[Twisted-Python] HTTPClient handling LF only servers

[Kevin Horn]

On Tue, Jan 25, 2011 at 1:53 PM, <exarkun at twistedmatrix.com> wrote:

> On 07:23 pm, kevin.horn at gmail.com wrote:
> >On Tue, Jan 25, 2011 at 11:54 AM, Jason J. W. Williams <
> >jasonjwwilliams at gmail.com> wrote:
> >>Hi Y'all,
> >Howdy, Jason!
> >>Certain servers (notably Hacker News) break HTTPClient because they
> >>use LF
> >>instead of CRLF to terminate header lines. I've uploaded a patch with
> >>tests
> >>to ticket 4814 (merged into ticket 2842).
> >>
> >>I'd appreciate a review and some discussion about integrating it.
> >>
> >>-J
> >I'm not really in a position to do a review right now, but since:
> >
> >1) 4814 is effectively a duplicate of 2842
> >2) 4814 has a clear problem decription
> >3) 4814 has a patch
> >4) 2842 is old, and confusing
> >
> >Can we go ahead and close 2842 in favor of 4814?  We have too many old
> >tickets floating around...
>
> #4814 is already closed.


Ah. Somehow I thought it was still open.


> If the #2842 description is confusing, then
> feel free to update it to be more clear. :)
>

Well, I can update the summary, but not the description, which is the really
bad part.


>
> We don't close old tickets as duplicates of new tickets, though.  That
> doesn't make sense, and it's annoying to the original reporter/cc list
> as well.
>

This certainly makes sense in theory, but when the last comment from the
nominal owner of the ticket is "I'm not sure what is supposed to be fixed
now.", I'm not sure how helpful it is.  Yes, I realize that you had
responded to it, and clarified the situation.  Just Saying.  There's a LOT
of mutant tickets out there, that don't mean what they say, or have
completely changed in scope.  And there are several which have been closed
in favor of newer, clearer tickets over the years.  I think we need to make
the above policy obvious someplace, like in the developer docs.   Maybe it's
in there, I haven't checked, but I don't recall reading it, and I've read
most of the documentation _many_ times.

In any case, I don't care _that_ much about _which_ one is closed as long as
one is.  Which appears to be the case.


> >
> >(To be clear, I'm not asking Jason to close this ticket...I'm asking
> >for the
> >list's opinion)
> >
> >Also 3833 is at least similar, if not another duplicate.  It would be
> >nice
> >if it could be worked on and/or closed along with this issue.
>
> Similar, certainly.  But it's for the other HTTP client implementation.
> There won't be any code in common in the resolution.
>
> I certainly agree that if it's worth fixing for one it's worth fixing
> for the other, though, and I hope someone will put #3833 up for review
> soon. :)
>
> Jean-Paul
>
> _______________________________________________
> Twisted-Python mailing list
> Twisted-Python at twistedmatrix.com
> http://twistedmatrix.com/cgi-bin/mailman/listinfo/twisted-python
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/twisted-python/attachments/20110125/7b6caff3/attachment-0001.html>