Re: XMLHttpRequest: uppercasing method names
I think it would be super useful if there were some standard built in functions for selecting properties. I would benefit greatly from being able to look up a prop with case insensitivity. I guess you COULD do this with proxies but I have to imagine it could be done faster if built in. It wouldn't be awful if I could query a nested object structure using something like css selecting or something like xpath but for us. On Aug 12, 2014 12:51 PM, Anne van Kesteren ann...@annevk.nl wrote: On Tue, Aug 12, 2014 at 6:10 PM, Takeshi Yoshino tyosh...@google.com wrote: On Wed, Aug 13, 2014 at 1:05 AM, Anne van Kesteren ann...@annevk.nl wrote: Well fetch() will never be able to explain synchronous fetches. But in general I agree that it would be sad if XMLHttpRequest could do more than fetch(). Per bz it seems we should just align fetch() with XMLHttpRequest and call it a day. wfm seems Safari is also going to be compliant to the latest spec. https://bugs.webkit.org/show_bug.cgi?id=134264 That's great! fetch() and XMLHttpRequest are now aligned. They both use a normalize concept defined in Fetch. -- http://annevankesteren.nl/
Re: XMLHttpRequest: uppercasing method names
On 2014-08-12 17:23, Boris Zbarsky wrote: On 8/12/14, 9:26 AM, Anne van Kesteren wrote: I somewhat prefer always uppercasing, but that would require changes to XMLHttpRequest. Gecko used to have the always uppercasing behavior for XHR and people complained about it until we aligned with the current spec. I don't think we particularly want to change behavior on developers again. -Boris +1. Please keep inconsistency with the base spec (HTTP) at a minimum. Best regards, Julian
XMLHttpRequest: uppercasing method names
In https://github.com/slightlyoff/ServiceWorker/issues/120 the question came up whether we should perhaps always uppercase method names as that is what people seem to expect. mnot seemed okay with adding appropriate advice on the HTTP side. The alternative is that we stick with our current subset and make that consistent across APIs, and treat other method names as case-sensitive. I somewhat prefer always uppercasing, but that would require changes to XMLHttpRequest. -- http://annevankesteren.nl/
Re: XMLHttpRequest: uppercasing method names
On Aug 12, 2014 9:29 AM, Anne van Kesteren ann...@annevk.nl wrote: In https://github.com/slightlyoff/ServiceWorker/issues/120 the question came up whether we should perhaps always uppercase method names as that is what people seem to expect. mnot seemed okay with adding appropriate advice on the HTTP side. The alternative is that we stick with our current subset and make that consistent across APIs, and treat other method names as case-sensitive. I somewhat prefer always uppercasing, but that would require changes to XMLHttpRequest. -- http://annevankesteren.nl/ Both seem like common enough answers to this question that I think either works. I prefer the later just for consistency sake with xhr and the off chance that we forgot to consider -something- with a change. If there's no really good reason to change it, least change is better IMO
Re: XMLHttpRequest: uppercasing method names
On Tue, Aug 12, 2014 at 3:37 PM, Brian Kardell bkard...@gmail.com wrote: If there's no really good reason to change it, least change is better IMO All I can think of is that it would be somewhat more consistent to not have this list and always uppercase, but yeah, I guess I'll just align fetch() with XMLHttpRequest. -- http://annevankesteren.nl/
Re: XMLHttpRequest: uppercasing method names
On Tue, Aug 12, 2014 at 6:26 AM, Anne van Kesteren ann...@annevk.nl wrote: In https://github.com/slightlyoff/ServiceWorker/issues/120 the question came up whether we should perhaps always uppercase method names as that is what people seem to expect. mnot seemed okay with adding appropriate advice on the HTTP side. The alternative is that we stick with our current subset and make that consistent across APIs, and treat other method names as case-sensitive. I somewhat prefer always uppercasing, but that would require changes to XMLHttpRequest. I prefer making them all case-insensitive, which I guess means always uppercasing. It's not a strong desire, but it seems silly to require a particular, unusual, casing for this kind of thing. ~TJ
Re: XMLHttpRequest: uppercasing method names
On Tue, Aug 12, 2014 at 10:55 PM, Anne van Kesteren ann...@annevk.nl wrote: On Tue, Aug 12, 2014 at 3:37 PM, Brian Kardell bkard...@gmail.com wrote: If there's no really good reason to change it, least change is better IMO All I can think of is that it would be somewhat more consistent to not have this list and always uppercase, Ideally but yeah, I guess I'll just align fetch() with XMLHttpRequest. Isn't it an option that we use stricter rule (all uppercase) for the newly-introduced fetch() API but keep the list for XHR? Aligning XHR and fetch() is basically good but making fetch() inherit the whitelist is a little sad. Some archaeology: - Blink recently reduced the whitelist to conform to the latest WHATWG XHR spec. http://src.chromium.org/viewvc/blink?view=revisionrevision=176592 - Before that, used this list ported to WebKit from Firefox's behavior http://trac.webkit.org/changeset/13652/trunk/WebCore/xml/xmlhttprequest.cpp - Anne introduced the initial version of the part of the spec in Aug 2006 http://dev.w3.org/cvsweb/2006/webapi/XMLHttpRequest/Overview.html.diff?r1=1.12;r2=1.13;f=h -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0124.html -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0126.html
