[fpc-pascal] fpc 2.0 and DOM/xml under OS/2 gives me "Process terminated by SIGSEGV"

Tomas Hajny XHajT03 at mbox.vol.cz
Fri Sep 9 23:35:36 CEST 2005


Date sent:      	Mon, 22 Aug 2005 20:01:38 +0200
From:           	Sřren Ager <sorenager at poboxes.com>
To:             	FPC-Pascal users discussions <fpc-pascal at lists.freepascal.org>
Subject:        	Re: [fpc-pascal] fpc 2.0 and DOM/xml under OS/2 gives me "Process
	terminated by SIGSEGV"
Send reply to:  	FPC-Pascal users discussions <fpc-pascal at lists.freepascal.org>
	<mailto:fpc-pascal-request at lists.freepascal.org?subject=unsubscribe>
	<mailto:fpc-pascal-request at lists.freepascal.org?subject=subscribe>


Hi Soren

> On 21-08-05 15.49, Sřren Ager wrote:
> 
> > Any ideas why?
> 
> Changing:
> 
>    DOMString = WideString;
>    DOMPChar = PWideChar;
> 
> to:
> 
>    DOMString = String;
>    DOMPChar = PChar;
> 
> in dom.pp fixed the problem. So I guess the WideString is broken in
> fpc for OS/2. Is there anyway to disable it so I don't run into this
> problem in the future?

Sorry, I had your e-mails (together with many other :-( ) in backlog 
from the time of my holiday...

For the first question - the best way to find exact place of SIGSEGV 
etc. in FPC programs under OS/2 is to compile everything with debug 
information (including RTL, FCL etc.) and either run them under 
(PM)GDB, or to open the generated core file in (PM)GDB afterwards and 
let it show you the exact location.

For the second one - I believe that it's a bug in common part of FPC 
RTL, in fact, that it crashes when WideString manager is not 
initialized. I didn't find time for implementing the OS/2 specific 
support for WideString, but it shouldn't crash nevertheless. I'd 
suggest to create a simplified test case and add it to our bug 
repository so it doesn't get forgotten.

Tomas




More information about the fpc-pascal mailing list