The class HttpHeader currently provides a mixed bunch of methods to
retrieve headers as a blob of ASCII in a String. This is causing a lot
of code almost-duplication, String copying, validation difficulties,
etc. Our Parser-NG project seeks to reduce this and make it a lot more
efficient by scoping the parse action inside Parser that can be audited
for accuracy isolated from the complications of header semantic handling.
Several of the IETF HTTPbis WG members have a parallel project underway
to refactor the HTTP headers into some form of consistent syntax. The
latest results of that background work is
<https://tools.ietf.org/html/draft-ietf-httpbis-header-structure>. PHK
submitted this draft, but it builds on a lot of work from Julien and others.
The implementation of this would be refactoring HttpHeader so most
headers can be parsed and syntax validated in one place (inside
HttpHeader) and callers receive a structured tree conforming as much as
possible to the common header structure instead of a String blob. The
callers then (at worst) only need to know what key or parameter it is
looking for and how to scan the tree for that instead of parsing the
whole String.
Anyone interested in joining the WG discussion of the draft and/or
working on the Squid refactoring with me?
Amos
_______________________________________________
squid-dev mailing list
[email protected]
http://lists.squid-cache.org/listinfo/squid-dev
