🚀 go-pugleaf

In article <O5R6LX@math.fu-berlin.de>, heiko@Harpo.Chemie.FU-Berlin.DE (Heiko Schlichting) writes:
|> Because the connection to Berlin are so slow, here are the first 20
|> lines from this INDEX to show the format:
|> 
|> #NAME           ftp.FU-Berlin.DE
|> #ADDRESS        130.133.4.50
|> #ORGANIZATION   Freie Universitaet Berlin
|> #LOCATION       Berlin, Germany; 52 31 N / 13 24 E
|> #ACCESS         no restrictions
|> #MAIL           ftp-adm@FU-Berlin.DE
|> #SERVER-TIMEZONE +0200 (MET DST)
|> #
|> #INDEX-TIMEZONE +0000 (UTC)
|> #VERSION        1.0 (28-Aug-1992); de-mirror INDEX format
|> #CREATED        07-Sep-1992 01:00
|> #
|> DR-X          0 24-Aug-1992 00:02 /
|> DRWX          0 04-Sep-1992 01:43 /incoming
|> DRWX          0 24-Aug-1992 16:57 /incoming/amiga
|> FRW-       4435 01-Jul-1992 16:48 /incoming/amiga/JOES.amiga.sites
|> FRW-     583680 24-Aug-1992 16:57 /incoming/amiga/MED_Songs.dms
|> FRW-        452 24-Aug-1992 16:57 /incoming/amiga/MED_Songs.readme
|> FRW-     168103 03-Jul-1992 12:23 /incoming/amiga/ReOrg_V2.3.lha
|> FRW-     210243 07-Jul-1992 17:51 /incoming/amiga/vortaro.lha
|> 
|> 
|> This format was discussed in the German FTP-mirror mailing list
|> (de-mirror@informatik.tu-muenchen.de, to join this list write to
|> de-mirror-owner@informatik.tu-muenchen.de - but the conversation language 
|> in this list is German). The name for this format comes from this
|> mailing list: "de-mirror INDEX format".

Sorry for the long delay, but I've been on holidays ...

Ok, some comments!

Unfortunately the Admins in Berlin could not be convinced to use a numeric
only date information! I have argued with better sorting capatibilities ...
no chance. So again my call: PLEASE USE ISO-DATE (see the FAQ for this group,
topic 2 as far as I can remember!)

Another topic I'd like to see some discussion about is the naming of files and
directories!
Here in Munich we used the above format at first :-)
We had the philosophy that no whitespace nor backspace should be allowed in
directory names or file names for the Unix-style INDEX.
This made the first character (F/D/L for File, Directory, Link) unnecessary,
as we added a '/' at the end of the name if it is a directory, a link is
recognisable by the name being three names with the 2nd name being "->".
The Admins in Berlin argued that with whitespace allowed it would at least
fail for links. I argued that even with the 'L' in front, you could not
exactly decide what the filename and what the linkname is in the following
example:   /a/b -> /c -> /d/e/f. 
We also omitted the RWX information, as according to our local philosophy
each directory called tmp, incoming or uploads or the like should be
writeable for everyone and all other directories should be read only.
x-permission for directories are always needed, cause the directory would be
worthless otherwise and should not appear in the INDEX file.
Same is for files without r-permission! They are worthless and shouldn't show
up in the INDEX file!
For this reasons I think the RWX information is superfluous.
I also tried to convince them not to allow comment-lines ... no chance.

Why do I argue again?
I want a format that is clear, simply and as short as possible and such
easy to parse.

Currently I use this format to generate daily diff files. These files can
easily be used to generate statistics on what has changed, what is new, what
has been deleted and how the size of the archives behaves ... Just a simple
awk script ;-)

Comments?

	\Maex
-- 
______________________________________________________________________________
 Markus Stumpf                        Markus.Stumpf@Informatik.TU-Muenchen.DE