Re: [whatwg] APIs to interrogate default outgoing HTTP headers, i.e. Accept-Encoding

[Nils Dagsson Moskopp]

Ron Waldon  writes:

> On Tue, 11 Aug 2015 at 08:31 Nils Dagsson Moskopp <
> n...@dieweltistgarnichtso.net> wrote:
>
>> I do not understand that use case. It reads incredibly convoluted to
>> me. The UA controls the transport anyway – it should not make any
>> practical difference to a script how the data is transmitted.
>>
> My use case is centred around trying to optimise network usage when
> requesting content from AWS CloudFront backed by S3.
>
> I 100% agree with you that this should not be a script's problem. However,
> it is. When the server (CloudFront in this case) has raw and GZIP'ed copies
> of content, and no automatic server-side selection between the two, the
> only way to optimise network usage is for the script to make this
> determination.
>
> Unfortunately, there is no way to gain access to the default Accept
> Encoding header from JavaScript, which is necessary to figure out whether
> to download raw of GZIP'ed content.
>
> So we currently do more hoop-jumping by serving a dynamic initial HTML,
> where the server constructing it can reflect the UA's Accept Encoding
> header back to the client in a generated script tag. It's yucky.

So the server is buggy (for your purposes) and you want to have another
feature to work around it. Introducing that will take considerable time
and resources. I suggest to just wait until the server is able to serve
compressed content transparently or use a more functional server setup.

> Beyond our own need for access to the Accept Encoding header, there may be
> other use cases that are supported by providing access to other headers.
>
>
>> Btw, why can AWS CloudFront not into compressed content?
>>
> http://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/ServingCompressedFiles.html
>> CloudFront doesn't compress the files itself
>> Amazon S3 doesn't compress files automatically
> AWS CloudFront will "do the right thing" if it is backed by a Custom Origin
> that honours the Accept Encoding header (not S3).
> We have repeatedly requested improvements from AWS, and they are likely on
> the way, but we have many hoops to jump until then.
>
> When last I checked, many sites using AWS CloudFront and S3, including the
> first-party AWS Console itself, do not serve GZIP'ed resources, which is
> sub-optimal.

It may take considerable longer time to spec a feature, ship it and wait
for user agents to catch up than to just fix your server setup. If your
setup really is so buggy that it cannot serve compressed content (Am I
understanding it correctly? It sounds so stupid!) switch to something
that can, instead of externalizing the costs to UA implementors and
future web developers who will have one more interface to learn.

-- 
Nils Dagsson Moskopp // erlehmann

Re: [whatwg] APIs to interrogate default outgoing HTTP headers, i.e. Accept-Encoding

[Ron Waldon]

On Tue, 11 Aug 2015 at 08:31 Nils Dagsson Moskopp <
n...@dieweltistgarnichtso.net> wrote:

> I do not understand that use case. It reads incredibly convoluted to
> me. The UA controls the transport anyway – it should not make any
> practical difference to a script how the data is transmitted.
>
My use case is centred around trying to optimise network usage when
requesting content from AWS CloudFront backed by S3.

I 100% agree with you that this should not be a script's problem. However,
it is. When the server (CloudFront in this case) has raw and GZIP'ed copies
of content, and no automatic server-side selection between the two, the
only way to optimise network usage is for the script to make this
determination.

Unfortunately, there is no way to gain access to the default Accept
Encoding header from JavaScript, which is necessary to figure out whether
to download raw of GZIP'ed content.

So we currently do more hoop-jumping by serving a dynamic initial HTML,
where the server constructing it can reflect the UA's Accept Encoding
header back to the client in a generated script tag. It's yucky.

Beyond our own need for access to the Accept Encoding header, there may be
other use cases that are supported by providing access to other headers.


> Btw, why can AWS CloudFront not into compressed content?
>
http://docs.aws.amazon.com/AmazonCloudFront/latest/DeveloperGuide/ServingCompressedFiles.html
> CloudFront doesn't compress the files itself
> Amazon S3 doesn't compress files automatically
AWS CloudFront will "do the right thing" if it is backed by a Custom Origin
that honours the Accept Encoding header (not S3).
We have repeatedly requested improvements from AWS, and they are likely on
the way, but we have many hoops to jump until then.

When last I checked, many sites using AWS CloudFront and S3, including the
first-party AWS Console itself, do not serve GZIP'ed resources, which is
sub-optimal.


> --
> Nils Dagsson Moskopp // erlehmann
> 
>

Re: [whatwg] APIs to interrogate default outgoing HTTP headers, i.e. Accept-Encoding

[Nils Dagsson Moskopp]

Ron Waldon  writes:

> I posted this at http://discourse.wicg.io/ a long time ago and forgot to
> email the list about it, so here goes...
>
> ## original post
>
> There's currently no good way to determine whether or not a browser /
> environment supports GZIP-deflated content entirely from the front-end.
> Servers can interrogate the Accept-Encoding header when they receive the
> request, but client-side JavaScript cannot see this value at all.
>
> This is important when using a CDN that doesn't facilitate selection of
> appropriately deflated content (e.g. AWS CloudFront). I've had projects
> where the initial HTML content is dynamically generated only so that the
> server can pass the Accept-Encoding header back to the client. That way,
> the client can adjust the other URLs it uses to pick pre-GZIPed files, e.g.
> blah.js.gz instead of blah.js all the time.

I do not understand that use case. It reads incredibly convoluted to
me. The UA controls the transport anyway – it should not make any
practical difference to a script how the data is transmitted.

Btw, why can AWS CloudFront not into compressed content?

-- 
Nils Dagsson Moskopp // erlehmann

[whatwg] APIs to interrogate default outgoing HTTP headers, i.e. Accept-Encoding

[Ron Waldon]

I posted this at http://discourse.wicg.io/ a long time ago and forgot to
email the list about it, so here goes...

## original post

There's currently no good way to determine whether or not a browser /
environment supports GZIP-deflated content entirely from the front-end.
Servers can interrogate the Accept-Encoding header when they receive the
request, but client-side JavaScript cannot see this value at all.

This is important when using a CDN that doesn't facilitate selection of
appropriately deflated content (e.g. AWS CloudFront). I've had projects
where the initial HTML content is dynamically generated only so that the
server can pass the Accept-Encoding header back to the client. That way,
the client can adjust the other URLs it uses to pick pre-GZIPed files, e.g.
blah.js.gz instead of blah.js all the time.

I was initially thinking that navigator.acceptEncoding could just be
specified to contain the default outgoing value of this header, but it
occurred to me that there are probably other headers where this is handy.

Should this be a function such as XMLHttpRequest.getDefaultRequestHeaders()?

Should all such headers just dangle from the navigator object as in my
previous example?

## summary of responses

http://discourse.wicg.io/users/stuartpb seemed interested at the time.

- it's probably more appropriate to store them in navigator instead of
XMLHttpRequest

- it's probably a good idea to discuss this with the Fetch specification
people

- for security reasons, certain headers (e.g. Cookies) should not be
available this way