Re: glibc-2.3 new malloc
>>>> "juhp" == Jens Petersen
<petersen(a)redhat.com> writes:
juhp> "Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
>> >>>> "juhp" == Jens
Petersen <petersen(a)redhat.com> writes:
>>
juhp> "Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
> >>
> >> 1. separate dump file (unclean, and paranoid sysadmins wet
> >> their pants when they see big globs of who knows what in a
> >> system bindir ;^)
>
juhp> How about putting the dump file is libdir then? ;)
>
> libdir is defined in the dump file. The only thing XEmacs knows at
> invocation time is where it is, itself.
juhp> I don't doubt that this has already been thought over.
juhp> Presumably this is so that xemacs can run both in the
juhp> buildtree and in the installtree?
juhp> Still how hard would it be to say first test the current
juhp> directory and then then look in libdir for a the dump file?
Very hard, because there's complicated algorithmics involved in
figuring out where libdir *is*. It's not just defined in the dump
file---there's code in the dump file to figure it out. The advantage
of the current approach is that it's very predictable.
--
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla