Hans Leidekker hans@it.vu.nl writes:
Fixes http://bugs.winehq.org/show_bug.cgi?id=1968 as well as a bunch of Valgrind warnings in the url test module.
-Hans
Changelog Make sure URL is terminated in InternetCrackUrlA.
The bug is in InternetCrackUrlW, it shouldn't require the string to be null-terminated.
On Monday 09 June 2008 17:09:52 Alexandre Julliard wrote:
The bug is in InternetCrackUrlW, it shouldn't require the string to be null-terminated.
Right, but I guess fixing that is no 1.0 material since there's at least 4 places where it assumes a null-terminated string, including a call to InternetCanonicalizeUrlW.
-Hans
On Tue, 10 Jun 2008, Hans Leidekker wrote:
On Monday 09 June 2008 17:09:52 Alexandre Julliard wrote:
The bug is in InternetCrackUrlW, it shouldn't require the string to be null-terminated.
Right, but I guess fixing that is no 1.0 material since there's at least 4 places where it assumes a null-terminated string, including a call to InternetCanonicalizeUrlW.
Why not? It's a bug fix and four places to fix sounds quite reasonable (>100 may not be). Maybe it's just me but I didn't feel like 1.0 would entail a deep deep freeze period where only critical/security bugs would be fixed.
Also I don't feel that the current state of Wine (or of the future Wine 1.0) is such a marvel of perfection that it justifies a deep freeze to make sure it's not going to get a tiny little scratch here or there.
So my understanding is that yeah, we're not going to introduce any big new functionality (DirectX 10 is out), or rewrite core parts of Wine (the DIB engine is out), or put the code through a prettifier, or switch to a new version control system (we're using Git already anyway). But this sounds like pretty regular bug fixing to me (unless I got the scope wrong).
Francois Gouget fgouget@free.fr writes:
Why not? It's a bug fix and four places to fix sounds quite reasonable (>100 may not be). Maybe it's just me but I didn't feel like 1.0 would entail a deep deep freeze period where only critical/security bugs would be fixed.
At this point the freeze is pretty strict, since the plan is to release next week.
Also I don't feel that the current state of Wine (or of the future Wine 1.0) is such a marvel of perfection that it justifies a deep freeze to make sure it's not going to get a tiny little scratch here or there.
Tiny scratches are not a problem, the problem is that in many cases you can't really know if a patch is going to cause only a tiny scratch or a major failure; the only way to know is to put it in and wait for the users' screams, and that's not a risk we want to take at this point.
So patches that are not obviously safe will have to wait; that sort of fix is certainly a candidate for inclusion in 1.0.1, once it has cooked for a bit in the post-1.0 devel branch.