[ftputil] Thanks & suggestion

[Stefan Schwarzer]

Randolph:

Sorry for the delay. I was away from home for almost a week.

On 2005-06-18 02:39, Randolph Fritz wrote:
> Just started using ftputil and I'm delighted with it.  Thank you for
> writing it and making it freely available!

Thanks for your thanks :-)

> Suggestion: for servers which support them, which seem pretty common,

The FTP server of my provider doesn't seem to support them. PureFTPd,
installed locally here, has support. It would be intesting to know what
ratio of FTP servers has MLSx support and how much they are in use.

> you might try the MLST and MLSD requests to gather directory
> information; they return what stat() requires in a uniform form (even
> use GMT).  For details on those, see <draft-ietf-ftpext-mlst-16.txt>.

That would bloat the design a bit. Currently, I always use

DIR -> output -> parser 1
DIR -> output -> parser 2
DIR -> output -> parser 3
...

where the parser depends on the server type (ms or unix).

If I use MLST and MLSD I have to add a "branch" instead of the DIR
invocation above: If MLSx are supported, use them and parse the result
with a MLSx parser. If not, use DIR and parse with the ms or unix
parsers (or possibly others in the future).

I wonder if it's worth the effort and complication of the design to
implement this because only one user of ftputil has complained that
neither of the current two parsers worked. He sent me a pure-Python
ftpparse.py module that supports several more formats but I couldn't use
it because it was GPL'd and I wouldn't have been able to include it in
ftputil which is BSD-licensed.

There were some parser problems for other users, but they resulted from
the wrong parser being used, not that neither of them was usable at all.

On the other hand, it _might_ be that people haven't been able to use
ftputil due to its current parsing limitations and never sent a feature
request.

ftputil development is rather driven by what _actual_ problems its users
have, not what _might_ be useful for someone. I follow the YAGNI
principle (see http://en.wikipedia.org/wiki/YAGNI and for a discussion
http://www.c2.com/cgi/wiki?YouArentGonnaNeedIt ) if I can.

Stefan