summaryrefslogtreecommitdiff
path: root/sw
diff options
context:
space:
mode:
authorTor Lillqvist <tml@iki.fi>2013-08-19 10:32:34 +0300
committerTor Lillqvist <tml@iki.fi>2013-08-19 10:58:37 +0300
commitbd507a0d093b7d5239992440a40d0a7453b38207 (patch)
tree513dc0e1a79abd1a9cb74deb6638d677c29cdff1 /sw
parent08452f1208a2ecbf7c6f15a96efeedbe5210ec83 (diff)
fdo#39007: Brute force fix: Unlink a potential restorecount.plist file
It is not completely clear to me whether we really still might be leaving such files around. After all, we now in *two* places (at least in current master) tell our NSWindows to not be restorable... Norbert's code in AquaSalFrame::initWindowAndView() *and* Herbert Dürr's code in -[SalFrameWindow initWithSalFrame:]. On the other hand, Cocoa does seem to create the file and keep it around for a short while before removing it (in responce to our setRestorable:NO calls?), so there is a slight time window where if LO crashes, the file will be left around. Such a file might also be around from an older LO version, or manually planted there by somebody wanting to reproduce the bug... Of course, the *real* fix for this problem would be to make LO a *proper* Cocoa document centric application, that would use NSDocument, NSApplication etc like native applications are supposed to, and that *would* handle window restoration the Cocoa way. I.e., work with the system instead of against it. Change-Id: I9ed058130ddddf49cf0221d899bef3e2654589c7
Diffstat (limited to 'sw')
0 files changed, 0 insertions, 0 deletions