Can you also summarize why we need in memory conversion as opposed to
simpler alternative?
So it is documented for the next time when we will have similar discussion
:)

On Tue, Oct 7, 2014 at 4:00 PM, Daniel Kurka <[email protected]> wrote:

> Hi all,
>
> summarizing the discussion I just had with Julien: We want to have the in
> memory conversion in GWT 2.7 (even if it is not battle tested yet), since
> this is the way to move foward. We can even do a followup release (2.8) as
> soon as google is using it in production and we sorted all issues out.
>
> If GSS is turned on we will:
>
>  - We will make a decision which generator to use based on file extension
>  - You can not mix gss / css on the same interface
>  - In UiBinder we will add a new attribute to tell that this interface
> needs gss (default css)
>  - If lib has gss in it the app using that lib needs to have it enabled.
>
> I think there is good value in including this in GWT 2.7 since we can ask
> for external feedback as well, while at the same time making sure its
> mature within google.
>
> -Daniel
>
>
>
> On Tuesday, October 7, 2014 8:55:27 AM UTC-7, Goktug Gokdogan wrote:
>>
>> I didn't take notes in the meetup so I cannot recall all the reasons
>> (perhaps someone else did?) but I think one of the reasons to introduce in
>> memory conversion was to get rid of the CSS generator code with the GSS
>> release and another one was the possible complications of coexisting css
>> and gss in the same page (e.g. name collisions).
>>
>> Given that we cannot get rid of the CSS generator in 2.7 and if you think
>> everything would be fine with css/gss mixed together (e.g. with prefixes),
>> I'm totally fine with kicking in-memory conversion out and it is definitely
>> not required for 2.7 even if we decide to keep it.
>>
>> However, arguably there might be still value in having an experimental
>> flag to enable GSS as it is not well tested yet.
>>
>> On Tue, Oct 7, 2014 at 1:37 AM, Julien Dramaix <[email protected]>
>> wrote:
>>
>>> I think we are making the things too complex. Initially, we wanted to
>>> remove the code of the existing generator for the CssResource and
>>> deprecated the existing syntax. It's why we have introduced the in-memory
>>> automatic conversion.
>>>
>>> Now, we've decided to keep the existing generator in GWT 2.7. So I think
>>> that we had better to remove this automatic conversion and choose the right
>>> generator in fonction of the file extension. People that want to use GSS
>>> have to use files with .gss extension. Old files with .css extension will
>>> continue to work because the current generator will be used for these kind
>>> of files. For the uibinder, we add a temporary attributes (and/or we can
>>> foreseen an configuration property that enable GSS by default in UiBinder.)
>>>
>>> That simplify a lot the implementation, removes three configuration
>>> properties (CssResource.enableGss, CssResource.legacy,
>>> CssResource.conversionMode) and we support all uses cases of application
>>> using third party libraries that will result of a mix of GSS and CSS. So
>>> it's simpler for the user and simpler for us.
>>>
>>> In the next release of GWT, when we remove the code for the existing
>>> generator, we
>>> will be able to reintroduce this automatic in-memory conversion if we
>>> want still to support the old syntax.
>>>
>>> Now I think we have to mark the current syntax of CssResource as
>>> deprecated in favor of GSS. In order to able to remove the generator in the
>>> next releases.
>>>
>>> Another idea: In order to ease the conversion of existing CssResource to
>>> GSS, we could maybe during the compilation convert every .css file with the
>>> converter and write the result on disk (in the "extra" directory) if the
>>> -extra flag is set.
>>>
>>> Julien
>>>
>>> On Mon, Oct 6, 2014 at 11:43 PM, 'Goktug Gokdogan' via GWT Contributors
>>> <[email protected]> wrote:
>>>
>>>> We worked on a migration plan a few weeks back. I don't think we need
>>>> to mix css and gss together inside the same app.
>>>>
>>>> At this point, the libraries have multiple reasonable options on
>>>> compatibility:
>>>>
>>>> 1- Provide both css and gss file for the resource (recommended).
>>>> The library will work fine regardless if the app enables gss or not as
>>>> the right file will be chosen for the app. That's what we will do for
>>>> internal resources in the SDK.
>>>>
>>>> 2- Keep providing the css file and make sure it works in strict mode.
>>>> The library will work fine in css mode and will work with gss only if
>>>> the legacy mode is enabled.
>>>>
>>>> 3- Keep the old css that requires lenient mode.
