When overwriting a package the sequence is: remove old, add new package.
On remove we add package to ToBeDeactivated set. This is ok.
On adding new package there is a check if package exists or not. On true
return. But package still in ToBeDeactivated set.
Fixes#10038
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
Signed-off-by: Axel Dörfler <axeld@pinc-software.de>
We can't delete old volume state while it's still active and can be
accessed until reboot.
Fixes#10898
Signed-off-by: Augustin Cavalier <waddlesplash@gmail.com>
Signed-off-by: Axel Dörfler <axeld@pinc-software.de>
The asynchronous listener is meant to be used when the functions are
meant to be called from the main thread (or another looper the listener
is explicitly atached to).
For this to work, the request should be attached to
listener->SynchronousListener(), which was not the case here. Assume it
is fine for the functions to be called from the HTTP thread directly
then (if I'm wrong, revert this and attach the synchronous listener to
the http request, however wihout further changes it will just move the
execution of the callbacks to the main thread, which may not be the
right place either).
This fixes a crash when exiting WebKit while playing a youtube video
(somehow the handler isn't cleanly deleted and removed from the main
looper there - going without a handler at all fixes this).
* Hasn't been used for quite some time
* Everything was ported over to a new ATA stack
some time ago.
* No huge regressions were seen from the new ATA
stack.
URL input is considered locked when it is focused and original text was
changed. Pressing ESC will now change URL input's text to the current URL.
Signed-off-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
This makes it a lot easier to use and test pc_serial, even on machines
with a single serial port (COM2 is getting even more rare than COM1
these days).
* built in non-strict mode: dovecot, qupzilla, r
* build failed: intltool (dep of gcab and transmission), python_html2text, vlc,
cython (dep of pyenet)
* not built because of an unsafe source: qutim
* removed compatibility packages for giflib, icu, tiff, libpng, libwebp, libbluray,
libpcre, libvpx, ncurses, readline.
* built in non-strict mode: dovecot, qupzilla, r
* build failed: intltool (dep of gcab and transmission), python_html2text, vlc,
cython (dep of pyenet)
* not built because of an unsafe source: qutim
* removed compatibility packages for giflib, icu, tiff, libpng, libwebp, libbluray,
libpcre, libvpx, ncurses, readline.
* built in non-strict mode: dovecot, qupzilla, r
* build failed: intltool (dep of gcab and transmission), python_html2text, vlc,
cython (dep of pyenet)
* not built because of an unsafe source: qutim
* removed compatibility packages for giflib, icu, tiff, libpng, libwebp, libbluray,
libpcre, libvpx, ncurses, readline.
* built in non-strict mode: dovecot, qupzilla, r
* build failed: intltool (dep of gcab and transmission), python_html2text, vlc,
cython (dep of pyenet)
* not built because of an unsafe source: qutim
* removed compatibility packages for giflib, icu, tiff, libpng, libwebp, libbluray,
libpcre, libvpx, ncurses, readline.
* The app_server isn't designed to support two fallback drivers, so
on systems using UEFI to boot, the framebuffer driver will often
win when other drivers would likely work on those systems.
Also remove the BMessageFilter EscapeFilter and use B_CLOSE_ON_ESCAPE
in the window constructor instead.
I moved fExpressionTextControl->MakeFocus(true); down to when the
window is Lock()ed. Otherwise the text control refuses to be focused...
In the current state, opening a link in a new tab will cause URL input to
become empty. This fixes the issue by changing the way locking works and
implementing a lock timeout. The text passed to SetText() will be instead
stored as previous text, so if the user presses ESC after editing the URL,
URL input will update to the latest sent text (i. e. current URL). Also
fixed two lines the style checker was complaining about.
Fixes#13548.
Signed-off-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Thus, BeOS compatibility is preserved (and there is no risk of
breaking GCC5<->GCC2 interoperation on hybrid builds.)
This commit only makes the actual change, the build fixes are
in the next commit.
Both the user-mode syscalls.h and the kernel-mode one define it
as an int32, not a time_t, and as it's a timezone offset not
an actual time, there's no reason it needs to be one.
