https://bugs.winehq.org/show_bug.cgi?id=57558
Bug ID: 57558 Summary: joy.cpl xinput joysticks circles are cropped by 1px on the bottom Product: Wine Version: unspecified Hardware: x86-64 OS: Linux Status: NEW Keywords: regression Severity: normal Priority: P2 Component: win32u Assignee: wine-bugs@winehq.org Reporter: rbernon@codeweavers.com Regression SHA1: 5924ab4c5155d97f4b23a33e97258c1ec382ea02 Distribution: ---
Created attachment 77637 --> https://bugs.winehq.org/attachment.cgi?id=77637 Patch to build a winejoy application to compare on Windows.
As described in the title, the circles are supposed to fit exactly in some bounding rectangle and they are now cropped by 1 pixel on the bottom side. This does not happen on windows with the same code.
This can be reproduced on Wine with `wine control joy.cpl` with a XInput-compatible gamepad plugged in, or with the attached patch.
Because joy.cpl gets overridden on Windows, I'm attaching a patch that builds a dedicated winejoy.exe program that does otherwise the same thing. Included is a hack that fakes XInput gamepad to be plugged in.
The blamed commit is the same as bug 57306, but that commit seems to be about crashes and memory corruption which doesn't look like the same problem.
https://bugs.winehq.org/show_bug.cgi?id=57558
--- Comment #1 from Rafał Mużyło galtgendo@o2.pl --- Bug 57306 had been (and - as of last check - still is) about memory corruption and asserts. The commit itself has been responsible for quite a few more benign (non-crash) regressions. Some were fixed, some still aren't.
(currently, besides the crash, one that grabbed a little of my interest is that drawing in mspaintxp, where straight edges of round rect/elipsis aren't being erased while being shrunk while drawing them; though I don't have 9.18 at hand to check if that is a regression though as I asked, the last answer *seems* to suggest they are; also, sometimes they aren't drawn in the first place...I suspect some integer math issues on the later part, the former is either the same or some incorrect clipping) (since quite a few years back I've played with porting from qt3 to qt4 kseg for my own needs (even got something working in the end) I'm quite aware how much seemingly minor changes may matter (like changing interpretation of the width of filled vs outline rectangle by 1px))
https://bugs.winehq.org/show_bug.cgi?id=57558
--- Comment #2 from Rafał Mużyło galtgendo@o2.pl --- In fact, see the end of discussion in bug 57296. About a coin toss if it's related.
https://bugs.winehq.org/show_bug.cgi?id=57558
--- Comment #3 from Hans Leidekker hans@meelstraat.net --- Bug 57306 points to the same regression commit.