:11 Uhr, schrieb Christopher Faylor:
> On Wed, Apr 07, 2010 at 10:43:08PM -0400, Mark Levedahl wrote:
>> On 03/31/2010 09:36 PM, Mark Levedahl wrote:
>>>> On Tue, Mar 30, 2010 at 08:08:37PM -0400, Mark Levedahl wrote:
>>>>> All of the above connections and i/o processing are done using
>>>>> bog-standard fork, pipe, open, read, write, and associated file
>>>>> commands. Git has no trouble opening all of the processes and pipes
>>>>> under Cygwin 1.7.2. However, in general before the data transfer is
>>>>> complete the demux process issues an error upon getting an incomplete
>>>>> packet (fewer bytes transmitted than declared in the header), and
>>>>> debugging shows the error code of the read() is ENOENT, indicating
>>>>> the
>>>>> pipe or socket has been terminated early. The index-pack process also
>>>>> issues an error due to receipt of an incomplete pack.
>> Given the above, it seems logical the problem is with pipes, and as
>> 1.7.4-1 is has fixes for pipes I tried it out and found a significant
>> improvement: using 1.7.2-1 I got 16 failures out of 20 attempts to clone
>> a 17 MByte repo on localhost, using 1.7.4-1 this is down to 3 failures
>> out of 20 attempts. (1.7.3-1 was not an improvement).
>
> And, yet, there were no fixes for pipes in 1.7.4.
Not to suggest anything particular to Mark, but rather as a general
comment: I wonder how many of the "1.7.X no better" can be attributed to
the replace-on-reboot failure up to and including setup 2.693. If people
rely on that rather than running "uname -a" or similar, those that were
told to reboot may be unaware they're using the older version.
--
Matthias Andree
--
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
- Raw text -