3:42 p.m.
XEmacs prints "Cache full, freeing GC" from time to time to stderr.
There are more debug_out calls in xgccache.c, but they are all #if'd
out. That is what my patch does with the one that isn't.
I'll commit tomorrow, if nobody objects.
src/ChangeLog addition:
2006-02-16 Marcus Crestani <crestani(a)xemacs.org>
* xgccache.c (gc_cache_lookup): Remove "Cache full" warning.
afs-xemacs-21.5 source patch:
Diff command: cvs -q diff -u
Files affected: src/xgccache.c
Index: src/xgccache.c
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/src/xgccache.c,v
retrieving revision 1.13
diff -u -r1.13 xgccache.c
--- src/xgccache.c 26 Nov 2005 11:46:11 -0000 1.13
+++ src/xgccache.c 16 Feb 2006 15:37:41 -0000
＠＠ -233,7 +233,9 ＠＠
cache->head = cell->next;
cache->head->prev = 0;
if (cache->tail == cell) cache->tail = 0; /* only one */
+#if 0
debug_out ("Cache full, freeing GC: %08lx\n ", XE_GCONTEXT(cell));
+#endif
XFreeGC (cache->dpy, cell->gc);
cache->delete_count++;
#ifdef GCCACHE_HASH
--
Marcus
4:45 p.m.
Thu, 16 Feb 2006 16:42:39 +0100, Marcus Crestani wrote:
XEmacs prints "Cache full, freeing GC" from time to time to
stderr.
There are more debug_out calls in xgccache.c, but they are all #if'd
out. That is what my patch does with the one that isn't.
I'll commit tomorrow, if nobody objects.
I think this ok when --enable-debug=no
--8<---------------cut here---------------start------------->8---
#ifdef XEMACS_DEBUG
debug_out ("Cache full, freeing GC: %08lx\n ", XE_GCONTEXT(cell));
#endif
--8<---------------cut here---------------end--------------->8---
src/ChangeLog addition:
2006-02-16 Marcus Crestani <crestani(a)xemacs.org>
* xgccache.c (gc_cache_lookup): Remove "Cache full"
warning.
afs-xemacs-21.5 source patch:
Diff command: cvs -q diff -u
Files affected: src/xgccache.c
Index: src/xgccache.c
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/src/xgccache.c,v
retrieving revision 1.13
diff -u -r1.13 xgccache.c
--- src/xgccache.c 26 Nov 2005 11:46:11 -0000 1.13
+++ src/xgccache.c 16 Feb 2006 15:37:41 -0000
＠＠ -233,7 +233,9 ＠＠
cache->head = cell->next;
cache->head->prev = 0;
if (cache->tail == cell) cache->tail = 0; /* only one */
+#if 0
debug_out ("Cache full, freeing GC: %08lx\n ", XE_GCONTEXT(cell));
+#endif
XFreeGC (cache->dpy, cell->gc);
cache->delete_count++;
#ifdef GCCACHE_HASH
--
Regards,
Andrey.
5:54 p.m.
>>>>"AS" == Andrey Slusar
<anrays(a)gmail.com> writes:
> XEmacs prints "Cache full, freeing GC" from time to time to stderr.
> There are more debug_out calls in xgccache.c, but they are all #if'd
> out. That is what my patch does with the one that isn't.
AS> I think this ok when --enable-debug=no
AS> --8<---------------cut here---------------start------------->8---
AS> #ifdef XEMACS_DEBUG
AS> debug_out ("Cache full, freeing GC: %08lx\n ",
XE_GCONTEXT(cell));
AS> #endif
AS> --8<---------------cut here---------------end--------------->8---
No. The symbol you probably mean is DEBUG_XEMACS. DEBUG_XEMACS is
enabled by default for beta builds, which means that this doesn't
quiet down my XEmacs, and I really don't want to read this output on
my console. I don't know what this output is supposed to mean, I get
it a lot and the fact that some cache is full doesn't seem to bother
my running XEmacs at all. And I don't want to turn off DEBUG_XEMACS
either since I really need it to track down allocator- and GC-related
problems.
As I already mentioned, there are a lot of these debug_outs in
xgccache.c, and they are all turned off by "#if 0". Seems to me, that
the one I spotted is a left over.
Can you clear me up what this warning message is supposed to tell me?
--
Marcus
4:26 a.m.
>>>> "Marcus" == Marcus Crestani
<crestani(a)informatik.uni-tuebingen.de> writes:
Marcus> As I already mentioned, there are a lot of these
Marcus> debug_outs in xgccache.c, and they are all turned off by
Marcus> "#if 0". Seems to me, that the one I spotted is a left
Marcus> over.
I did some (not yet complete) work with the gc cache, and for my own
convenience did it in the sjt-xft branch, and it's probable that it's
a leftover from that. It doesn't really convey any information by
itself, it's more like a trace. Xft uses an internal pool of GCs or
something like that, so this code probably shouldn't get hit very
often if you're using Xft. Or possibly it got moved out of a
different #if* ... anyway, I don't see them so probably it's something
I missed when I #if'd the rest. Sorry.
Of course it's harmless to disable it with #if 0. I'd appreciate it
if you not completely remove it in pending review of whether they're
useful to me at all anymore.
--
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
7:50 a.m.
sjt> I did some (not yet complete) work with the gc cache, and for my own
sjt> convenience did it in the sjt-xft branch, and it's probable that it's
sjt> a leftover from that. It doesn't really convey any information by
sjt> itself, it's more like a trace. Xft uses an internal pool of GCs or
sjt> something like that, so this code probably shouldn't get hit very
sjt> often if you're using Xft. Or possibly it got moved out of a
sjt> different #if* ... anyway, I don't see them so probably it's something
sjt> I missed when I #if'd the rest. Sorry.
I get this message from time to time in a non-xft build. As far as I
remember, I did never get this message in a xft-enabled build. If you
need more information about my configuration, let me know.
sjt> Of course it's harmless to disable it with #if 0. I'd appreciate it
sjt> if you not completely remove it in pending review of whether they're
sjt> useful to me at all anymore.
I have no plans to remove it completly, #if 0 is fine. I'll commit
this afternoon.
--
Marcus
11:01 a.m.
>>>> "Marcus" == Marcus Crestani
<crestani(a)informatik.uni-tuebingen.de> writes:
Marcus> I get this message from time to time in a non-xft build.
Marcus> As far as I remember, I did never get this message in a
Marcus> xft-enabled build.
Good, that confirms my suspicions about why I don't see it.
Marcus> I have no plans to remove it completly, #if 0 is fine.
Marcus> I'll commit this afternoon.
Don't let me rush you. ;-)
--
School of Systems and Information Engineering http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
6637
days inactive
6638
days old
6 comments
3 participants
participants (3)
-
Andrey Slusar -
Marcus Crestani -
Stephen J. Turnbull