[new http] Normalize line continuations in response headers.

[new http] Normalize line continuations in response headers.

BUG=1272571
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=1818

[Evan Martin, 2008-09-05 00:29:09]

http://codereview.chromium.org/458/diff/1/7
File base/string_tokenizer.h (right):

http://codereview.chromium.org/458/diff/1/7#newcode196
Line 196: typedef StringTokenizerT<std::string, std::string::const_iterator>
StringTokenizer;
80 chars, yo.

[darin (slow to review), 2008-09-05 01:07:21]

Awesome, LGTM!

http://codereview.chromium.org/458/diff/1/3
File net/http/http_util.cc (right):

http://codereview.chromium.org/458/diff/1/3#newcode358
Line 358: // (leading LWS implies a line continuation)
nit: please indicate that line continuations should have already been joined.

http://codereview.chromium.org/458/diff/1/5
File net/http/http_util.h (right):

http://codereview.chromium.org/458/diff/1/5#newcode49
Line 49: static bool IsLWS(char c);
nit: please add a comment explaining what this means.

[wtc, 2008-09-05 01:59:33]

http://codereview.chromium.org/458/diff/1/3
File net/http/http_util.cc (right):

http://codereview.chromium.org/458/diff/1/3#newcode319
Line 319: raw_headers.append(line_begin, line_end);
We should replace the LWS at the beginning of the
continuation line by a space ' '.  See RFC 2616
Section 2.2:

   A recipient MAY replace any linear white space with a single SP before
   interpreting the field value or forwarding the message downstream.

and

   A CRLF is allowed in the definition of TEXT only as part of a header
   field continuation. It is expected that the folding LWS will be
   replaced with a single SP before interpretation of the TEXT value.

[eroman, 2008-09-05 22:26:34]

Spoke with wtc about the white space normalization.

The plan is this is ok to submit, and I will subsequently change the policy and
update the tests.

Doing it in two passes helps me ensure details don't get lost, and keeps the
diffs simpler.

http://codereview.chromium.org/458/diff/1/7
File base/string_tokenizer.h (right):

http://codereview.chromium.org/458/diff/1/7#newcode196
Line 196: typedef StringTokenizerT<std::string, std::string::const_iterator>
StringTokenizer;
On 2008/09/05 00:29:10, Evan Martin wrote:
> 80 chars, yo.

Done.

http://codereview.chromium.org/458/diff/1/3
File net/http/http_util.cc (right):

http://codereview.chromium.org/458/diff/1/3#newcode358
Line 358: // (leading LWS implies a line continuation)
On 2008/09/05 01:07:22, darin wrote:
> nit: please indicate that line continuations should have already been joined.

Done.

http://codereview.chromium.org/458/diff/1/5
File net/http/http_util.h (right):

http://codereview.chromium.org/458/diff/1/5#newcode49
Line 49: static bool IsLWS(char c);
On 2008/09/05 01:07:22, darin wrote:
> nit: please add a comment explaining what this means.

Done.