I will send you email regard Spans.

@tasn How about Break discussion of each task to its own subtask.

This is fixed from Harfbuzz Side.
https://github.com/harfbuzz/harfbuzz/issues/1910

This is clear things up, to know that we are based on current existing stuff

This behavior produces incorrect issues on the widget

Lets discuss this one first :

interface Efl.Text.Font
Can also be applied to annotations
functions and properties:
@property font

@tasn
Again Font should not be interface, Are there font-able objects other than font object?
If my class use font object then he composite inside, not adopt new interface, do you have any example from other platforms ?

It's already explained with examples in the original post.

What post ?

I think it is best to discuss other parts too, current discussion could be discussed in better way.

@tasn This is the example:

@cedric just small thing, If you build without this patch, do you still get error ?

@cedric I think you will get same error, even if you get latest.
the reason is you build latest harfbuzz library, and with it you will get error on evas.
We already raise an issue on Harfbuzz side https://github.com/harfbuzz/harfbuzz/issues/1910

I mean why we need to attach this funcinality (Create and manipulate markups ) to be part of canvas, why cannot to expose it, and everyone can use it canvas and other objects and users (everyone), Why do we need to hide it in TextBlock

+GTK
https://developer.gnome.org/gtk3/stable/GtkTextBuffer.html
(Iterators and Markup are implemented on Markup Object(GtkTextBuffer) level )

In my example, why did you assume data is in HTML format, it could be anything andmaybe building it without annoatation could be very hard (you need to do it from scratch).

So each user want to create Markuptext should create his own Markup builder ?! I do not think this is not good for user

I understand use case as something user can do.

simple use case:
I want to create multible markup text in runtime. (for example I got input form user or service) and I want to cache these markup to be used in future.
And I got these values in backthread and create markup text in the backthread

Now things are more clear.

I do not get the difference between Annotations and Markup Object.
Why do not we make annoations work on Markup Object, instead of working on TextBlock.

Yes and no. I'm not going to create it, but it could be created and it would use exactly the same interfaces (so functions) for both the MarkupText object and the Textblock object.

First lets forget about Documnet and Manager, I think it will complicate things on this stage.

This is seems fine to me.
Just simple note if you think it is right (else forget about it), move the if (stl && !stl->readonly) inside the function _edje_textblock_style_member_add

If the style is readonly then we know for sure it dosen't have any text_class/color_class.
In edje_textblock_styles.c I do not see check for color_class (only text class)

In D9636#178869, @smohanty wrote:

and if you are concern about those 2 allocation you can put a patch to check at start of the function to see if styles is empty and return.

for example of there are no (edf->styles) then we are doing useless allocation, which is against the Idea of this patch

@zmike can you please close this one

The main Idea for efldocument and layoutmanager, is to break textblock into multiple objects, and expose functionality (for future purposes like extending functionality or optimizations)
EflDocument
By Document I mean something similar to :
https://doc.qt.io/qt-5/qtextdocument.html (or like https://developer.apple.com/documentation/appkit/nsdocument if we want to extend it more)
Where this class is responsible for store and parse formatted text, build text-nodes and format-nodes, cursor works on efldocument.

TextBlock internally contains efldocument, and text formatting and cursor functionality are processed by efldocument 
- Suppose user want edit text in background thread, then he does not need textblock, he can work directly on efldocument
- multiple textblock  object can share same efldocument
- efldocument will contains extended functionality to save and load files from Disk and Memory.

Hello @tasn, these are most comment I think of right now :

We create Issue on Harfbuzz:
https://github.com/harfbuzz/harfbuzz/issues/1910

I still did not get the error, and both harfbuzz and fontconfig are enabled, so I try to rebase

rebase

I think single_line and password should be processed same way as textWrap, where these properties founded in both efl_text and efl_ui_text

@bowonryu Does the test depends on local environment ? like ("/images/DejaVuSerif.ttf", "/images/NanumGothic.ttf" )

Will we rename all canvas objects inside efl.Canvas namespace to have canvas word at the end like (text_canvas, image_canvas, rectangle_canvas, ... etc) ?

this is weird I did not get any errors in test suite !

In D9536#176933, @cedric wrote:

I would really appreciate a new test in expedite.

update test plan code

@cedric I did not get exactly what you mean, can you please be clear how to add example in tree(what tree :) )

@smohanty ,
I will take a look at edje parts, but the main issue with edje it has its own logic where it keep set markup_set evas_object_textblock_text_markup_set and for some reason it will force textblock to relayout and create text items. I suspect that the reason set inside Edje logic not the TextBlock, but anyway lets discuss them next week :)

@smohanty thanks for the example, For me now it is more clear where is the problem.

In D9534#176771, @smohanty wrote:

In D9534#176763, @ali.alzyod wrote:

I think we should have some information, and sample code about optimization (time reduction) value.

Well I ran the Vtune Profiler to collect the memory allocation and this one was one of the hotspot.

Test Env: Tizen Mobile
just create a empty window with one button(hello world ). this reformat_line calls 23 times (which means we are parsing the style text 23 times and half of the time the button dosen't have even text).
If you add a genlist to the window the number goes above 2000 and depending on what style text you are using the number of allocation was 16000 (32000 malloc/free).

In D9533#176772, @smohanty wrote:

@ali.alzyod

I think its a issue in Textblock . If the text style hasn't changed then it shouldn't do reformatting the style text. If you need a proof i can create a test case and send it to you which does reformatting although style hasn't changed.
I printed the style text most of the time it was same. and i didn't found any code that checks if style test has changed .. if anything triggers relayouting it just blindly parse the text without checking if the style has changed.

I think we should have some information, and sample code about optimization (time reduction) value.

In D9533#176761, @smohanty wrote:

@ali.alzyod ,

The use case am talking about happens a lot when textblock part is created from edje.  to reproduce the issue create a empty window application and add a button and put some logs in the _format_fill() to see how many times we do reformatting.

Hello @smohanty how are you ?

update for legacy only

In D9491#176193, @segfaultxavi wrote:

In D9491#175958, @ali.alzyod wrote:

@segfaultxavi Why do you think r * 76 causes overflow ?

Because r is a char, or what am, I missing?

remove UNUSED for used variables

Auto testing is hard to implement for this function since no path lead to this function, but we may need to use it in future.

@segfaultxavi Why do you think r * 76 causes overflow ?

@cedric this is other hard to add for auto test

@cedric Thank you for your comments :) and explanations. And I am sorry if it bother you

@raster I saw the filters, and they are very strong feature, but as name suggest they are filters.
But If user is looking at deeper control for manipulate or process raster image level, the he want to manipulate pixels directly in C.

@zmike I mean, I can not see abstract class to present Image class, canvas is related to view image. but I think there should be Image class to manipulate and interact directly with image object.