Previously, HdsSchema directories were "built" with the Clean action and
manually created in their respective commands. This patch creates the
directory as a separate target to be cleaned. The MkDir1 action is used
directly since $(1) is already rooted and the HdsSchema actions use the
ungristed directory as a dependency.
This also fixes a build failure in Ham since it does not allow multiple
targets as part of 'together' actions.
Change-Id: Ibc818375477b5f1712ac7c7cec8a1f8812f62b65
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5630
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Reviewed-by: Stephan Aßmus <superstippi@gmx.de>
This reverts commit 7fe1087ed4.
Reason for revert: could cause underflow
Change-Id: Ib2263f3ab17ae377b0565d020410bae616385cfe
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5577
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This is needed on ARMv7 as inode number is stored using STRD instruction
which can result in alignment fault even when SCTLR.A bit is disabled.
see ARMARM section A3.2.1, Unaligned data access
Change-Id: Ib17bf3f78e84ad7987d8158426bc9b27ded28a8e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5623
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Fixed bug on reading files inside Haiku using CAT command on shell
Fixed bug on reading small directories inside Haiku
Node directories can now be read fine, which fixes#17839
LongBlock verification is now through common VerifyHeader function
Change-Id: I1eb651bd084580c86d0facb919825e94488e43d2
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5546
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
* 90% of our logs start with some context on what is
generating a log message (thing:)
* Things following this logging model however do "thing [time"
which is inconsistent
* Being consistent will allow us to start scanning logs in a
smart way and try to analyze patterns of what is throwing
errors, etc in an automatic way.
aka /^(service/driver/etc)\:/
Change-Id: I1ef2df4f17f70f858a485554a4e8a3f87f1a69c8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5568
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Oscar Lesta <oscar.lesta@gmail.com>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
This is for BeOS application compatibility. BeOS apps expect the label
to be drawn a few pixels higher and may not draw correctly otherwise.
Font height / 6 scales with font size so that the label moves up
proportionally at larger font sizes e.g. 21pt. 1/6 is a better offset
for ArtPaint Global settings and DVDRip 1.0.
Change-Id: I6d852111d75a97f34539f6e94c2e2408b6c15808
Reviewed-on: https://review.haiku-os.org/c/haiku/+/3017
Reviewed-by: John Scipione <jscipione@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Header files need to be included into a source file, not depended on.
Using depends changes *.cpp from a source file to a generated file. Jam
has a bug in dependency resolution that allows this to work, but this
breaks in Ham.
Change-Id: I584c719d353adbfdb68ddad5d05e7085f635ae28
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5625
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Stephan Aßmus <superstippi@gmx.de>
Instead of checking against the current block's DSO generation,
the DSO that the operator[] was invoked on was being checked instead.
This could lead to spurious deletions of TLS data in the case where
the generation count changed between accesses of the TLS variable
(currently that only happens when modules are unloaded.)
May fix#17896, but I did not have a reliable way to reproduce that.
Change-Id: Ie1ee51c1fbbe638918fd5a010047a2b56d481f2b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5604
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
* Reduce default search scope to: "name", "summary", and "provides".
* Add a new "--search-scope=<scope>" option that allows to either limit
the search to packages names (-s name), or to replicate the previous
behavior of searching everywhere in the package info (-s full).
Motivation for the change:
The ever increasing number of packages available for Haiku was impacting
pkgman search usefulness, as it easily returns too many packages (what a
wonderful problem to have :-D).
Limiting the default search scope helps with that, and the old behavior
is kept "just an option away".
Change-Id: I5b456b90137237134eee7ca7ee501bf8e3767440
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5616
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
* Disable BMenuField unless a sound event is selected.
* Changes in the MenuItem enable/disable the play button as necessary.
(previously only changes in the EventList did that).
* Enable the play button when BFilePanel is closed (after selecting valid file).
Fixes the (still reproducible) UI/UX portion of #17150.
Change-Id: I5b5d00cc7fce00a0d0f908bbaff883169f38f855
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5613
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Quite a few source files were missing when collecting strings
for translation. Just do the catalogs for all applicationSources.
Change-Id: I452bed61b50d5693b46d63cf5773b49f75ef2aef
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5622
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Introduce common page fault handler so we don't have to do code duplication.
Change-Id: Ib9fa6a3d38e137e43911a24a3405c7c3f1de5908
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5619
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Based off of tray padding instead of hard-coding 12. Calculate tray
padding once in AttachedToWindow() and use that value everywhere.
Create some class float member variables to keep track of these values.
