John Polstra <[EMAIL PROTECTED]> wrote:
> In article <[EMAIL PROTECTED]>, Alexander N. Kabaev
> <[EMAIL PROTECTED]> wrote:
>
> > Why FreeBSD does not link libgcc into shared libraries by
> > default? Everyone else is doing that. Linking shared libraries
> > with libgcc seems to be the ultimate
In article <[EMAIL PROTECTED]>,
Maxim Sobolev <[EMAIL PROTECTED]> wrote:
> After the crt changes the following piece of code, which worked previously,
> gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> "__register_frame_info' error message (yeah, I know that it's better to check
>
>
> I don't think it is allowed to do that. See the SUSv2 description:
>
> http://www.opengroup.org/onlinepubs/007908799/xsh/dlerror.html
>
Sorry, you are right of course. The attached patch removes the _rtld_error
call at the end of the symlook_default function which seems to be unnecessar
In article <[EMAIL PROTECTED]>, Alexander N. Kabaev
<[EMAIL PROTECTED]> wrote:
> I think dymanic loader should clear it's error_string variable
> before returning success to the caller.
I don't think it is allowed to do that. See the SUSv2 description:
http://www.opengroup.org/onlinepubs/0
John,
I think dymanic loader should clear it's error_string variable before returning
success to the caller. I am halfway done preparing the patch to do that. Is
that the bug you mentioned or there is something else? Anyway, do you want to
see my patch when it is ready or you prefer to take care
In article <[EMAIL PROTECTED]>,
Alexander N. Kabaev <[EMAIL PROTECTED]> wrote:
> > Nope, explicitly linking shared module with -lgcc doesn't solve
> > this problem.
>
> Then it is not new crt* code that is broken
That's correct. The new crt code just uncovered the bug in the
dynamic linker.
Jo
In article <[EMAIL PROTECTED]>,
Maxim Sobolev <[EMAIL PROTECTED]> wrote:
> The full sources of this testcase can be found at:
> http://people.freebsd.org/~sobomax/dlbug.tar.gz .
Thanks for the nice test case. It made my job a lot easier. :-)
The bug is in the dynamic linker. I'll append the
> Nope, explicitly linking shared module with -lgcc doesn't solve this problem.
>
> -Maxim
Then it is not new crt* code that is broken
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
John Polstra wrote:
> In article <[EMAIL PROTECTED]>,
> Maxim Sobolev <[EMAIL PROTECTED]> wrote:
> > After the crt changes the following piece of code, which worked previously,
> > gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> > "__register_frame_info' error message (yeah, I kn
"Alexander N. Kabaev" wrote:
> Why FreeBSD does not link libgcc into shared libraries by default? Everyone
> else is doing that. Linking shared libraries with libgcc seems to be the
> ultimate work-around. Are there any compatibility problems which are keeping
> FreeBSD from doing that?
>
> > O
In article <[EMAIL PROTECTED]>, Alexander N. Kabaev
<[EMAIL PROTECTED]> wrote:
> Why FreeBSD does not link libgcc into shared libraries by
> default? Everyone else is doing that. Linking shared libraries
> with libgcc seems to be the ultimate work-around. Are there any
> compatibility problems
Why FreeBSD does not link libgcc into shared libraries by default? Everyone
else is doing that. Linking shared libraries with libgcc seems to be the
ultimate work-around. Are there any compatibility problems which are keeping
FreeBSD from doing that?
> OK, thanks. I'll take a look at it as soo
I've spent couple hours already trying to reproduce the error and so far
failed miserably. Am I the only one who does not see this problem at all?
On 06-Nov-00 Maxim Sobolev wrote:
> After the crt changes the following piece of code, which worked previously,
> gives a 'host: dlopen() failed: ./
In article <[EMAIL PROTECTED]>,
Maxim Sobolev <[EMAIL PROTECTED]> wrote:
> After the crt changes the following piece of code, which worked previously,
> gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> "__register_frame_info' error message (yeah, I know that it's better to check
>
Kris Kennaway wrote:
> On Mon, Nov 06, 2000 at 10:14:32AM +0200, Maxim Sobolev wrote:
>
> > Jordan Hubbard wrote:
> >
> > > > After the crt changes the following piece of code, which worked previously,
> > > > gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> > > > "__register_frame
On Mon, Nov 06, 2000 at 10:14:32AM +0200, Maxim Sobolev wrote:
> Jordan Hubbard wrote:
>
> > > After the crt changes the following piece of code, which worked previously,
> > > gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> > > "__register_frame_info' error message (yeah, I know
Jordan Hubbard wrote:
> > After the crt changes the following piece of code, which worked previously,
> > gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> > "__register_frame_info' error message (yeah, I know that it's better to check
> > handle == NULL first, but it's the way some
Jordan Hubbard wrote:
> > After the crt changes the following piece of code, which worked previously,
> > gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> > "__register_frame_info' error message (yeah, I know that it's better to check
> > handle == NULL first, but it's the way some
> After the crt changes the following piece of code, which worked previously,
> gives a 'host: dlopen() failed: ./module.so: Undefined symbol
> "__register_frame_info' error message (yeah, I know that it's better to check
> handle == NULL first, but it's the way some apps work).
Huh! So that's w
19 matches
Mail list logo