[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Libtool 2.2.6 uses unknown -b option w/HP-UX cc

From: Ralf Wildenhues
Subject: Re: Libtool 2.2.6 uses unknown -b option w/HP-UX cc
Date: Mon, 10 Nov 2008 20:09:33 +0100
User-agent: Mutt/1.5.18 (2008-05-17)

Hi Daniel,

* Daniel Richard G. wrote on Mon, Nov 10, 2008 at 07:08:41PM CET:
> I am building a shared library with Libtool 2.2.6 on an HP-UX system, using 
> the native HP C compiler. This is what I am seeing:
> ----
> libtool: link: mv -f ".libs/libfoo.expT" ".libs/libfoo.exp"
> libtool: link: cc -b -Wl,+h -Wl,libfoo.sl -Wl,+b -Wl,/tmp/project/lib -o 
> .libs/libfoo.sl  .libs/xxjni.o .libs/mcatobj.o   -L/opt/lib/pic -L/opt/lib 
> /opt/lib/pic/libxxlib.a -L/opt/lib /opt/lib/libxml2.a -lz /opt/lib/libiodbc.a 
> /opt/lib/libiodbcinst.a -lcl -lpthread /opt/lib/libpcre.a -lm -lc  +DA2.0
> cc: warning 422: Unknown option "b" ignored.
> /usr/ccs/bin/ld: Unsatisfied symbols:
>    main (Not referenced yet! Probably due to -u option)
> gmake[5]: *** [libfoo.la] Error 1

Does it work if you replace -b with -Z in archive_cmds,
archive_expsym_cmds inside the libtool script?

> I believe this is a system version issue---Libtool is assuming that it is 
> working with a newer version of the C compiler that does support -b.

I agree.

> Some details on the versions involved:
> % uname -a
> HP-UX hrdygrdy B.11.00 A 9000/785 2003934647 two-user license
> % cc -V -c hello.c -o /dev/null
> cpp.ansi: HP92453-01 A.11.01.20 HP C Preprocessor (ANSI)
> ccom: HP92453-01 A.11.01.20 HP C Compiler

Works here on 11.00 with
cpp.ansi: HP92453-01 B.11.X.35175-35176.GP HP C Preprocessor (ANSI)
ccom: HP92453-01 B.11.X.35098-35101.GP HP C Compiler

OTOH this compiler on 10.20 warns about -b, too:
cpp: HP92453-01 A.10.32.30 (WM) HP C Preprocessor
ccom: HP92453-01 A.10.32.30 HP C Compiler

> If I recall correctly, this issue has come up before, with the 1.5.x 
> series.

Hmm.  I don't recall.

> Libtool was inferring that the compiler was new enough just from 
> the system version---which works in most cases, but this system was a 
> counterexample where that premise doesn't hold true.

Seems like we need a compile test.

Thanks for the bug report!


reply via email to

[Prev in Thread] Current Thread [Next in Thread]