Fixes#17923
Change-Id: I3e767fdc8d2775e5f0dfff3275aeb0b686a8ea76
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5610
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
* set PXN for all page tables below KERNEL_BASE
* also set PXN for physical page mapper
PXN, Privileged execute-never
When the PXN bit is 1, a Permission fault is generated if the processor
is executing at PL1 and attempts to execute an instruction fetched from
the corresponding memory region.
Change-Id: I3056cbed151004ac9edfbc81ebeada328aeb603c
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5607
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Use the Privileged Only Thread ID Register aka TPIDRPRW to store
the current thread pointer.
The Privileged Only Thread ID Register is only accessible
in privileged modes, and is read/write.
see: ARMv7 Architecture Reference Manual,
section B3.12.46 CP15 c13 Software Thread ID registers
Change-Id: I5273bee8a80b78cdc547b2f6c96632d120eb3d55
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5608
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Mostly for HiDPI, but this also cleans up the code in general, too.
* Apply a scale-factor when loading and serializing column size/offsets
in ViewState.
* Do not hard-code offset values for default columns but let
BPoseView::AddColumn() compute them manually instead.
* Change BPoseView::AddColumn() to support being called during
early initialization and utilize it in SetUpDefaultColumns...().
* When adding deserialized columns to a BPoseView, always realign
the offsets instead of just doing sanity checking. The rationale here
is that the first time the TitleView is touched, it will do this
anyway, and cause "snapping" or drawing glitches if the alignment
is not as expected.
If it was actually somehow an intended feature that non-adjacent
columns can be displayed, changes are needed in TitleView and PoseView
to properly support this without triggering redraw glitches.
I still saw some very slight redraw glitches (e.g. column contents
shifting by 1px) at higher font sizes; however, Tracker before this
commit had far more of those glitches (this commit seems to resolve
a number of them.)
Tested with 12pt, 18pt, and 20pt font sizes; this seems to work
pretty well. But if you see strange behavior or more redraw glitches
than there were previously, please do file a ticket.
ViewState now always holds the unscaled icon size and PoseView
computes and caches the actual display icon size.
Anyone who used Tracker on HiDPI before this commit will likely need
to reduce the icon size on their desktops and any open windows.
Change-Id: I8689ea6baa408d07c715703e8f8bf8f29bf40f60
std::find() never returns NULL but the last element in case of failure
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5602
Reviewed-by: Máximo Castañeda <antiswen@yahoo.es>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
* All MenuItem variants which draw icons now derive from IconMenuItem
and use its functions to draw and otherwise manage their icons. This
resolves a number of TODOs and reduces code duplication.
* Use BControlLook::ComposeIconSize() to compose icon sizes throughout.
* Remove unused methods from IconMenuItem.
"Shutting down" is too reminiscent of a system shutdown.
We're just quitting the app. Going with a generic "Cleaning up".
Correcting the app name, HaikuDepot.
Change-Id: Ifadf533170dc89befc971ea86b7140d3aa7fcef8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5605
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
If the new row is not in an open branch, it cannot possibly have
caused the focused row to have moved, and so we need neither check
nor perform an invalidation if that is the case.
This code has been this way since it was originally imported into
Vision's (!) CVS repository in 2001 from some Be sample code area,
and has not been substantially changed since it was imported into
Haiku's repository in 2004.
This saves quite a number of app_server round-trips and redraws
when importing lots of rows to the list, which fixes the other half
of the hang when changing functions in the call stack in Debugger.
This table will have thousands of items in every branch, most with
extremely long names, so resizing all its columns to the preferred size
takes an unbelievable number of app_server round-trips to measure text.
For modules with any significant usage of C++, there are going to be
functions with names so long that it will make the table quite
unwiedly to work with, anyway, when columns are automatically resized.
This fixes half of the whole-application hang that happens when changing
between functions in the call stack.
The buttons are now aligned to the scroll bar and window resize knob,
by using BControlLook::GetScrollBarWidth. (Hmm, perhaps it might make
more sense to just have scrollbar sizes be a kind of insets, which can
be gotten via B_USE_SCROLL_BAR_INSETS, and then drop this extra method
altogether? After all there are a lot of applications which align to
scroll bars...)
Resolve an issue with rendering package
icons when "Disable user addons" is
turned off.
Relates to #17862
Change-Id: Ic2d17626c40a97598613b8f21218d6829de97bbf
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5556
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jacob Secunda <secundaja@gmail.com>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This was a holdover from BeOS, which did this as inline assembly.
We do not on any platform other than 32-bit x86, and even there
we may have preferred to do things a little differently on non-BeOS
ABIs.
Most things ported from other systems, or even native apps, are going to
use _Thread_local variables anyway, which will bypass this system
altogether.