Re: [Technical Errata Reported] RFC7230 (4667)

[Alex Rousskov <[e-mail]>]

On 04/15/2016 06:11 AM, Amos Jeffries wrote:

> Note that the Eratta proposed syntax, including the extra suggestions
> will still not successfully match what Squid is sighting. We would also
> need to allow BWS / OWS / SP trailing the chunk-size value when no
> chunk-ext is present.

As Roy has eloquently said, "Don't confuse the various lenient ways in
which implementations parse HTTP with the requirements on generating
HTTP messages that are defined by the ABNF". This errata is about fixing
[other parts of] the ABNF for the reasons specified in the errata. Squid
will be fixed for other reasons, regardless of the errata acceptance or
the final ABNF shape.

Needless to say, somebody may submit another errata to accommodate
chunk-size space padding without chunk extensions. However, they would
face even greater obstacles justifying such change because it is
debatable whether the "implied *LWS" rule applies to that case -- CR is
not a "token" or "quoted string" that the poorly worded LWS rule kind of
requires... The best they can hope for, after hearing the "be lenient"
argument, is adding an informal RFC sentence alerting implementations of
known space padding issues.


Cheers,

Alex.