This is the mail archive of the
binutils@sourceware.cygnus.com
mailing list for the binutils project.
Re: [new test case jluu@mainsoft.com: Re: Possible linker problem]
- To: Jose Luu <jluu at mainsoft dot com>
- Subject: Re: [new test case jluu@mainsoft.com: Re: Possible linker problem]
- From: "H . J . Lu" <hjl at lucon dot org>
- Date: Wed, 15 Mar 2000 07:24:38 -0800
- Cc: Ian Lance Taylor <ian at zembu dot com>, geoffk at cygnus dot com,binutils at sourceware dot cygnus dot com
- References: <031001bf8e7e$c064fad0$2a0da8c0@fred.mainsoft.com>
On Wed, Mar 15, 2000 at 02:02:41PM +0100, Jose Luu wrote:
>
> For us it is a normal option, because it is the behavior of the Windows
> linker, and all these Windows applications that we are porting to Linux
> needs it (for behavior rather than speed), especially the complex ones, like
> the gigantic professional CAD programs.
>
> >
> >Maybe we should just detect this case and issue a warning.
> >Particularly since H.J. reports that the Solaris linker doesn't do a
> >better job.
> >
>
> A warning will be welcomed. We will still have a problem with the g++ RTTI
> information which uses this mechanism, I have to think about it further.
>
We have several options:
1. You can use the symbol versioning. You only export the symbol you
want to export. The versioning scripts take regular expressions (?)
for symbol names.
2. We can add a switch to ld to do data copy for -Bsymbolic. If it
causes any conflicts, you can use the symbol versioning to solve
them.
3. Ian mentioned that ld could do the hardwork in the main executable
to detect -Bsymbolic.
#3 may be the best. But it will take time. #2 should be easier to do.
H.J.
H.J.