This is the mail archive of the 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: RFD: cygipc ENOSYS patch


On Tue, Nov 05, 2002 at 10:38:41AM -0500, Jason Tishler wrote:
On Wed, Oct 30, 2002 at 08:35:41PM +0100, Peter Eisentraut wrote:
> I'm getting tired of the "initdb hangs" class of complaints.  Why
> doesn't the relevant function fail if the ipc-daemon isn't running?
> Can anything be done in that area?
> Motivated by the above, I finally looked into this long standing
Cygwin PostgreSQL problem.  Specifically, Cygwin PostgreSQL will hang
and consume all available CPU cycles if cygipc's ipc-daemon is not
running when either initdb or postmaster is started.
> I'm proposing that the attached patch (against cygipc-1.11-1) be
applied to cygipc since it solves the Cygwin PostgreSQL hang problems.
Specifically, this patch changes cygipc's shmget() and semget() to
return ENOSYS instead of EACCES if ipc-daemon is not running.
See the following for the full details:

The Cygwin PostgreSQL and KDE communities seem to be happy with the
above.  Would you be willing to accept such a patch into cygipc CVS?

If so, then please let me know because I just found some more places
where ENOSYS should be returned instead of EACCES.  However, I would
like to gauge your interest before pursuing this further.

I've no objection. But would somebody PLEASE tell me what the HELL we're waiting for with respect to the 64bit key_t change in newlib?

I've had a new release of cygipc waiting for months, predicated on that change, which I THOUGHT would happen "real soon".

Now, I have to revert all of that (I didn't make a branch point at the time -- why bother? the change was "imminent") before I release any other updates. and following.


Unsubscribe info:
Bug reporting:

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