From: Pete B. <pb...@gm...> - 2010-08-17 10:26:17
|
On 2010.08.17 03:26, Michael Plante wrote: > Xiaofan Chen wrote: >>> That being said, I have no issues with the patches or your methods >>> to create the binary snapshots. It is kind of test binary only and not >>> considered to be official binary anyway. > > Oh, I missed this. Is this not being submitted as a patch? Right now, the whole Windows backend is experimental, so by extension, I'd say any Windows binary we produce is. But of course, then there's "official" experimental and non-official experimental. As I consider that Peter or Daniel would be the ones generating official binary snapshots (and preferably those would need to include Linux and Darwin binaries as well), I don't see the snapshots on my page as "official". However, in the absence of anything better, they are currently fitting that role. Now, I'm not going to force anybody to use multilib to generate MinGW binaries, so whether this patch is needed for the official binary production is another story. But I do consider that some of our developer-users might be recompiling MinGW-w64 from source (because that's the only way to get multilib right now, and there are of course all sorts of other good reason to recompile MinGW-w64), and if they do that, the patch is very much needed. Thus, depending on how fast MinGW-w64 fix their headers / how soon we can get the rest of the Windows backend integrated, I'm seeing it as "official until further notice", and will submit it as a patch if needed (but after integration is over). Regards, /Pete PS: I'm planning to add usbi_err() lines to the macros (which is something I should have done from the start) and limit the scope of the workaround later on today. |