It's been rumoured that Jonathan Corbet said:
>
> I ran gnucash through xmon for a bit this evening. The immediate result is
> that I didn't find any obvious leaks by looking at the X protocol. Every
> CreatePixmap was followed by a FreePixmap, every CreateGC by a FreeGC, and
> so on. It looks like it's all right.
>
> I did notice a tremendous number of such calls. Simply moving the pointer
> across the register window can generate dozens of them - with no visible
> changes in the register window itself. Something is watching for pointer
> motion and creating/freeing lots of pixmaps in response, to no evident
> purpose.
Ouch! This might be something we should aks our freinds at helixcode to
look into.
It *might* explain the memory leak: its not really a leak, its
fragmentation. memory might get filled with lots of little chunks,
most of which are empty, none of which are big enough to hold the
next request, ergo, sbrk() to get more memory. But I thought that
the x server used separate pools for gc's etc precisely in order to
avoid this problem ...
--linas
--
Gnucash Developer's List
To unsubscribe send empty email to: [EMAIL PROTECTED]
