On Tue, 12 Nov 2002, Jarkko Hietaniemi wrote:
On Mon, Nov 11, 2002 at 02:10:43PM -0500, Karl Matthias wrote:
of the files I actually need. The whole unicode tree is somewhere near
5MB which would be about 25% of our entire ditribution. That's not
5MB? Really? That small? Ahhh, I guess that this is the unicore/
directory. You should be able to just nuke it.
Yeah I was talking about the unicore directory. But then I
discovered even more.
acceptable, so I'm trying to figure out which files we actually need from
that tree. We will never use the Unicode support so if there is some way
If one is really never going to use any of the Unicode support,
after install one should be able to
rm -rf $Config{installprivlib}/Encode
rm -rf $Config{installprivlib}/unicore
rm -rf $Config{installprivlib}/Unicode
rm -rf $Config{installprivlib}/utf8*
rm -rf $Config{installarchlib}/Encode*
rm -rf $Config{installarchlib}/Unicode
rm -rf $Config{installarchlib}/auto/Encode
rm -rf $Config{installarchlib}/auto/Unicode
Another way would be to disable the Encode and Unicode::* extensions
already during the configure/build, that way only the unicore and
utf8* removals from the above would be needed. But judging by your
5 MB figure it sounds like that's what you have done.
Yeah I tried building it without the Encode and other Unicode
related extensions. What we do is actually just package up the parts we
use, so I am not _ex_cluding things, but rather having to figure out what
to _en_clude. Even some of the packages to trace module dependencies
don't seem to be able to wade through this stuff.
to actually just remove (not that I've found so far), that would be
So you have tried removing the unicore/ directory? If so, what happened?
Well after much frustration and head pounding I finally figured
out that the newer version of one of the modules we use was calling
another module which uses the Unicode stuff and that's why it kept showing
up as required. I am now going to try to revert the module version and
see if I can still get it to work on 5.8 without Unicode.
Thanks for your suggestions.
Karl
