RE: libtool support for intel icc compiler

[Himanshu_Khurana]

Robert:
Yes it works


I included the line 

_LT_AC_TAGVAR(archive_cmds_need_lc, $1)=no

Both under 
case $CXX
icpc)


AND

case $CC
icc|ecc)


leaving the following piece as it is
 
> # Assume -lc should be added
>  _LT_AC_TAGVAR(archive_cmds_need_lc, $1)=yes


RESULT 
make was errorless and so was the libtool usage with my programs!!

regards
Himanshu 



-----Original Message-----
From: Allan Sandfeld Jensen [mailto:address@hidden
Sent: Wednesday, April 09, 2003 3:24 AM
To: Robert Boehne; Himanshu_Khurana
Cc: address@hidden; Roberto Bagnara
Subject: Re: libtool support for intel icc compiler


On Tuesday 08 April 2003 15:57, Robert Boehne wrote:
> Himanshu:
>
> Does this patch work?
>
> Robert
>
> Index: libtool.m4
> ===================================================================
> RCS file: /cvsroot/libtool/libtool/libtool.m4,v
> retrieving revision 1.314
> diff -u -r1.314 libtool.m4
> --- libtool.m4        5 Apr 2003 19:32:58 -0000       1.314
> +++ libtool.m4        8 Apr 2003 13:54:50 -0000
> @@ -4850,6 +4850,7 @@
>      linux*)
>        case $CC in
>        icc|ecc)
> +     _LT_AC_TAGVAR(archive_cmds_need_lc, $1)=no
>       _LT_AC_TAGVAR(lt_prog_compiler_wl, $1)='-Wl,'
>       _LT_AC_TAGVAR(lt_prog_compiler_pic, $1)='-KPIC'
>       _LT_AC_TAGVAR(lt_prog_compiler_static, $1)='-static'
>
It's not the right patch. The problem is that people keep changing the way 
archive_cmds_need_lc is calculated and overrides whatever I set it to (it is 
set to no in the beginning as default). Sometimes it works in libtool HEAD 
sometimes it doesnt. I would rather remove this part:

> # Assume -lc should be added
>  _LT_AC_TAGVAR(archive_cmds_need_lc, $1)=yes

Since it is obviously wrong and breaks linking on some platforms.

Another way to protect us from random breakage in libtool, could be to 
introduce an extra variable "archive_breaks_with_lc" that overrides need_lc 
if set to yes. 

`Allan