Re: GTK+ 2.10.10 released

[Martyn Russell <martyn imendio com>]

Sergei Steshenko wrote:
> I think you are missing the point.
> 
> On my memory there have been two release announcements
> mentioning several memory leak related fixes.
> 
> On the other hand, gtk+ FAQ "5.28. Why does GTK+/GLib leak memory?
> claim gtk+ does not leak memory - the "It doesn't" part.
> 
> So I'm waiting ON the "It doesn't" part to be fixed.
> 
> Because, as release announcements show, it does.
> 
> Should I file a bug report against gtk+ website, specifically, against
> the FAQ 5.28 ?
> 
> As I wrote earlier, I would feel having moral right to write that a suite
> doesn't have memory leaks provided in, say, a year there has been no memory
> leaks confirmed bugs, and I would write the claim more cautiously anyway.

The point here is that a lot of people have written to mailing lists
saying that the toolkit leaks for them, etc, and it turns out that
actually, they are usually not using it correctly or don't understand
that "Both GLib and the C library (malloc implementation) will cache
allocated memory". There is a good reason for this, it has a performance
benefit.

I think that perhaps that part of the FAQ could be worded differently to
explain that there might be bugs in the toolkit, to say "It doesn't" is
quite a claim. So if I were you, yes, I would report a bug about the
documentation on the site and suggest a better wording for it.

-- 
Regards,
Martyn