>>>> The library will work fine in css mode and *may* work with gss
>>>> if lenient legacy mode is enabled.
>>>>
>>>> More problematic part is the css inside uibinder. Like you said,
>>>> simplest option is to introduce a temporary attribute in <ui:style> tag
>>>> that marks the content as gss. In this case it is more tricky for the
>>>> library to support compatibility mode similar to the mode (1) above. Either
>>>> the lib should write the style that both compiles with css and gss or if it
>>>> is not feasible then it should extract the resource to a file and follow
>>>> the two file approach described above.
>>>>
>>>>
>>>> On Mon, Oct 6, 2014 at 12:54 PM, Julien Dramaix <[email protected]>
>>>> wrote:
>>>>
>>>>> > Hmm but still a bit of work if a library has lots of inline styles
>>>>> and wants to use GSS stuff.
>>>>> If you want to use GSS in all of your inline styles, you have first to
>>>>> convert them to GSS and it will take more time than simply adding an
>>>>> attribute that say that this inline style is using GSS.
>>>>>
>>>>> The problem with UiBinder is that we don't know if the developer is
>>>>> using CSS or GSS because  the CssResource interface is generated and we
>>>>> cannot based our logic on the file extension as we do with normal
>>>>> CssResource.
>>>>>
>>>>> I have still to investigate but I'm pretty sure that adding an
>>>>> attribute will be the easiest way to support correctly GSS in uibinder
>>>>> inline style.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>  --
>>>>> You received this message because you are subscribed to the Google
>>>>> Groups "GWT Contributors" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>>> an email to google-web-toolkit-contributors+unsubscribe@
>>>>> googlegroups.com.
>>>>> To view this discussion on the web visit https://groups.google.com/d/
>>>>> msgid/google-web-toolkit-contributors/CABb_3%3D4gaeww%
>>>>> 3DY4cTkvNXXKKNmig6_s1CaOpwNAAemHXemb%2BjQ%40mail.gmail.com
>>>>> <https://groups.google.com/d/msgid/google-web-toolkit-contributors/CABb_3%3D4gaeww%3DY4cTkvNXXKKNmig6_s1CaOpwNAAemHXemb%2BjQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>>> For more options, visit https://groups.google.com/d/optout.
>>>>>
>>>>
>>>>  --
>>>> You received this message because you are subscribed to the Google
>>>> Groups "GWT Contributors" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>> an email to google-web-toolkit-contributors+unsubscribe@
>>>> googlegroups.com.
>>>> To view this discussion on the web visit https://groups.google.com/d/
>>>> msgid/google-web-toolkit-contributors/CAN%3DyUA2FYLyv%
>>>> 3DWYfOi0TqtojPbCTPkgzApW6J3BwnmJP19HVAQ%40mail.gmail.com
>>>> <https://groups.google.com/d/msgid/google-web-toolkit-contributors/CAN%3DyUA2FYLyv%3DWYfOi0TqtojPbCTPkgzApW6J3BwnmJP19HVAQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>>> .
>>>>
>>>> For more options, visit https://groups.google.com/d/optout.
>>>>
>>>
>>>  --
>>> You received this message because you are subscribed to the Google
>>> Groups "GWT Contributors" group.
>>> To unsubscribe from this group and stop receiving emails from it, send
>>> an email to [email protected]
>>> .
>>> To view this discussion on the web visit https://groups.google.com/d/
>>> msgid/google-web-toolkit-contributors/CABb_3%3D6X1b-
>>> ZW21UjF3MPABTsofnqQ1G1BUH00fk8pR9KTh2yQ%40mail.gmail.com
>>> <https://groups.google.com/d/msgid/google-web-toolkit-contributors/CABb_3%3D6X1b-ZW21UjF3MPABTsofnqQ1G1BUH00fk8pR9KTh2yQ%40mail.gmail.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>>> For more options, visit https://groups.google.com/d/optout.
>>>
>>
>>  --
> You received this message because you are subscribed to the Google Groups
> "GWT Contributors" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/google-web-toolkit-contributors/7187f2b7-024e-4c5a-94d4-71e6f5f58456%40googlegroups.com
> <https://groups.google.com/d/msgid/google-web-toolkit-contributors/7187f2b7-024e-4c5a-94d4-71e6f5f58456%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups "GWT 
Contributors" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/google-web-toolkit-contributors/CAN%3DyUA3ECj8K_sugcOA6AMODwUV46aKgf%3DqtQgn9wvuKXph-dw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to