This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.

Bug 27853 - Yet another small fixes
Summary: Yet another small fixes
Status: RESOLVED FIXED
Alias: None
Product: WHATWG
Classification: Unclassified
Component: Fetch (show other bugs)
Version: unspecified
Hardware: PC Windows NT
: P2 normal
Target Milestone: Unsorted
Assignee: Anne
QA Contact: sideshowbarker+fetchspec
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-01-19 12:09 UTC by Arkadiusz Michalski (Spirit)
Modified: 2015-01-20 09:25 UTC (History)
1 user (show)

See Also:


Attachments

Description Arkadiusz Michalski (Spirit) 2015-01-19 12:09:22 UTC
I caught a few more things:

===

For Request() constructor, step 3.,  "referrer is client" << client is initial value for "referrer" so probably can be skip here:
https://fetch.spec.whatwg.org/#concept-request-referrer

Btw, why "context" (https://fetch.spec.whatwg.org/#concept-request-context) hasn't initial value, but we see that "context frame type" has. If you set initial for "context" (let say fetch) then  "context is fetch" in Request() constructor, step 3., can be removed too. 


2. "Let request be input's associated request,..." << this second request can/should refer to https://fetch.spec.whatwg.org/#concept-request-request

In step 13., "force cache" should not be wrapped by <dd>?

In step 21.4 you use more restrict append algo (https://fetch.spec.whatwg.org/#concept-headers-append), but for Response() constructor step 7.3 append is just https://fetch.spec.whatwg.org/#concept-header-list-append. This two constructor have very similar steps for body and I wondering why they use a different append, it's intentional?

===

For "header list" (https://fetch.spec.whatwg.org/#concept-headers-header-list), you can add ", which is initially empty".

=== 

https://fetch.spec.whatwg.org/#http-access-control-allow-origin
"Indicates whether a resource can be shared, via returning the literal value of the Origin reques...." << not `Origin`?

===

Why in fill algo (https://fetch.spec.whatwg.org/#concept-headers-fill) header's key must be converted to ByteString, but for sequence (step 2.) that operation is not needed?

IDL definition:
sequence<sequence<ByteString>>
OpenEndedDictionary<ByteString>

===

In basic fetch algo (https://fetch.spec.whatwg.org/#concept-basic-fetch), for "blob" scheme, you define /length/ variable (step 3.), but use it only one time (step 5.). I think you can skip step 3. and write just: 
4. Append `Content-Length`/blob's size to response's header list.

After this we get shorter algo. Notice, that for blob's type you don't define any variable.
Comment 1 Anne 2015-01-20 09:25:27 UTC
https://github.com/whatwg/fetch/commit/bb2c619b74ffc5bc3d048fb75fbb5a89a0d673b2

> Btw, why "context" (https://fetch.spec.whatwg.org/#concept-request-context)
> hasn't initial value, but we see that "context frame type" has.

I want those creating requests to think about context.

> In step 21.4 you use more restrict append algo
> (https://fetch.spec.whatwg.org/#concept-headers-append), but for Response()
> constructor step 7.3 append is just
> https://fetch.spec.whatwg.org/#concept-header-list-append. This two
> constructor have very similar steps for body and I wondering why they use a
> different append, it's intentional?

Yes, Request is much more security sensitive.


> Why in fill algo (https://fetch.spec.whatwg.org/#concept-headers-fill)
> header's key must be converted to ByteString, but for sequence (step 2.)
> that operation is not needed?
> 
> IDL definition:
> sequence<sequence<ByteString>>
> OpenEndedDictionary<ByteString>

I'm awaiting some resolution to bug 27008.