Re: crosstool: wrong assumption made wrt compression method of sourcebinaries (bz2 vs gz)

[Dan Kegel <dkegel at ixiacom dot com>]

Wolcott, Ken (MED, Compuware) wrote:
>   Looks like I found a bug in crosstool :-)  Not sure, please verify.
>
>   Not all compressed source binaries are compressed with bzip2, some are 
> compressed with gzip.  Therefore looking for glibc2.2.5.tar.bz2 fails where 
> glibc2.2.5.tar.gz would succeed.  Perhaps there should be an attempt to 
> download *.tar.bz2 first, and if that fails, then an attempt to download 
> *.tar.gz, prior to failing completely.  Also, failing completely, should exit 
> the script?  Can't find the required component locally, can't find the 
> required component on the web, can't build the cross compiler...rather than 
> failing later because the subcomponent isn't there?

getandpatch.sh tries .tar.bz2 first, then .tar.gz, just as you would like:

getUnpackAndPatch ftp://ftp.gnu.org/pub/gnu/glibc/$GLIBC_DIR.tar.bz2 || getUnpackAndPatch ftp://ftp.gnu.org/pub/gnu/glibc/$GLIBC_DIR.tar.gz

so I'm not sure why it didn't work for you... maybe there is a bug.

It relies on 'set -e' above to abort the script if they
both fail; maybe I should add
   || abort "can't download $GLIBC_DIR"
to the end of that line to make sure I'm not misunderstanding how set -e
and shell functions interact.

Also, in my next version I'll abort properly if the tarballs don't
unpack due to corruption; currently it just sails right through
(due to, yep, my misunderstanding of how set -e and shell functions interact).


BTW, if you're having trouble building busybox, just comment out the
line in ptx.config that enables busybox... you ought to sail through
to building ncurses then.
- Dan



------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com