Mail Archives: djgpp/2013/12/12/12:45:07
It's not as simple unfortunately, since the problems occurs on the
user's computer (which is not much djgpp-aware), I don't have much
freedom in debugging :/
I guess I will have to find some 386 machine to reproduce the exact
problem myself (IF it happens to be a CPU problem - this is only my
supposition so far, because it fails on one particular machine that
happens to be of an antique age).
Yes, the application is freely available (GNU/GPL) - it's the FreeDOS
package manager:
http://fdnpkg.sourceforge.net/
The latest source code is on the svn here (includes my i386 build of
watt32):
http://sourceforge.net/p/fdnpkg/code/HEAD/tree/
cheers,
Mateusz
On 12/12/2013 02:06 PM, Juan Manuel Guerrero wrote:
> Am Dienstag, 10. Dezember 2013 15:21:11 UTC+1 schrieb Mateusz Viste:
>> Hi there,
>>
>>
>>
>> I have some troubles using watt32 with DJGPP. I have two applications
>>
>> that work fine on a pentium-grade machine, but both crash on a i386 when
>>
>> the Watt32 lib is called.
>>
>>
>>
>> At first I though that it is because I was using a Watt32 lib compiled
>>
>> for Pentium (the Watt32 identification string was saying '(Pent)'), so
>>
>> then I added -march=i386 when compiling Watt32, unfortunately the
>>
>> applications are still crashing (with a SIGSEGV, General Protection
>>
>> Fault to be exact).
>>
>>
>>
>> The ident. string of the Watt32 build I am using now is this:
>>
>>
>>
>> Watt-32 (2.2.10), GNU-C 4.7.1 (386), djgpp 2.03, Dec 9 2013
>>
>>
>>
>> Am I alone with that problem? Or maybe - by any chance - anyone already
>>
>> noticed this and would be kind enough to point me in the right direction? :)
>>
>>
>>
>> Or - have anyone already seen Watt32 working on a 386-class PC?
>>
>>
>>
>> I should add that I don't have an i386 machine myself to test - the
>>
>> crashes have been reported to me by a user of my program, so I can't
>>
>> even reproduce the problem at will... This same user says that "wget,
>>
>> which relies on watTCP works perfectly fine on that machine" - so my
>>
>> next step will be to try to use WatTCP instead of Watt32, and see if it
>>
>> changes anything... but I am totally in the dark here.
>>
>>
>>
>> cheers,
>>
>> Mateusz
>
>
> I do not have a i386 to test the code. Your 386 version dies with SIGSEGV.
> Can you use bfdsymify to identify the code lines where it crashes?
> Is the application available that you try to compile?
>
> Regards,
> Juan M. Guerrero
>
- Raw text -