Re: XMLHttpRequest: uppercasing method names
On Aug 12, 2014 11:12 AM, Takeshi Yoshino tyosh...@google.com wrote: On Tue, Aug 12, 2014 at 10:55 PM, Anne van Kesteren ann...@annevk.nl wrote: On Tue, Aug 12, 2014 at 3:37 PM, Brian Kardell bkard...@gmail.com wrote: If there's no really good reason to change it, least change is better IMO All I can think of is that it would be somewhat more consistent to not have this list and always uppercase, Ideally but yeah, I guess I'll just align fetch() with XMLHttpRequest. Isn't it an option that we use stricter rule (all uppercase) for the newly-introduced fetch() API but keep the list for XHR? Aligning XHR and fetch() is basically good but making fetch() inherit the whitelist is a little sad. Some archaeology: - Blink recently reduced the whitelist to conform to the latest WHATWG XHR spec. http://src.chromium.org/viewvc/blink?view=revisionrevision=176592 - Before that, used this list ported to WebKit from Firefox's behavior http://trac.webkit.org/changeset/13652/trunk/WebCore/xml/xmlhttprequest.cpp - Anne introduced the initial version of the part of the spec in Aug 2006 http://dev.w3.org/cvsweb/2006/webapi/XMLHttpRequest/Overview.html.diff?r1=1.12;r2=1.13;f=h -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0124.html -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0126.html fetch should explain magic in XMLHttpRequest et all.. I don't see how it could differ in the way you are suggesting and match
Re: XMLHttpRequest: uppercasing method names
On Wed, Aug 13, 2014 at 12:15 AM, Brian Kardell bkard...@gmail.com wrote: On Aug 12, 2014 11:12 AM, Takeshi Yoshino tyosh...@google.com wrote: On Tue, Aug 12, 2014 at 10:55 PM, Anne van Kesteren ann...@annevk.nl wrote: On Tue, Aug 12, 2014 at 3:37 PM, Brian Kardell bkard...@gmail.com wrote: If there's no really good reason to change it, least change is better IMO All I can think of is that it would be somewhat more consistent to not have this list and always uppercase, Ideally but yeah, I guess I'll just align fetch() with XMLHttpRequest. Isn't it an option that we use stricter rule (all uppercase) for the newly-introduced fetch() API but keep the list for XHR? Aligning XHR and fetch() is basically good but making fetch() inherit the whitelist is a little sad. Some archaeology: - Blink recently reduced the whitelist to conform to the latest WHATWG XHR spec. http://src.chromium.org/viewvc/blink?view=revisionrevision=176592 - Before that, used this list ported to WebKit from Firefox's behavior http://trac.webkit.org/changeset/13652/trunk/WebCore/xml/xmlhttprequest.cpp - Anne introduced the initial version of the part of the spec in Aug 2006 http://dev.w3.org/cvsweb/2006/webapi/XMLHttpRequest/Overview.html.diff?r1=1.12;r2=1.13;f=h -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0124.html -- http://lists.w3.org/Archives/Public/public-webapi/2006Apr/0126.html fetch should explain magic in XMLHttpRequest et all.. I don't see how it could differ in the way you are suggesting and match Which do you mean by fetch? http://fetch.spec.whatwg.org/#dom-global-fetch or http://fetch.spec.whatwg.org/#concept-fetch. fetch() and XHR share the fetch algorithm but have different bootstrap and hooks.
Re: XMLHttpRequest: uppercasing method names
On Tue, Aug 12, 2014 at 5:15 PM, Brian Kardell bkard...@gmail.com wrote: fetch should explain magic in XMLHttpRequest et all.. I don't see how it could differ in the way you are suggesting and match Well fetch() will never be able to explain synchronous fetches. But in general I agree that it would be sad if XMLHttpRequest could do more than fetch(). Per bz it seems we should just align fetch() with XMLHttpRequest and call it a day. -- http://annevankesteren.nl/
Re: XMLHttpRequest: uppercasing method names
On Wed, Aug 13, 2014 at 1:05 AM, Anne van Kesteren ann...@annevk.nl wrote: On Tue, Aug 12, 2014 at 5:15 PM, Brian Kardell bkard...@gmail.com wrote: fetch should explain magic in XMLHttpRequest et all.. I don't see how it could differ in the way you are suggesting and match Well fetch() will never be able to explain synchronous fetches. But in general I agree that it would be sad if XMLHttpRequest could do more than fetch(). Per bz it seems we should just align fetch() with XMLHttpRequest and call it a day. wfm seems Safari is also going to be compliant to the latest spec. https://bugs.webkit.org/show_bug.cgi?id=134264 -- http://annevankesteren.nl/
Re: XMLHttpRequest: uppercasing method names
On Tue, Aug 12, 2014 at 6:10 PM, Takeshi Yoshino tyosh...@google.com wrote: On Wed, Aug 13, 2014 at 1:05 AM, Anne van Kesteren ann...@annevk.nl wrote: Well fetch() will never be able to explain synchronous fetches. But in general I agree that it would be sad if XMLHttpRequest could do more than fetch(). Per bz it seems we should just align fetch() with XMLHttpRequest and call it a day. wfm seems Safari is also going to be compliant to the latest spec. https://bugs.webkit.org/show_bug.cgi?id=134264 That's great! fetch() and XMLHttpRequest are now aligned. They both use a normalize concept defined in Fetch. -- http://annevankesteren.nl/