Re: CRLF requirement

Julian,

On Jun 30, 2014, at 7:50 AM, Julian Reschke <julian.reschke@gmx.de> wrote:
> On 2014-06-30 13:40, Anne van Kesteren wrote:
>> On Mon, Jun 30, 2014 at 1:34 PM, Julian Reschke <julian.reschke@gmx.de> wrote:
>>> On 2014-06-30 13:23, Anne van Kesteren wrote:
>>> Assuming that the statement above is true: because there are some existing
>>> implementations that require CRLF and which would be broken by just sending
>>> CR or LF.
>> 
>> It seems for robustness you'd want to parse using newline (per initial
>> email), but require CRLF for transmission. (The tight coupling the
>> HTTP specification has with parsing and format has always seemed like
>> a bug.)
>> ...
> 
> Do you have any evidence that accepting anything but CRLF is needed in practice? If yes, we'd probably want to add something to the prose about the message format.

In my experience, many servers and clients accept just "LF".  And for CUPS we always generate CR LF but accept both CR LF and LF alone.

_________________________________________________________
Michael Sweet, Senior Printing System Engineer, PWG Chair