Re: pixmap/pixbuf confusion

[Michael Koppelman <michael aps umn edu>]

On Sunday, July 13, 2003, at 11:40 AM, Peter "Firefly" Lund wrote:

> Good thing other people's even humbler opinion prevail.
>
> Clearing it would be pointless (since one would want to override that
> before using the pixmap anyway), more code, and be inefficient.  
> Perhaps
> not by much but by some.  Most of all, it would be absolutely 
> pointless.

Probably true.

> That it did seem to grab the current window was just one way a "random"
> uninitialised buffer could turn out. Deal with it.

It didn't "seem" to grab the current window, it did grab the current 
window 100 times out of 100. You only notice a problem if you change 
the size of the pixmap to a size different from the window. From my 
limited data (and without looking at the source) it seems to me that 
the data is initialized with the current window. Otherwise it is a 
miraculous (but not impossible) coincidence that with a thousand other 
processes going it happened to use my window's old memory every time.

I did deal with it and it works fine. Uninitialized data creates bugs. 
I mistakenly assumed that the pixmap would be initialized with some 
rational value (such as the GC background value). Please forgive my 
imposition on your precious time.

Cheers,
Michael