'That should compass well via gzip. '

I think I was slightly west-of-north with that typo. Should be:

  That should compress well via gzip.



On Thu, Feb 27, 2020 at 9:01 AM Greg Dove <[email protected]> wrote:

> We definitely had a discussion about this in the past.
>
> I think the use of fonts/font icons etc absolutely makes sense where
> possible. This is common for development in javascript in general (outside
> Royale), and is also often based on the application 'theme'
>
> For actual real embeds though, I think it makes sense to have some sort of
> a solution at some point. To me, the dataURI encoded format seems like it
> would make the most sense for JS. For bitmap base images (png, jpg) this
> should still be base64 I think, I don't understand how that could be
> improved upon. For svg maybe it can be regular text instead of base64. That
> should compass well via gzip.
>
> Having the smallest possible javascript is not always the most optimal
> approach in any given situation.
> I think it was either Harbs or Josh (I can't remember) who pointed out
> that competing technologies can do this 'embedding', and they intentionally
> add smaller (e.g. less than 10kb) images as base64 encoded dataURIs to the
> javascript or css. The reason for this is that it is more optimal than
> separating it out as a separate http load. IIRC the justification was that
> http request/response overhead and any other latency associated with
> accessing the file remotely is heavier than the 33% extra bytes in the
> base64 encoded version loaded as part of the single request for the
> javascript or css file. Apparently Webpack applies this type of rule to how
> it manages images although I am sure it is configurable  - that part is
> what I remember from the earlier discussion, because it was news to me at
> the time.
> From another angle, if you have your application images hosted on the same
> subdomain as your application or even on a separate single subdomain in
> general, I'd suggest that having many of these small images 'embedded'
> could make a big difference to the startup readiness of some apps because
> of the limitations on concurrent http requests by browsers (this was no
> different in flash/Flex). That could also probably be relieved somewhat by
> 'embedding' a lot of the smaller images.
>
> I don't know how any of this relates to browser cache, that might change
> things as well, because after all, unless we are using cachebusters to
> avoid caching, even the application javascript itself should come from
> there most of the time for a stable production app. Once that has happened
> I think people tend to be more concerned about overall 'performance' than
> they would be about 'size'.
>
> my 2 cents...
>
> On Thu, Feb 27, 2020 at 4:28 AM Alex Harui <[email protected]>
> wrote:
>
>> No support for embed at this time.  We could make a nicer error I suppose.
>>
>> There is embedding support in HTML/JS/CSS, but AIUI, it requires Base64
>> encoding of the asset, which makes it larger.  It just isn't clear to me
>> that it is worth it.  We don't generate code right now either because I
>> think we should agree on what the "right" code is, if any.  IMO, even
>> converting embeds to an Image subclass that loads the image isn't the right
>> pattern because it will cause re-layout and bad measurements at startup.  I
>> think smallest size, fastest performance is to convert images to SVG but
>> I'm not sure we can do that automatically.  In the framework itself, most
>> of the symbols that were embedded were replaced by Unicode characters for
>> arrows and cursors.  Symbol fonts are also a possibility.
>>
>> HTH,
>> -Alex
>>
>> On 2/26/20, 6:26 AM, "Carlos Rovira" <[email protected]> wrote:
>>
>>     IIRC, we don't support Embed, or at least don't remember to see
>> nothing
>>     about that in the last months.
>>     I think it's some of the things will need to be converted with the
>> rest of
>>     needed tweaks when come from Flex.
>>
>>     El mié., 26 feb. 2020 a las 15:02, Yishay Weiss (<
>> [email protected]>)
>>     escribió:
>>
>>     > Makes sense.  I was wondering mainly about some emulation to keep
>> the
>>     > syntax. It looks like the compiler is trying to do something with
>> these
>>     > declarations, but it could be unintended.
>>     >
>>     > From: Andrew Wetmore<mailto:[email protected]>
>>     > Sent: Wednesday, February 26, 2020 3:30 PM
>>     > To: Apache Royale Development<mailto:[email protected]>
>>     > Subject: Re: Embed in Royale
>>     >
>>     > I don't believe Embed is required or would be a benefit in Royale,
>> if you
>>     > are going to compile an app to run on a browser using JavaScript.
>>     >
>>     > This page [1] on Flex equivalents says
>>     >
>>     > "Embedding assets was necessary in Flex so the application could
>> quickly
>>     > access images and other assets that were part of a compiled Flash
>> SWF. To
>>     > make it easy to display on a button an icon that was in a SWF, we
>> wrote
>>     > something like: <mx:Button icon="@Embed('icons.swf#our_button')"/>
>> Since
>>     > JavaScript does not support the concept of embedding assets, doing
>> the same
>>     > thing in Royale is a simple URL reference: <mx:Button
>>     > icon="icons/our_button.png"/> (once you have placed the relevant
>> image in
>>     > the *icons* folder).
>>     >
>>     > [1]
>>     >
>>     >
>> https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fapache.github.io%2Froyale-docs%2Fmigrate-an-existing-app%2Fflex-equivalents&amp;data=02%7C01%7Caharui%40adobe.com%7Cad5395cd042147026e6708d7bac7c2c8%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C637183240009896732&amp;sdata=LbzMSZ6uMQO5pJPivTBuer7EJC3HnOsDuEGiJbXGDuo%3D&amp;reserved=0
>>     >
>>     > On Wed, Feb 26, 2020 at 9:18 AM Yishay Weiss <
>> [email protected]>
>>     > wrote:
>>     >
>>     > > Does Embed work in Royale?
>>     > >
>>     > > I’m working on a POC to port an app that has some css that looks
>> like [1]
>>     > > this, but the compiler complains with messages such as
>>     > >
>>     > >     [mxmlc] Error: The definition of base class BitmapAsset was
>> not
>>     > found.
>>     > >
>>     > > Which makes sense as BitmapAsset does not exist in Royale src.
>>     > >
>>     > > Am I right in assuming that this is not yet  implemented?
>>     > >
>>     > > Thanks.
>>     > >
>>     > > [1]
>>     > > .closeBtn {
>>     > >        upSkin:        Embed('/assets/16x16/button_cancel.png');
>>     > >        overSkin:      Embed('/assets/16x16/button_cancel.png');
>>     > >        downSkin:      Embed('/assets/16x16/button_cancel.png');
>>     > >        disabledSkin:  Embed('/assets/16x16/button_cancel.png');
>>     > > }
>>     > >
>>     >
>>     >
>>     > --
>>     > Andrew Wetmore
>>     >
>>     >
>> https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fcottage14.blogspot.com%2F&amp;data=02%7C01%7Caharui%40adobe.com%7Cad5395cd042147026e6708d7bac7c2c8%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C637183240009896732&amp;sdata=JEwrl31eWYLMBfr3JaMi0JCFRK3Z0H3VaoBQU5URGKo%3D&amp;reserved=0
>>     >
>>     >
>>
>>     --
>>     Carlos Rovira
>>
>> https://nam04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fabout.me%2Fcarlosrovira&amp;data=02%7C01%7Caharui%40adobe.com%7Cad5395cd042147026e6708d7bac7c2c8%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C637183240009896732&amp;sdata=9UlK0PWh8hDpOPzriLfrWoTduvyUHzbpTm%2BpRYEMpmE%3D&amp;reserved=0
>>
>>
>>

Reply via email to