libtool-patches
[Top][All Lists]
Advanced

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

Re: -msg_* flags for cc on osf


From: Ralf Wildenhues
Subject: Re: -msg_* flags for cc on osf
Date: Fri, 6 Jul 2007 17:38:51 +0200
User-agent: Mutt/1.5.15 (2007-05-20)

* Peter O'Gorman wrote on Fri, Jul 06, 2007 at 05:10:46PM CEST:
> On Fri, 2007-07-06 at 17:05 +0200, Ralf Wildenhues wrote:
> > 
> > Well I checked now.  You can't allow through -msg* because the system
> > that you would break would be just this very compiler:
[...]
> I think you are misreading the patch.

Yes, I misread it.  Apologies.

> Right now, libtool allows any flag
> starting with -m through to the linker. The issue that Bob encountered
> was that the -msg_disable flag was being sent to the linker, but its
> argument was not. The patch *stops* flags starting with -msg_ from
> getting passed to the linker when creating a shared library. It has
> nothing to do with the arguments really.

Ah, ok.  So then I was looking at the wrong thing, and the correct
question to ask is: can any of the sensible arguments to the -msg_*
options be misinterpreted by libtool as not to throw away?

Example:
  libtool -n --mode=link $CC -o liba.la a.lo -rpath /foo \
          -msg_enable -pg -msg_dump -p

would allow through -pg and -p, although cc would only interpret -p as
option, but -pg as argument to -msg_enable.  (I'm only taking these two
flags as an example.)

The -msg_* options that take arguments take either a list of numbers, or
some name specifying a message group or a level.  So it seems they will
always be ignored by libtool.  So it seems your patch is safe to apply.

Cheers,
Ralf




reply via email to

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