GNU bug report logs - #9941
Handle truncated member names in LIBADD archive libraries

Previous Next

Package: libtool;

Reported by: "Daniel Richard G." <skunk <at> iSKUNK.ORG>

Date: Wed, 2 Nov 2011 21:00:02 UTC

Severity: normal

Full log


View this message in rfc822 format

From: "Daniel Richard G." <skunk <at> iSKUNK.ORG>
To: "Bob Friesenhahn" <bfriesen <at> simple.dallas.tx.us>
Cc: 9941 <at> debbugs.gnu.org
Subject: bug#9941: Handle truncated member names in LIBADD archive libraries
Date: Thu, 03 Nov 2011 13:14:10 -0400
On Thu, 2011 Nov  3 08:51-0500, Bob Friesenhahn wrote:
>
> This is an interesting problem.  There would be similar problems if
> the underlying filesystem did not support long file names (as was
> common on SYSVR.3 systems).

I'm not sure the build process would even work on a system like that!
Though it would be interesting to try.

> What happens if several of the original object file names are
> identical for the first 15 characters?

Oh, I was afraid someone was going to ask that...

    $ ls
    LongObjectNameIsLong1.o    LongObjectNameIsLong2.o

    $ ar cru liblong.a *.o
    ar: filename LongObjectNameIsLong1.o truncated to LongObjectNameI
    ar: filename LongObjectNameIsLong2.o truncated to LongObjectNameI

    $ ar t liblong.a 
    LongObjectNameI
    LongObjectNameI

    $ mkdir extract; cd extract

    $ ar x ../liblong.a 

    $ ls
    LongObjectNameI

Nothing that can't be addressed by implementing an analog of Windows
"PROGRA~1"-style name mangling, but I was hoping to avoid that
approach :(


--Daniel


-- 
Daniel Richard G. || skunk <at> iSKUNK.ORG
My ASCII-art .sig got a bad case of Times New Roman.




This bug report was last modified 13 years and 229 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.