http://bugs.winehq.org/show_bug.cgi?id=5253
--- Comment #117 from Saulius K. saulius2@gmail.com 2010-08-07 02:03:31 --- Folks, lets stop reporting workardounds. There are plenty of them here already. Thanks:)
(In reply to comment #110 from Benjamin Horstman 2009-05-03)
Presumably this bug is caused by incomplete fonts in wine; if free alternatives can't be found perhaps it could at least be modified in a way that doesn't cause crashes?
Benjamin, the first your statement needs confirmation. In negative case the second statement would mean workaround for the bug. We need to replicate Windows behaviour instead.
Lets follow to the summary of Jesse Allen's investigation.
(In reply to comment #64 from Jesse Allen 2006-09-15)
Aha! I know why I need 26 more lines! For whatever reason, the font chosen makes *all* the geometry of the bitmaps scale to fit it and with the sserife.fon 26 of the screen's lines ends up off the screen, which would be a problem if not all of it fits on the buffer of the surface.
Observable using the opengl backend which won't crash btw.
Nice. So to behave like Windows we need (1) to know what font the game requested + what font was chosen by (various) Windows font substitution engines.
(2) The next step is to investigate whether Wine returns the same result (testcase is probably still absent from the test suite).
(3) If the results are same, then Wine font being chosen or font rendering code should be modified so the things don't scale up too much (additional testcases required).
(4) If the results differ, then Wine font substitution engine should be modified to choose the required fonts.
(5) Then if games still suffers the crash, we should return to step 3.
PS: code for test cases would be very nice regardless any yes-no switches here.