Re: [ANN][RFC] cygipc-1.11 at cygutils

[Charles Wilson <cwilson at ece dot gatech dot edu>]

Jason Tishler wrote:

> Chuck,
> 
> [Sorry for the sluggish response time...]
> 
> On Thu, Nov 29, 2001 at 12:41:04PM -0500, Charles Wilson wrote:
> 
>>Horak Daniel wrote:
>>
>>
>>>>What does postgresql do if there is no system-wide union semun?
>>>>
>>>There is a check for union semun in configure and then if it does not
>>>find a system wide union semun it defines its own as
>>>
>>>[snip]
>>>
>>Ah, then cygipc-1.11 won't work for postgresql.  Cygwin (and Linux) 
>>implementations of IPC need a fourth union memeber,
>>[snip]
>>This sounds like a vote for 'return to 1.10 behavior'.  Any other comments?
>>
> 
> I just tried the simplistic approach: configure, make, make check.
> PostgreSQL 7.1.3 builds cleanly and passes all regression tests.  So it
> seems that cygipc 1.11 is still compatible with PostgreSQL.


I guess postgresql never calls sem_ctl() with the IPC_INFO or SEM_INFO 
cmd.  (Those commands cause the libcygipc code to try to access union 
semun's __buf field, which may fail if the calling program doesn't 
declare union semun "properly")

This is the "doctor, it hurts when I..."  "don't do that" solution.

It relies on postgresql not exercising certain code paths.  Bad and flaky.

--Chuck






--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/