Thread: Program crash on Mac

Program crash on Mac

From
"Adam H. Pendleton"
Date:
I was sure that I sent this to the list, but maybe it got lost in all
the list confusion.  On the Mac, when adding a new server, I click the
button or select the menu item, the add new server dialog appears.  No
matter what I enter, and it doesn't matter if I click Cancel or OK, I
get a segfault with this error:

*** malloc [23866]: Deallocation of a pointer not malloced: 0xbffff0c0;
This could
be a double free(), or free() called with the middle of an allocated
block; Try setting
environment variable MallocHelp to see tools to help debug

Having changed all the wxLogInfo() calls to wxLogDebug() in the relevant
areas of the program (so they would actually print out, since settings
don't seem the be saved on the Mac), this is the last thing printed
before the error:

13:52:59 Debug: Destroying a connect dialog

This piece of output is located in the frmConnect destructor, all by
itself.  So I can only assume that the offending code is either after
the destructor is called, or perhaps is located in the parent object of
frmConnect.  Wherever it is, my head hurts from debugging, and besides,
I have real work to do (meaning work that pays the bills -- no more real
than that :-).  Maybe someone else can take a look at this.  If not, I
will start looking at it again tomorrow.

ahp


Re: Program crash on Mac

From
"Dave Page"
Date:

> -----Original Message-----
> From: Adam H. Pendleton [mailto:fmonkey@fmonkey.net]
> Sent: 20 May 2003 19:00
> To: pgadmin-hackers@postgresql.org
> Subject: [pgadmin-hackers] Program crash on Mac
>
>
> This piece of output is located in the frmConnect destructor, all by
> itself.  So I can only assume that the offending code is either after
> the destructor is called, or perhaps is located in the parent
> object of
> frmConnect.  Wherever it is, my head hurts from debugging,
> and besides,
> I have real work to do (meaning work that pays the bills --
> no more real
> than that :-).  Maybe someone else can take a look at this.
> If not, I
> will start looking at it again tomorrow.

Hi Adam,

I'll happily try to debug it (tomorrow probably), but cannot do much
unless I can recreate the problem which I can't not having a Mac. Any
chance of VNC access to yours?

Regards, Dave.


Re: Program crash on Mac

From
Jean-Michel POURE
Date:
Le Mardi 20 Mai 2003 20:24, Dave Page a écrit :
> I'll happily try to debug it (tomorrow probably), but cannot do much
> unless I can recreate the problem which I can't not having a Mac. Any
> chance of VNC access to yours?

Hi Dave,

I have a MacOsX 10.2 waiting for you. The machine is running, unfortunately I
did not install OpenSSH and VNC yet.

Adam, what do you recommand for remote access : shall I install Fink or native
packages. Any ideas? I won't be able to spend time installing before
thursday.

Do you think a MacOs X public machine is still needed? With Adam debugging
like hell, I don't know if it is still worth setting a public machine. Maybe
it can be used to generate daily MacOs X packages from CVS. What do you
think?

Cheers,
Jean-Michel

Re: Program crash on Mac

From
"Adam H. Pendleton"
Date:
Jean-Michel POURE wrote:

>I have a MacOsX 10.2 waiting for you. The machine is running, unfortunately I
>did not install OpenSSH and VNC yet.
>
>Adam, what do you recommand for remote access : shall I install Fink or native
>packages. Any ideas? I won't be able to spend time installing before
>thursday.
>
>Do you think a MacOs X public machine is still needed? With Adam debugging
>like hell, I don't know if it is still worth setting a public machine. Maybe
>it can be used to generate daily MacOs X packages from CVS. What do you
>think?
>
>
>
I for one would vote for a public MacOS X box, if for no other reason
than it would allow more than just one person (me!) to debug.  It might
also make it easier to test new features on a Mac platform as well, so
that we are sure they work before we commit them.  As for remote access
packages, I have copies of Timbuktu for both Windows and Mac that would
be good for remote access.  They are native packages.  Personally I have
avoided using fink only because I prefer to develop in as native an
enviroment as possible.  It keeps the amount of "well it works on my
Mac" problems to a minumum.

I would be more than happy to provide you the software you need, and I
can also perform any installation you need done, if you are busy.

ahp



Re: Program crash on Mac

From
"Dave Page"
Date:

> -----Original Message-----
> From: Jean-Michel POURE [mailto:jm.poure@freesurf.fr]
> Sent: 20 May 2003 19:53
> To: pgadmin-hackers@postgresql.org
> Subject: Re: [pgadmin-hackers] Program crash on Mac
>
>
> Le Mardi 20 Mai 2003 20:24, Dave Page a écrit :
> > I'll happily try to debug it (tomorrow probably), but
> cannot do much
> > unless I can recreate the problem which I can't not having
> a Mac. Any
> > chance of VNC access to yours?
>
> Hi Dave,
>
> I have a MacOsX 10.2 waiting for you. The machine is running,
> unfortunately I
> did not install OpenSSH and VNC yet.

Cool, you got the upgrade then?

> Do you think a MacOs X public machine is still needed? With
> Adam debugging
> like hell, I don't know if it is still worth setting a public
> machine.

I would sure like to play with an OS X box anyway - even if it's just to take a look. Actually Adam's machine isn't
appropriatefor net access anyway, so it would be useful for me to help out with any debugging. 

> Maybe
> it can be used to generate daily MacOs X packages from CVS.
> What do you
> think?

That's certainly not a bad idea. There is a build-snapshot script in CVS already that would do most of the work, and
rsyncor ftp could be used to copy the results onto snake.pgadmin.org. 

Regards, Dave.