Summary:
Both font and font_size are already added into the style text
in _edje_format_reparse() function and there we update the tag->font_size
as well as tag->font member. so I think it is unnecessary to
add again which has memory as well as parsing performance impact.
Note :
someone please update this cryptic comment
/* Add font name last to save evas from multiple loads */
how this is going to help saving multiple load.
Reviewers: ali.alzyod, Hermet, raster, cedric
Reviewed By: Hermet
Subscribers: segfaultxavi, cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9543
Reviewed-by: Felipe Magno de Almeida <felipe@expertisesolutions.com.br>
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Differential Revision: https://phab.enlightenment.org/D9529
the main user of textblock_style_set() api is the edje whcih keeps its owen edje_textblock_style
tags( string_shared string) by changing the textblock to keep the string_shared string will improve
the chance of sharing the same string hence reducing memory.
By removing the Eina_StrBuf usage inside the loop in textblock_style_set() api we can avoid lot
of temporary memory allocation and deallocation hence will improve performance.
Note: I see lot of places we use Eina_Strbuf inside a loop (eina_strbuf_new() does 2 allocation)
We need to be extra carefull while reviewing when the code uses those construct to see if its really necessary.
Data: it reduces memory allocation by 7000 in elementary_test launch time.
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D9545
Evas_TextBlock_Style has no idea about the text_class tag its a garbage value to it.
So keep the text_class tag in the edje level and update the text style property in the
final style string when necessary.
Because text_class id tends to be unique by removing from the final style string
enable it to be shared (string shared string).
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D9544
asan is most unhappy about using a priori stack frame's data for
this.. it seems check uses the char * buf directly as-is without
duplicating it... so we can't sensibvly use local stack data. that is
what asan is saying... and this makes our tests not work under asan to
begin with... nto to mention other possible issues i have yet to see
as i got to this one first.
coverity doesnt like inconsistent behavior of code, so make it
consistent even if pointless as this is during init when we wont have
threads .... yet.
CID 1403903
in the case pipes fail to create we'll close the wrong ones... this
fixes that. it also happens because i didn't use names consistently.
now it does so it's easier to keep right.
thanks coverity.
fix CID 1396994
on read/decode we can avoid lots of little locks and unlocked by
having the rwlock go to the outside and a single lock+unlock (read) on
the dictionary. it blocks for longer by has less atomics/fence points
as a result and thus decodes faster where changes that we are writing
while decoding is insanely low so no point worrying here.
this is mostly to verify that the internal label is sizing as expected
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D9523
ensure that scroll-based popups don't accidentally trigger a full
recalc and (wrong) size change during group_calc by hitting the base
popup size_set() implementation and setting the needs_calc flag or
by using the base popup calc code when it should not be used
this resolves a corner case sizing issue in the text_alert 2 popup
case in elm_test which seems to have been present for a long time
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D9538
env vars should not be retained when this subsystem is not "active"
fixes unit test running
ref 9149767184
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D9537
This changes the behavior of the existing env var to only check
unimplemented functions in stable APIs by default. Beta checks
can be enabled with an additional environment var, so use
EOLIAN_CLASS_UNIMPLEMENTED_WARN for stable and
EOLIAN_CLASS_UNIMPLEMENTED_BETA_WARN for extra beta checks.
Summary:
textblock format parsing creates/delets eina_TempStr for each
token. Eina_TempStr is even worse than malloc/free because it also
takes a lock/unlock along with malloc/free each time we create it.
Just use a stack bufefr and create string in it if the string is too big
then it will fall back to heap which is anyway we are doing right now.
Tested this in Tizen List view the number of allocation reduced by 16000.
Reviewers: Hermet, ali.alzyod, woohyun
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9534
Summary:
If the color_class is not overridden by the object level the
hash will be empty but still we do the expensive call to _edje_hash_find_helper()
find the color_class in an empty hash. by checking if the hash is empty
and returning early we save lot of unnecessary hash computaion and lookup.
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9532
Summary:
Check if hash is empty before computing the hash key and look inside the
hash to find data.
Note: could have called the eina_hash_population() api but didn't
because of extra function call.
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9531
if a proxy is not being proxied, it's optimal to create a surface for only
the necessary dimensions and then only draw within those dimensions.
when a proxy is clipped to a size smaller than the proxy object, the required
size for the proxy render becomes smaller as the proxy has less visible area.
this enables us to draw only the clipped region and thus gives a performance
boost
this can only be enabled if the clipper is marked as static
Reviewed-by: Hermet Park <hermetpark@gmail.com>
Differential Revision: https://phab.enlightenment.org/D8881
there's no point in allocating a massive mask surface if it's going to
be clipped to a smaller size, so instead just allocate the smaller size
and position it where the clipped size would be
this can only be enabled if the clipper is known to not be changing size,
as performance would be impacted if the clipper was forcing a full mask
redraw due to regular resizing
Reviewed-by: Hermet Park <hermetpark@gmail.com>
Differential Revision: https://phab.enlightenment.org/D8841
this provides a hint for rendering that the object is not going to resize
for as long as the flag is set and allows for some optimizations to
be made during rendering based on this knowledge
@feature
Reviewed-by: Cedric BAIL <cedric.bail@free.fr>
Differential Revision: https://phab.enlightenment.org/D8887
do it once and remember the result from the first one. drops overhead
for sure by a chunk i actually could see in perf reports like about 1-2%
of cpu...
Summary:
this brings a spec test suite for Efl.Ui.Selectable
Depends on D9517
Reviewers: zmike, cedric
Reviewed By: zmike
Subscribers: #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9518
Summary:
The internal wrapper was generating the argument types directly instead
of passing through the translation generator `grammar::c_type`.
This caused the type in the `caller` callback to be different from the
actual C type of the declared function pointer, like in `@out` parameters.
Reviewers: tasn, felipealmeida
Reviewed By: felipealmeida
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D9524
also.. note the badness of the code design mixing a global singleton
with a "per struct" set of data like fd handlers for the same devices
initted only once but... anyway. it's messy.
Jaehyun Cho
subhransu mohanty
Cedric BAIL
Carsten Haitzler
Mike Blumenkrantz
Daniel Kolesa
Hermet Park
Boris Faure
Marcel Hollerbach
Lauro Moura