Re: Re: Proposed Windows-specific change: Enable crash dumps (like core files)

On 19/12/2010 7:51 PM, Magnus Hagander wrote:

>> Great. I pulled the latest from your git tree, tested that, and got much
>> better results. Crashdump size is back to what I expected. In my test code,
>> fcinfo->args and fcinfo->argnull can be examined without problems.
>> Backtraces look good; see below. It seems to be including backend private
>> memory again now. Thanks _very_ much for your work on this.
>
> Ok, great. I think that leaves us at least complete enough to commit -
> we can always improve things further as we get some more real world
> testing.
>
>
>> fcinfo->flinfo is still inaccessible, but I suspect it's in shared memory,
>> as it's at 0x00000135 . Ditto fcinfo->resultinfo and fcinfo->context.
>
> Hmm. Not sure why those would be in shared memory, that seems strange.

OK, I'll have to do some more digging on that, then. I'm getting on a
plane in about 2 hours, but will be bringing Visual Studio snapshots, a
postgres git tree, an XP vm, etc with me, so time permitting I should be
able to keep on with this.

>> Anyway, here's an example of the backtraces I'm currently getting. They're
>> clearly missing some parameters (in shm? Unsure) but provide source
>> file+line, argument values where resolvable, and the call stack its self.
>> Locals are accessible at all levels of the stack when you go poking around
>> in windbg.
>
> Yeah, they're still very useful. Is that a release or debug build?

That's a release build. I'm intentionally testing with release builds
because I want something that's useful on real-world end-user systems,
and I'm aware that few Windows users will build Pg for themselves.

(That said, the Perl-based build scripts are some of the least painful
build tools I've ever worked with on Windows. Only CMake can rival them
for low-pain Windows compilation.)

--
Craig Ringer

Tech-related writing at http://soapyfrogs.blogspot.com/