possible bug in pure lfs

James Iwanek chthon at chthon-uk.com
Sun Mar 30 03:39:18 PST 2003


Greg Schafer wrote:

> On Mon, Mar 24, 2003 at 01:16:33AM +0000, James Iwanek wrote:
>> this might just be a problem on my system, but when building binutils
>> using:
>> 
>>     mkdir ../binutils-build &&
>>     cd ../binutils-build &&
>>     CFLAGS="-O2 -pipe" ../binutils-2*/configure --prefix=/usr
>>     --enable-shared && make LDFLAGS="-s" &&
>>     make check &&
>>     make install &&
>>     rm /usr/lib/libiberty.a
>> 
>> and building gcc using --enable-version-specific-runtime-libs
>> 
>> code generated over two consecutive chapter6 builds is not consistant
>> 
>> ive attached a diff between binutils built in the first ch6 and the
>> second ch6 to show the problem
>> 
>> btw yesterday i build plfs using the standard instructions and there were
>> no differences that affected the final programs so its definantly
>> something to do with the ommision of tooldir from binutils and the
>> addition of --enable-version-specific-runtime-libs to gcc
> 
> I cannot see the point in comparing the build directories. What matters is
> the installed binaries. If they differ, then refer back to the build dirs
> to discover the reason. That is how I've found all the bugs so far.
> 
> Greg

ill assume that if stripped (strip -s) the installed binaries and the
binaries in the build dirs are the same (except for some .a files). when i
checked my last build a LOT of executables (bash, bison, flex and others)
differed between two builds.

btw the only reason i diff betwen the build dirs is that its quicker than
putting all the executables in one place and `cmp`ing it ;-) (atleast for
me)

FYI: it seems that it was the fact that i had not built bison flex and m4 in
ch5 that was causing the problem (the differing config.logs told me that
;-) ) - i was using binutils-2.13.2.1 so the note in the hint should be
changed.

-- 
Jay
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list