>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
>>>> "Adam" == Adam Duck
<duck(a)informatik.uni-frankfurt.de> writes:
Adam> I find this patch
suboptimal as the subprocess doesn't work
Adam> all the time -- at least here on my FreeBSD box. It takes
Adam> approx. 10x longer to unbzip2 edict.bz2 than on 21.4 (in
Adam> which I didn't experience any hangs).
Stephen> Did you experience hangs in 21.5 before the patch?
Stephen> What do you mean by "doesn't work all the time"? Sometimes
it hangs,
Stephen> sometimes it doesn't?
Sorry, I thought it was all obvious. I have been indeed experiencing
hangs with 21.5 (21.4 worked fine, but I didn't install the latest few
releases). I thought it didn't matter which process but fetching mail
always worked. So, as soon as I saw something with "process" in this
list, I cvs'upped and rebuilt. The problem was gone. But, the process
itself doesn't use 100% CPU time all the time. It begins, then stops,
my CPU graph doesn't show anything and then goes up and down until the
file is unzipped. So, it takes far more time to unzip edict.
Adam> Can _I_ do something about it (apart from coding it myself?
Adam> :-) ).
Stephen> Profile it and see where the time's going (M-: (profile FORMS) RET or
Stephen> M-x profile-key-sequence SINGLE-COMMAND, where SINGLE-COMMAND may be
Stephen> bound to a sequence of several keys). Then M-x profile-results.
Oh yes, I tried. Unfortunately xemacs crashes -- and doesn't leave a
core-file, which is strange as I do have many core files laying around
here on my FreeBSD-box (from XEmacs, too). I will submit a bug report.
Bye, Adam.
--
Adam Duck (duck(a)cs.uni-frankfurt.de)
Bockenheimer Landstr. 135 / Zi. 211
60325 Frankfurt/Main
______________________________________
If it jams, force it. If it breaks, it needed replacing anyway.