a) Possible, but unlikely, I would think. I did test in 32-bit 0.7, 32-bit
0.8RC1, and 64-bit 0.8RC1, and they all succeeded.
b) Again, possible, but I would think unlikely. Two 32-bit Windows versions
suceeded.
c) Definitely possible. The possibility of an over/underflow here under
certain situations (as I suggested in my prior e-mail). However, if I
recall, someone tried it on their Linux machine and encountered the same
error, which is what leads me to suspect it is:

d) A platform specific bug. The Windows and OSX/Linux code have different
paths for managing the enblend-specific swap file, for swapping data in and
out. As I mentioned in my prior e-mail, based on the strace, from where the
code was failing it looked like an over/underflow.

If someone can point to me to the most-up-to-date version of building on OS
X, I'll give it a go on my Intel OS X 10.5 box here and see if I can't
attach some gdb watchpoints to track it down. It may take a few days to get
everything built though, fwiw.

Regarding malloc on OS X, I too am suspicious of it/realloc, just knowing
the trouble it's caused people, but if it's repro-ing on Linux as well, then
I don't think it's something directly specific to OS X.

> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On
> Behalf Of George Row
> Sent: Sunday, May 17, 2009 7:54 AM
> To: Harry van der Wolf
> Cc: [email protected]
> Subject: Re: Again: [hugin-ptx] Alpha Channel Masking
> 
> 
> Harry,
> 
> Sorry to hear about all your disasters - they serve as a reminder to
> the rest of us (well me anyway)  to make regular back up copies.
> 
> Thanks for all the effort that you and others have put into this
> problem.
> 
> It seems to me that there are three possible explanations here.
> 
> The fact that the 64bit build that you used was from version 0.7
> could mean  that:
> 
>  a) the error is triggered by a bug that was introduced into the
> enblend code since version 0.7
> 
>  b)  this particular project makes such demands for masking that they
> can only be serviced in a 64 bit environment
> 
>  c) there is a bug that triggers in a 32 bit environment that does not
> trigger in a 64 bit environment
> 
>  (The fact that it worked for Ryan on a 64 bit windows system lends
> support to b) or c)  or perhaps suggests that a) is a platform
> specific bug?)
> 
> I would guess that we could rule out option a) by trying the project
> with the Hugin 0.7 version in a 32 bit environment - would it run in
> 32 bit mode if I used it on my G5 Mac under Mac OS X 10.4.11?    - If
> so I will give it a try.
> 
> all the best
> 
> George
> PS Sorry about the false start on this message - it seems that some
> typing mistake of mine is a Google mail shortcut for "send"
> 
> 2009/5/17 George Row <[email protected]>:
> > Harry,
> > Sorry to hear about all your disasters - it serves as a reminder to
> > the rest of us to make regular back up copies.
> >
> > Thanks for all teh effort that you and others have put into this
> problem.
> >
> > It seems to me that there are two possible theories here.
> >
> > The fact that the 64bit version that you used was from version 0.7
> > could mean  that either:
> >  the error is triggered by a bug that was introduced in some code
> since
> >
> >
> > 2009/5/17 Harry van der Wolf <[email protected]>:
> >> Hi All,
> >>
> >> I also tried with George's set. I do have an Intel core duo which is
> a 64bit
> >> system (and running Leopard). It only has 2GB memory. I used the 0.7
> >> official version which is a 32/64 bit version which runs in 64 bit
> on my
> >> system.
> >>
> >> It took some more time than expected. So therefore an intermediate
> mail with
> >> feedback.
> >> Evening 1: just before starting the run my external HD crashed where
> I had
> >> downloaded George's 2GB source images (and more important: my own
> data).
> >> After opening the external casing I could smell that it was
> completely
> >> fried. Connecting it to the IDE of a pc in my home gave the same
> result:
> >> nothing :-(
> >> Evening 2: I let it run overnight. In the morning it turned out that
> my
> >> macbook had completely shutdown somewhere at 80% (I guess). I even
> needed to
> >> remove the battery and then reinsert it to make it "come to live"
> again.
> >>
> >> Night 3: It took3 hours but it succesfully finished.
> >>
> >> I will try another run tonight or tomorrow night with a standard 32
> bit
> >> version.
> >>
> >>
> >> Harry
> >>
> >
> 
> 

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"hugin and other free panoramic software" group.
A list of frequently asked questions is available at: 
http://wiki.panotools.org/Hugin_FAQ
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at http://groups.google.com/group/hugin-ptx
-~----------~----~----~----~------~----~------~--~---

Reply via email to