Re: Line termination philosophy [Re: I-D ACTION:draft-ietf-secsh-publickeyfile-08.txt]

[der Mouse <mouse%Rodents.Montreal.QC.CA@localhost>]

> There's a simple way to solve this debate.

> Is anyone on this list (which seems to include most SSH implementors)
> going to have their SSH implementation inconvenienced by the current
> text?  That is, does it affect any real SSH implementation?  If it
> directly affects your code, please let us know.

Probably not.  It certainly won't affect the code I write; it'll just
mean the resulting files won't conform unless the text file format
underlying the interfaces I use actually has line terminations (which
is something I can't easily determine when writing at the C level).

> (As I've already mentioned, my code runs on OSes that optionally have
> record-style files, and one that AFAIK only has record-style files,
> and I'm not inconvenienced in any way, the spec is fine for me).

No, it won't inconvenience you, unless you bother to actually implement
the spec as written, instead of its apparent intent - and probably not
even then unless you're on an OS where the local convention for text
files isn't to use line terminators.  And even *then*, it won't unless
you care about the difference between a file that uses the local
conventions for text files and a file that manages to use line
terminators despite the local conventions being otherwise.

>> It prompted me to go read 2822 and 2045 with this in mind, and I am
>> surprised and discouraged to find that those specs do include such
>> specification and that when they are supposedly applied to local
>> storage, they have thus been (almost universally, in my experience)
>> "implemented" by ignoring their line-termination aspects.
> "Gee, look at all these specs and implementations (many of which have
> been around for years), they're all interpreting them incorrectly
> except for me".

More likely, just implementing their apparent intent rather than
implementing the spec as written.  If the current language stands,
that's basically what I'll do for publickeyfile (but I'll be careful to
not claim conformance, of course).

> Peter (who still can't believe we're actually debating something like
> this.  It's trigraphs all over again).

I wouldn't be, except that I'd like to be able to conform without
having to go to extreme lengths to do so (like making sure that if it's
built on VMS it uses a stream-LF file rather than a varying-record file
for output, something I'm not even sure how to do from C).

I also want the spec to be as good as possible.  It really is quite
good; this one point of line terminations, small as it is, is the
largest (in my opinion) flaw I found in the whole thing.

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse%rodents.montreal.qc.ca@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B