On Mon, Sep 3, 2018 at 7:40 AM, Ivan Zhakov <i...@visualsvn.com> wrote:
> On 25 August 2018 at 16:14, Rainer Jung <rainer.j...@kippdata.de> wrote: > > > > There's a list of changes done by Ivan Zhakov starting with 1785072 in > early > > 2017 to improve Win32 file I/O performance. Maybe Ivan can tell, whether > > they can and should be backported to 1.7 (and 1.6?). > > > In my opinion, r1785072 should not be backported to 1.7 and 1.6, as it > is not bug fix. As far I remember these win32 file I/O performance > changes are quite dependend on each other and some refactorings made > on trunk. I think it's not good idea to backport them one-by-one. > Better way to release APR trunk at some point, instead of backporting > non big fix changes. The issue is that you slipped one bugfix into the optimization changeset, and this was addressed by jfclere in a different manner since 1.6.3. I've reverted to the conditions expressed in trunk and current readwrite.c which you introduced in your patch to buffer.c (both in file_io/win32). With that change, I'm aware of no other bug fixes we are blocking on, so I'll proceed this afternoon with the tag and roll of apr 1.6.4, unless someone raises a red flag in the next hour or two.