Mailing-List:	contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender:	cygwin-developers-owner AT sourceware DOT cygnus DOT com
Delivered-To:	mailing list cygwin-developers AT sourceware DOT cygnus DOT com
From:	fortinj AT ibm DOT net
Message-ID:	<37267705.6D7987A4@ibm.net>
Date:	Tue, 27 Apr 1999 22:48:37 -0400
Reply-To:	fortinj AT ibm DOT net
X-Sender:	"" <@smtp-gw01.ny.us.ibm.net> (Unverified)
X-Mailer:	Mozilla 4.5 [en]C-gatewaynet (Win98; U)
X-Accept-Language:	en
MIME-Version:	1.0
To:	Chris Faylor <cgf AT cygnus DOT com>
CC:	DJ Delorie <dj AT delorie DOT com>, cygwin-developers AT sourceware DOT cygnus DOT com
Subject:	Re: cygwin1.dll startup and GDB
References:	<37265A6E DOT 813F8ED8 AT ibm DOT net> <199904280131 DOT VAA28701 AT envy DOT delorie DOT com> <37266BBC DOT F8F24A3A AT ibm DOT net> <199904280206 DOT WAA05486 AT envy DOT delorie DOT com> <372672D8 DOT 7CFA4EF1 AT ibm DOT net> <19990427223431 DOT A28362 AT cygnus DOT com>

Chris Faylor wrote:
> 
> On Tue, Apr 27, 1999 at 10:30:48PM -0400, fortinj AT ibm DOT net wrote:
> >DJ Delorie wrote:
> >>
> >> >       Either way, mapname will be NULL since NULL is being passed as the
> >> > first parameter to open_shared.
> >>
> >> Not always.  When NULL is passed, we musn't call OpenFileMapping(),
> >> because it doesn't fail gracefully when passed NULL.
> >>
> >
> >Sorry, I meant in the specific case I had described later where NULL
> >was passed.
> 
> Maybe we're agreeing here, but I just wanted to point out again that
> DJ's analysis of the situation was correct (not a surprise).  The
> problem was that OpenFileMapping should never be called when name ==
> NULL.  This was a bug that my recent change introduced.  Prior to my
> change, that segment of code never saw a name == NULL.  After my change,
> this segment of code was *validly* passed a NULL but it was not dealing
> with it the way I intended.
> 
> Anyway, that means that either his change or mine should rectify the
> problem.
> 
> I'm not sure why you are arguing the point rather than trying his
> suggestion.
> 
> cgf

Actually, I agree totally.  In my post I wasn't sure if I had made it
clear I was talking about a specific case, and DJ seemed to be
explaining the general case.  

Besides, asking a question, valid or not, usually increases the
knowledge of the person asking.  There are many different pieces to this
project, and I felt it was a reasonable question.  I  don't like to
blindly change code without at least a hazy understanding of what's
going on.

I made the changes you suggested in your other post, and gdb now works
as expected.


John Fortin

- Raw text -