Ian Abbott wrote:
> 

[snipped]

> I'd say ignore my previous "proposal", fix what's there already, and then maybe
> tune the safe/unsafe character handling for
> particular file-systems (with options to override this) later.

I'm not a developer, but it seems to me that a problem is more general
than just safe/unsafe character handling/mapping.

>From my user's point of view there are two types of resources: remote
and local. Both types of resources have names. Remote resource names I
type when I invoke wget and I see them inside HTML cross references.
Local resource names I see (or wish to see) in my local file system
after wget finishes. Remote resources may have naming conventions
different than local ones.

What I need as a user that is an unambiguous mapping between remote and
local resources. I don't mind about any internal representation. I only
want to know rules which allow me to find a local resource which
corresponds to a remote one and vice versa.

>From HTTP perspective I'd expect wget not only a) to download remote
resources, b) assign them local names and c) save them locally. I'd also
expect it 6) to fix cross references in locally saved resources. Less or
more these all have already done.

>From FTP perspective I'd still expect a), b) and c). And there is also
one point where cross referencing is used. It is .listing file. So I's
expect the same behaviour as in HTTP case.

Unfortunately at the moment, wget saves remote names to .listing file
but it doesn't fix every local names properly. For example, it fixes
space characters in local directory names to '%20' but leave them
untouched in local file names. And wget always forgets about that when
it reloads the same remote resource next time. It doesn't used the
naming conversion rules in reverse direction to determine that it has
already loaded the resource. Details I'm giving in separate message.

So, I'd try to summarize my very long monolog (forgive me for it):

1. Well-defined mapping rules are necessary (possibly tuneable by
users), so Ian's proposal looks very promising;
2. Two way unambiguous conversion remote->local and local->remote will
allow wget a) not to miss any resources and b) not to do the same job
twice.
3. The same behaviour for both HTTP and FTP.

Thank you for your patience.

Alexey

Reply via email to