While calculating buffer size for IN reports, only "Input" report type
should been taken in account. Complex HID devices, such as "Gaming"
keyboards and mice, often have "Feature"-reports declared (with size way
bigger than "typical" kbd IN report) for vendor-specific device
configurations.
But "Feature"-reports are sent over control channel only and can't
appear on interrupt channel.
Should fix#14919, #17937, #17699
Change-Id: I4b9eb51938ca4aba2bc34247d00ae164eb2c19fc
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5771
Reviewed-by: Jérôme Duval <jerome.duval@gmail.com>
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
This methods do not need DoublyLinkedList class state.
sGetLink field that actually implement GetPrevious/GetNext methods is already static.
Change-Id: Ie0b40f7f1b72d640d75403905b8944666874dc87
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5796
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The last thttpd update was less than perfect. A bad format string was
causing crashes on directory listings, and some of our changes were
removed, like PoorMan's setting for directory listings, leaks and thread
safety.
Fixes: #17329
Change-Id: I84d9862a0ebd5492a2542cf9776462e7fe7dbe77
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5772
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Not particularly high priority (this being a test) but... I came across
these while searching for code still needing updates after the changes
on BPartition::ContentName().
Change-Id: I05e8d6c02ba34688cee4f0b55e4e782243b0af4d
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5769
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
system_time returns the number of microseconds since booting. This is
not what's needed for the thumbnail generation timestamp attribute, as a
result the attribute was always considered out of date and the stored
attribute would never be used.
Change-Id: I3728077c484f341b765700532d3f986e64f165ad
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5767
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
Reviewed-by: John Scipione <jscipione@gmail.com>
This commit formats all the netservices2 code with the `haiku-format` tool from
https://github.com/owenca/haiku-format (commit aa7408e), with the following
customizations:
* SpaceBeforeRangeBasedForLoopColon is set to false
* Braces before a catch block are not wrapped
* Most headers, except for ExclusiveBorrow.h, have been manually reformatted
to adhere to Haiku's header format (issue #19 in the repository)
Change-Id: I693c4515cf26402e48f35d1213ab6d5fcf14bd1e
vnode deletion is prevent either by owning a reference to the node,
or by being at least a reader of the vnodes lock. Thus, in acquire_vnode,
we have to inc_ref_count while still holding the lock in order to prevent
a race.
This function is used so rarely and quite deep inside FS drivers that
I'm not sure this race would ever have been a problem. Nonetheless
the old code was incorrect.
Most of the time, that is harmless and will just cause a slight delay
before the vnode is removed and we return NULL instead of finding it.
However, in rare circumstances involving renames, we can wind up in
a deadlock with the thread that is trying to remove the vnode, and
would have to wait all the way to the timeout (a full ten seconds!).
The only vnodes not about to disappear from the table that can be both
"removed" and "busy" seem to be special vnodes like pipes, which will
be in an "unpublished" state while they are initially "busy" which we
can check for, in case something wants to wait for them.
The "dirconc" test readily triggered a pathological case of this behavior.
Before this commit, it ran for over 15 minutes before I killed it (and
it was not close to done at that point, either.) After this change,
it completes successfully in around 3 minutes or so on my test VM.
Thanks to mjg@freebsd.org for pointing out this testcase and its
misbehavior on Haiku!
Change-Id: Id1accf0aaf0724e1aec927a437d3a2ac1596cd98
Generate a name indicating the volume size and filesystem.
Remove code that was doing that or similar things in various
filesystems.
Change-Id: I6b993735e58cdfaf1f19af575e918614c7fe5679
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5381
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
Reviewed-by: John Scipione <jscipione@gmail.com>
When parsing a raw HTTP body without a known size, whether or not a request is complete depends on the
end of the connection. Make sure that the parser updates to complete when all the bytes in the buffer have
been read and the connection closed.
Change-Id: I6f055b43ffe4a44da65da85c19b71304d804c800
Directory Iterator class represents one instance which could be Short, Extent, Leaf, Node or B+Tree, a good case for making it an abstract class
Change-Id: I925255caf4c4f8bc01a975740ef2ebf0bb2e1b49
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5764
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Opening and closing a file from the results window with Pe, for
example, was triggering unnecessary new grep searches, when only
the attributes or other metadata on the file got changed (and
not its content).
That new search was also changing the focus from the results view
to the search field (making one lose the position on that list,
specially annoying with large number of results).
Avoid that by not monitoring/reacting to B_ATTR_CHANGED, and not
reacting to B_STAT_CHANGED if the only flag is B_STAT_CHANGE_TIME.
Change-Id: I56d34c93da94acf36890abe458da45d26a334593
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5763
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
The GetNextNetwork() method is really inefficient: it fetches all the
networks at once from the kernel every single time and then winds
up returning only one of them. In parts of the GUI that iterate over
all networks more than once per refresh (sometimes within a loop, even!)
this was often a noticeable lag on the GUI, especially with OpenBSD
drivers which have extra overhead to do struct translation in the
ioctl handler.
Now, we have a way to fetch all scan results at once and just iterate
over them as many times as we need, and this is what NetworkStatus
and Network preferences now do, saving lots of time and effort.
These were not maintained here, merely mirrored from the Userguide Translator,
and rather infrequently at that. But as they were built with Haiku,
that meant new versions of these packages were built every hrev,
which meant a lot of wasted space and bandwith for updates.
Now there is a separate "userguide" repository which will collect the
files exported from the Translator, and recipes at HaikuPorts to build
the packages, instead, so we can delete these files from this repo.
They are now created from a separate repository and via HaikuPorts.
Since the HaikuPorts packages are not yet in the BuildPackageRepository,
the actual section that adds the packages is commented out (but it only
runs for release-* targets anyway.)
The "welcome", "userguide", and "quicktour" scripts, which are symlinked
from the desktop in release builds, are added to the Haiku package's
"regular" profile instead (they do not actually require the other packages
to be installed but will detect if they are not and launch the online
versions instead.)
This is normally done just after init, but in the case of deferred init,
there will be nothing to print at that point.
Currently, there is no way to access that info after the system has
booted. listdev should be extended, or the info should be published in
PCI device nodes to view in the Devices app.
Change-Id: Id89377ccf8bb967abae96194e503fffc06477acf
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5758
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: waddlesplash <waddlesplash@gmail.com>
This change refactors various parts of the HTTP parsing. The HttpParser now
tracks what part of the message needs to be received next, and throws an error
if the object is used in the wrong way (when the caller requests to parse the
wrong part of he message).
The metadata about the transmission is now also saved in te parser. There is
partial work in there to start exposing the 'bytes written', which in
compressed streams is expected to differ from the bytes read. This is not
used yet.
This also simplifies the state tracking done by BHttpSession::Request.
Change-Id: I8532c6a5c8776456ea8bbccd6df7a44bac92b60d
- Made GetNext() function in Node Attributes efficient by reducing disk seeks to number of Node entries
Change-Id: I4e2bf8fb1898676c30eca2c6d6f39dc1999ae2f8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/5612
Tested-by: Commit checker robot <no-reply+buildbot@haiku-os.org>
Reviewed-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
Before implementing interrupts, all I/O was polled.
Now we only use polling as fallback if interrupts mode
does not work for some reason.
Should fix#17484.
* Set only a minimum width and then resize to preferred on window.
* Set a minimum height for the text view itself.
* Use the new TextView controls instead of the old BTextView.
* Adjust other size constraints as needed.
Fixes#17998.
