On Sun, 2018-08-26 at 17:59 +0200, Markus Koschany wrote: > On Thu, 23 Aug 2018 17:56:01 +0300 Uoti Urpala <uoti.urp...@pp1.inet.fi> > wrote: > > Neverball automatically saves a replay of the latest run on disk while > > playing. In the Debian package, the binary constantly calls fsync() on > > this file, which very seriously degrades performance. If the issue is > > not obvious when trying to reproduce, try on a machine with as slow a > > spinning disk as possible. > > > > This is likely a libphysfs issue. Neverball code contains no direct > > fsync() calls. Neverball upstream has changed the default filesystem > > backend to be stdio instead of libphysfs, and current upstream code > > (with no physfs) does not display the problem.
> Why did you report this bug against neverball instead of libphysfs when > all the issues you describe are related to libphysfs? Neverball is broken so at least a bug blocked by a physfs bug would be appropriate in any case, and I assume that the easiest way to fix the problem would be to change the Neverball package to stop using libphysfs.
