Board index » kylix » Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Re: Workaround for kylix de{*word*81} on kernels > 2.6.10


2006-02-17 11:30:00 PM
kylix0
Yep great! Thanks A LOT!!!!!!
2.6.13-15.7
 
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Quote
This works for us under both kernel 2.6.12 and 2.6.16-rc3. The de{*word*81}
doesn't hang anymore and works just fine.
works fine here 2.6.12-1.1378_FC3smp :)
tnx guys :)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Works greate under 2.6.13-15.7-smp (SUSE 10.0)
- Qt2Clx works (original)
- Qt3Clx works
- Libc.system() (pipe-problem) has no problem
- Debugging from executable through SharedObject and back works
- Debugging from exe through Package and back works
What do we need next?
- Binding Kylix against newer WineLib :-)
Regards,
Andreas Hausladen
 

{smallsort}

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Den Jean wrote:
Quote
PS: export LD_ASSUME_KERNEL=2.4.1 should not be forgotten
What Kylix developer hasn't exported this :-)
--
Regards,
Andreas Hausladen
(www.kylix-patch.de.vu - unofficial Kylix 3 patches)
(andy.jgknet.de/blog)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Paulo Marques schrubte am Freitag 17 Februar 2006 16:12:
Quote
Thanks for any feedback,
Doesnt tested yet, nevertheless i want to thank you very very much for your
work. Its very good to have some motivated and skilled people around.
kind regards, Jens
--
Die oben angegebene Adresse ist OK, wird aber nicht gelesen. Wenn direkter
Mailkontakt gewünscht wird, bitte spamtrap gegen jens<punkt>nixdorf
austauschen. Danke.
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Quote
I would like to know if this works for everyone, or if there are still
problems with this approach with some distributions.
works fine on 2.6.12-12mdk (Mandrake/Mandriva 2006.0)
Thank you very much !!!
kind regards,
Den Jean
PS: export LD_ASSUME_KERNEL=2.4.1 should not be forgotten
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Andreas Hausladen wrote:
Quote
Works greate under 2.6.13-15.7-smp (SUSE 10.0)

- Qt2Clx works (original)
- Qt3Clx works
- Libc.system() (pipe-problem) has no problem
- Debugging from executable through SharedObject and back works
- Debugging from exe through Package and back works


What do we need next?
- Binding Kylix against newer WineLib :-)
and bplvisualclx-6.9 against qt3 ;)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Very great news ! Many thanks !
is the C++ part works ??
Regards,
Lionel
"Paulo Marques" < XXXX@XXXXX.COM >a écrit dans le message de news:
XXXX@XXXXX.COM ...
Quote

Hi, all

Me and Paulo Costa have been trying to find out why the kylix integrated
de{*word*81} hangs with recent kernels. You can follow the thread on LKML
here:

lkml.org/lkml/2006/2/14/419

We've now been able to make the de{*word*81} work with any recent kernel (I
hope) using a stub that intercepts system calls made by the de{*word*81}, so
that it simulates the old kernel behavior.

I've posted the file with the stub to borland.public.attachments.

To make the de{*word*81} work:

- unzip the stub file
- place the ptrace_interposer.so in /usr/lib (for instance)
- do export "LD_PRELOAD=/usr/lib/ptrace_interposer.so" before calling
"startdelphi"

This works for us under both kernel 2.6.12 and 2.6.16-rc3. The de{*word*81}
doesn't hang anymore and works just fine.

I would like to know if this works for everyone, or if there are still
problems with this approach with some distributions.

Thanks for any feedback,

--
Paulo Marques - www.grupopie.com
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Paulo Marques wrote:
Quote
I would like to know if this works for everyone, or if there are still
problems with this approach with some distributions.

The de{*word*81} works fine,
but code browsing (ctr-clicking on a unit or procedure)
does not work on new kernels.
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Lionel Reynaud wrote:
Quote
is the C++ part works ??
Compiler: Yes
Linker: On my system it crashes.
--
Regards,
Andreas Hausladen
(www.kylix-patch.de.vu - unofficial Kylix 3 patches)
(andy.jgknet.de/blog)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Hi Andreas,
Quote
>is the C++ part works ??

Compiler: Yes
Linker: On my system it crashes.

So we always have trouble with ilink. I read somewhere that you try to use
ldd in place of ilink. Did you find a good issue ?
Regards,
Lionel
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Lionel Reynaud wrote:
Quote
So we always have trouble with ilink. I read somewhere that you try to
use ldd in place of ilink. Did you find a good issue ?
I only wrote that it meight be possible to use ldd but I haven't tried it,
yet.
--
Regards,
Andreas Hausladen
(www.kylix-patch.de.vu - unofficial Kylix 3 patches)
(andy.jgknet.de/blog)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Quote
>So we always have trouble with ilink. I read somewhere that you try to
>use ldd in place of ilink. Did you find a good issue ?

I only wrote that it meight be possible to use ldd but I haven't tried it,
yet.


One "d" too many, right?
ldd - print shared library dependencies
ld - the GNU linker
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

theo wrote:
Quote
ldd - print shared library dependencies
ld - the GNU linker
And I always asked me why the printer started to output strange letters
when I wanted to link an application :-)
--
Regards,
Andreas Hausladen
(www.kylix-patch.de.vu - unofficial Kylix 3 patches)
(andy.jgknet.de/blog)
 

Re:Re: Workaround for kylix de{*word*81} on kernels > 2.6.10

Quote

The de{*word*81} works fine,
but code browsing (ctr-clicking on a unit or procedure)
does not work on new kernels.


?? I can't confirm this