Fwd: [Gc] FW: GC: Time for GC final release? (patch for cleanups)

NIIBE Yutaka gniibe at fsij.org
Mon Jul 26 18:06:42 PDT 2010


Ivan Maidanski wrote:
> And this patch solves the cleanup problem ("apparently on Solaris
> and probably on Linux we can't collect while a thread is exiting,
> since signals aren't handled properly, this gives currently gives
> rise to deadlocks") you observed before, right?

It is true that the symptom of the cleanup problem I observed before
is solved by the patch.

But it is *not* me who report Solaris deadlock issue you quote, so I can't
say about this.

On March, I was working for Gauche 0.9, and found the failure on threads
with GC 7.1 (in this case on HPPA):
     https://buildd.debian.org/build.cgi?pkg=gauche;ver=0.9-8

Then, I found the problem of the cleanup.  The problem itself was
reported by Shiro Kawai on November 2009:

     (0)
     http://www.hpl.hp.com/hosted/linux/mail-archives/gc/2009-November/003550.html

I made patches two times:

     (1)
     http://www.hpl.hp.com/hosted/linux/mail-archives/gc/2010-March/003836.html

     (2)
     http://www.hpl.hp.com/hosted/linux/mail-archives/gc/2010-April/003841.html

For Gauche on Debian, I used the patch of (1), then (2).  It is the
file debian/patches/12_libgc_fix__EXCEPTIONS.patch in the distribution
of Gauche on Debian.

Then, you gave me your patch last week.  I backport your patch for GC
7.1 and replace my patch by yours for Gauche on Debian.  It is Gauche
0.9-16, and works fine for all supported platforms.

    https://buildd.debian.org/build.cgi?pkg=gauche;ver=0.9-16

That's all I can say.
-- 


More information about the Gc mailing list