This is the mail archive of the cygwin-developers@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: 1.3.4 status?


----- Original Message -----
From: "Robert Collins" <robert.collins@itdomain.com.au>
To: "Jonathan Kamens" <jik@curl.com>
Cc: <cygwin-developers@cygwin.com>
Sent: Tuesday, October 23, 2001 10:41 AM
Subject: Re: 1.3.4 status?


> My 2c on this is that this could be a lot worse than a malloc issue...
> even though it is occuring at process exit.
>
> GCC optimisation can change the code substantially as you step up
                             machine
> layers.
>
> i.e. on common problem squid had was that a function ended up XOR'ing
         e
> variable foo with itself, before trying to use it! (Oh, it was _not_
> meant to be 0). That resulted in the *BSD's requiring special
configure
> lines to disable -O2 for that OS release, and yet another gcc version
> test in the configure script.
>
> So, I'd start of by hand checking the faulting line of assembly, to
see
                off
> that is *should* work if everything where normal, and then work back
       it
> through the stack trace doing the same thing. If you get past the
exit()
> stuff, then malloc is a thing to try. I'm not sure which is faster,
this
                                                         way
> is just my 2c.
>
> Rob

That was almost illegible - sorry!


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]