[rfc-i] EOL in draft-flanagan-plaintext-04

Robert Sparks rjsparks at nostrum.com
Tue Oct 28 12:49:07 PDT 2014


On 10/28/14 2:44 PM, Brian E Carpenter wrote:
> Well, since it was me who asked for the old EOL text to be kept,
> I would like to know why you think the inconsistent EOL problem has gone
> away. As far as I know, it's still alive and kicking when files are copied
> between any Unix-like system and Windows, and can be seen trivially by
> any user of NotePad on Windows when the file has LF alone to signal a
> new line.
>
> Maybe the old text goes too far, but I think the problem does need to
> be flagged somehow.
As I said, I don't mind warning people that they might run into a file
where EOL has been changed, but we should define what we're going
to publish unambiguously.
>
>     Brian
>
> On 29/10/2014 08:08, Joe Hildebrand (jhildebr) wrote:
>> +1.  Pick a line ending and stick to it.  To start painting the bike shed,
>> I like simple "\n"s.
>>
>> Modern transports such as HTTP will transmit the file with a length and
>> encoding information.  FTP can be used in a similar manner if set to
>> "binary".  Telnet isn't really about transferring files, without escapes
>> to [xyz]modem land.  SMTP can deal with base64-encoded binary just fine.
>> My point here: the protocol used to transmit the file has almost nothing
>> to do with the format we pick.  If you use a protocol that mangles the
>> file in transit, or configure a protocol to do so, then you shouldn't be
>> surprised if the signature doesn't match.
>>
>> On 10/28/14, 6:44 PM, "Robert Sparks" <rjsparks at nostrum.com> wrote:
>>
>>> Section 4.3 of -04 spends a couple of paragraphs talking about having
>>> different EOL markers on different OSes, and what some transports might
>>> want to see. I think this is confusing what the RFC Editor will produce
>>> and publish with what people might find lying around after it's been
>>> moved off the RFC Editor's website.
>>>
>>> I think this document should specify exactly what the plaintext
>>> publication format will contain (and what will be signable).
>>> I _don't_ think the plan is to publish versions with the various line
>>> ending styles. Unless that's wrong, please pick one and say what it is.
>>>
>>> It's ok to warn people that their own tools might change the line
>>> endings for them, but don't make it look like the variants are part of
>>> what's published.
>>>
>>> RjS
>>> _______________________________________________
>>> rfc-interest mailing list
>>> rfc-interest at rfc-editor.org
>>> https://www.rfc-editor.org/mailman/listinfo/rfc-interest
>>>
>>



More information about the rfc-interest mailing list