When passing an owned acessor from a converted collection we need a way
to unpin the passed data when the accessor is freed.
This commits adds a thin wrapper around the CArray accessor that unpins
the data when freed.
Summary:
When pulling in OpenSSL and building against the
port there are some linker complications.
For example, doing a TLS HTTP transaction will
crash in libcurl which was built against
OpenSSL in base. If we link against FreeBSD's
OpenSSL these issues are resolved.
Test Plan:
On FreeBSD use the openssl port and build EFL
against it.
1) Upload a screenshot in E (segv).
2) Click on a link to image in terminology over
HTTPS (segv).
Reviewers: bu5hm4n, raster
Reviewed By: raster
Subscribers: cedric, Peter2121, #reviewers, q66, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D10789
we also need to reemit the position of current active element. Otherwise
the indicator will display a wrong value.
This is already done in any other spotlight manager.
fixes T8499
Differential Revision: https://phab.enlightenment.org/D10782
before this commit, push would add before the current element, pop would
return to the next higher element.
after this commit, push would add after the current element, pop would
return to the previous element.
ref T7991
Differential Revision: https://phab.enlightenment.org/D10781
Revert "eina: remove no longer used function _eina_thread_queue_msg_block_real_free"
This reverts commit 695b44526c.
Revert "eina/threadqueue: use mempool_del for hash free function"
This reverts commit b0cb3b935a.
Revert "eina_thread_queue: use normal mempools for block allocation"
This reverts commit 14ae3e3dec.
Why? Threadqueue is a highly performance sensitive API.
_eina_thread_queue_msg_block_new() may be called quite often. Doing a
hash lookup to then find a mempool handle to then allocate from was
not the same as what was there and was going to be far more costly.
This would have actual performance impact as we have to compute a hash
and rummage through a hash, hunt for an environment var too. The
original code looked at a spare block pool where blocks *MAY* be of
different sizes (not always the same size so using a mempool is
actually wrong and will stop threadqueue from being able to send
larger messages at all). If you send large messages, larger blocks would
have been allocated and put in this pool. In almost all cases the first
item in the pool would be big enough so we don't hunt and the find pulls
out the first memory, resets the fields that are needed and returns that
block. If it needs a bigger one, it does hunt. This is going to be
rare that such big blocks are needed so I never tried to optimize this
(but it could be done with an array of sizes to make a walk to find
the right sized element cheap if the need arises).
Performance dropped quite a lot. On aarch64 The above mempool usage
dropped message rate from 1037251 msg/sec to 610316. On x86 it was even
worse. It dropped from 2815775 msg/sec to 378653.
So backing this out sees the message rate is 7.4 times faster and on
aarch64 it's 1.7 times faster.
So moving to a mempool was actually just wrong (size is not always the
same). Also this ended up with a mempool of 64k for thread queue blocks even
if we only sent messages sporadically, as opposed to a single 4kb
block. So backing this out saves memory by only having 1 or 2 4k blocks
around most of the time, not a 64k mempool.
So the above patch then follow-on patches were done without accounting
for the performance implications. There were good reasons to do what I
did - because this code was highly tuned even to the point where I
used atomics instead of locks specifically to cut down some contention
overhead. Beware when you change something that there may be steep
performance implications. 7.4 times faster to go back to what was
there is a great example.
This file is identical with test_naviframe.c and has no reference in our
build system. Found by the use of the Duplo duplicated code checker
tool.
Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D10777
The filter_event function calling a lot of times when it runs.
This can help performance by reducing the number of calls to the efl_data_scope_get() function.
