Part of the origin of the problem is that the HTTP 1.0 Header has the
misspelling. See
http://en.wikipedia.org/wiki/HTTP_referer for the history. I think I
made an effort long ago to correct it in our codebase where it was not a
direct reference to the header name, but we may have regressed.
--a.
On 3/24/13 8:09 AM, Dave wrote:
On Sun, Mar 24, 2013 at 9:51 AM, Glen Mazza <[email protected]> wrote:
I'll look at the source more thoroughly; we seem to be 50/50 between the
two spellings. The *only* reason why I might want to apply it to 5.0.2 is
because this change affects several files and it makes it easier to
backport *other* patches and security fixes when the underlying files are
identical.
Very good point.
- Dave
On 03/24/2013 09:22 AM, Dave wrote:
I could go either way on this one, so if you want to take action please
decide and go ahead.
However, I don't think we should address this in 5.0.x -- I think we
should
only address security issues and critical bugs there and a spelling error
does not feel like a critical bug.
Thanks,
- Dave
On Sun, Mar 24, 2013 at 7:25 AM, Glen Mazza <[email protected]> wrote:
Hi, I'd like to implement https://issues.apache.org/****
jira/browse/ROL-1229 <https://issues.apache.org/**jira/browse/ROL-1229><
https://**issues.apache.org/jira/browse/**ROL-1229<https://issues.apache.org/jira/browse/ROL-1229>
,
on at least 5.1 (possibly 5.0.2 as well). Any critical backwards
compatibility problems if I do so? Another alternative is to close this
issue as a "Won't Fix". (As "referer" is an HTTP term:
http://en.wikipedia.org/wiki/****HTTP_referer#Origin_of_the_****
term_referer<http://en.wikipedia.org/wiki/**HTTP_referer#Origin_of_the_**term_referer>
<http://en.**wikipedia.org/wiki/HTTP_**referer#Origin_of_the_term_**
referer<http://en.wikipedia.org/wiki/HTTP_referer#Origin_of_the_term_referer>
,
misspelled or not.)
Thanks,
Glen
