The OS is Unix-like, but the filesystem is not. The 'real' directory path delimiter is ':' not '/'. OSX makes a good job of presenting ':' as '/' and accepting '\' as ':' where it should. Maybe working though URL-encoding is one case where it or wget gets lost.

David


At 21:27 +1100 31/12/05, John Horner wrote:
I hope you don't mind me asking a wget question. I have a feeling it may hang on some Mac-specific issue which I'm not getting.

I'm trying to download http://www.guardian.co.uk/pda/avantgo/ for offline reading. I'm using wget's --restrict-file-names flag to make sure that problem characters aren't written to file names.

Some of the links, for instance "Top Stories", still don't work when browsing the files.

I'm following a link to:

file:///path/to/0,11725,ID%5E3983240;NAME^Top+stories,00.html

but Firefox can't access it.

The file is there, as

file:///path/to/0,11725,ID%5E3983240%3BNAME%5ETop+stories,00.html

where the colon and caret have been URL-encoded.

Is this a bug in wget? It's sanitising the filename but not linking to it correctly?

I'm asking here because someone may know the answer, but also because --restrict-file-names=unix is supposed to fix all problems with filenames on "Unix-like OS'es" according to the manpage.

--
David Ledger - Freelance Unix Sysadmin in the UK.
Chair of HPUX SysAdmin SIG of hpUG technical user group (www.hpug.org.uk)
[EMAIL PROTECTED]
www.ivdcs.co.uk

Reply via email to