Reviewed-by: Hermet Park <hermetpark@gmail.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D10437
Small patch to fix issue where merge of patch did not go so well, and
thus caused Coverity to report an issue where window was not being
initialized.
ref T8016
fixes CID1408311
They are just helpers, so a user does not need to handcreate spotlight
managers. This also leaves us the possibility of letting
spotlight_managers beeing @beta.
ref T7991
Differential Revision: https://phab.enlightenment.org/D10773
this introduces the test requested in D10701. And checks that all
signals are correctly delivered even if animations are frozen
Differential Revision: https://phab.enlightenment.org/D10767
the spotlight now is setting the overall correct min and max size on
itself. Additionally, the page size is now clamped to the size of the
container.
Correct min size of the container is defined to the MAX min size of all
the content.
The correct max size of the container is defined to the MIN max size of
all the content.
ref T7991
Differential Revision: https://phab.enlightenment.org/D10766
before we inherited from layout, however the theme of the layout was
basically just the event part, and the holder part for the indicator.
The indicator part is going to be refactored into something else
anyways, since the indicator should be useable on other widgets as well.
Which means, only the event part is left, which is only used by the
scroller spotlight manager, (and now moved there).
With the move from this away we are saving round about 0.2KB of pure
edje accounting. Additionally, we are saving in perf 4% that is spend in
_efl_canvas_layout_efl_gfx_entity_size_set, which also makes this less
CPU intensive when resizing (Or even just starting is also enough).
ref T7991
Differential Revision: https://phab.enlightenment.org/D10765
the main widget before presented a layout to hold the indicator,
however, for the causual case, this is not really needed. The indicator
itself is anyways going to be a more general prupose widget soon, where
the layout here can be taken as an starting point.
Additionally, this fixes general displaying of the indicator, before the
box padding refactor, a padding would have changed the minsize of the
mix, this is not the case anymore, which forces us to calculate the
minsize of the indicator theme.
Differential Revision: https://phab.enlightenment.org/D10764
for now the eventrect was a swallow part, created by the container, set
to the edje layout. However, the only real user for this is the scroll
spotlight manager.
Which means, we have mostly unneeded element resized by edje, which is
quite an overhead. With this commit, this is moved to the scroll
manager, which makes the usage with stack and plain less memory heavy.
Differential Revision: https://phab.enlightenment.org/D10763
if something is just repositioned, but not resized, it would not have
been placed correctly. This fixes that.
Differential Revision: https://phab.enlightenment.org/D10762
Summary:
In simple words:
when we create efl_canvas_text object, efl_text_password_get will return TRUE.
which is wrong,
efl_text_password_get should return FALSE, unless user change the password using efl_text_password_set
Reviewers: woohyun, zmike, segfaultxavi, tasn
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D10735
again 37b55172b0 broke min size calc.
this time it was for mouse cursors. the entry cursor was 0 sized after
commit 37b55172b0. this disables the
"don't cacle if its 0 sized) which isnt a valid thing to skip - you
caqn min size calc a 0x0 edje ... and it should then return a valid
min size given that design/layout.
This icon is used whenever an icon cannot be found. It's most
commonly found within E itself. The previous iteration has been
with us since at least Black and White (theme), which is pre
E17. Tried to keep it simple yet informative.
Summary:
Value types are already assumed to be stored by pointer (e.g.
`int val = *(node->data);`)
This commit just changes the current usage of the `ptr` modifier in the
ptr, not affecting the parser.
Reviewers: q66, segfaultxavi, bu5hm4n, felipealmeida
Reviewed By: q66
Subscribers: cedric, #reviewers, #committers, brunobelo
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D10769
This fixes the issue of zero division when columns is zero.
Sometimes columns can be zero when "pd->viewport.w = 0".
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D10768
Since the vector file data is shareable among the multiple vg instances,
vfd could keep the requested frame data already by the other instance.
This case vector cache quickly return the vector data withouth any further progress.
Lauro Moura
Yeongjong Lee
Alastair Poole
Xavi Artigas
Woochanlee
Christopher Michael
Marcel Hollerbach
Carsten Haitzler
Stefan Schmidt
Ali Alzyod
Hermet Park
Daniel Kolesa
WooHyun Jung