mi <[EMAIL PROTECTED]> writes:
> About the output of 'wwwoffle -purge'.
>
> Maybe not much people are reading the output at all,
It isn't really there for people to read all the time. It is just
there to tell you what happened in the rare case that you are
interested.
> Some notes on the layout of the wwwoffle -purge output.
>
> What does 'not purged' mean ?
> I believe 'purging' is the process of entering a http directory to
> look if there's something to do.
> So, 'not purged' would mean 'directory not checked'. Is this correct?
Not quite, it means that nothing was deleted. There might be some
files compressed, but nothing would have been deleted. You will only
ever see this if the use-url option is disabled.
All directories are "checked" even if that just means that junk files
are deleted from them and compression takes place (if enabled).
> Now that i study the purge output more regularily, i note that i
> mainly look for 3 categories of domains, which i imnagine could be
> listed on top by wwwoffled.
Or by a script that processes the WWWOFFLE output.
> (1) The deleted domains.
> (2) The top stored (harddisk usage) domains, as a kind of percentile
> (Is this the word ? eg, x domains resembling 30% of the complete http disk
> usage.)
> (3) Default Age domains. I need to decide if some of them are more
> important and deserve a 'Hold' or higher expiring age.
> The complete domain list should be shown afterwards (redundant - why
> not), sorted alphabetically.
>
> In order to postprocess the wwwoffle-purge output, one would have to
> cut the relevant domain list ougt of the other info, usually with a
> program function. I would be glad if i could rely on this extracting
> function even with further wwwoffle versions, which might change the
> header / trailer layout. Perhaps the wwwoffle list output could be
> separated by fix markers (like LIST BEGIN and LIST END). Perhaps
> it's even possible (by an option switch) to have it logged to a file
> separatly by wwwoffled itself.
There is always the chance that the format of the purge output will
change in future versions of WWWOFFLE. I think that they way things
are at the moment there is enough information available to make a good
parsing of the output without needing to have markers at the beginning
and end of the list. The lines that you are looking for will always
match 'Purged' or 'Not purged' at the beginning of the line for
example.
> Also i would be delighted if storage sizes could be formatted 'human
> readable' (KB MB GB) wherever possible, especially in the statistics
> (summary) at the end.
By keeping the sizes in the same units everywhere it becomes very
obvious which are the big directories and which are the small ones.
You only need to look at the length of the numbers. Without this it
is possible to confuse directories with 100 KB as bigger than those
with 10 MB or 1 GB.
--
Andrew.
----------------------------------------------------------------------
Andrew M. Bishop [EMAIL PROTECTED]
http://www.gedanken.demon.co.uk/
WWWOFFLE users page:
http://www.gedanken.demon.co.uk/wwwoffle/version-2.8/user.html
