[Clipart] HASH bug: diagnosed (at least partly)

[Jonadab the Unsightly One]

"Jonadab the Unsightly One" <jonadab at bright.net> writes:

> Oh, yeah, we're cooking with gas, now.  *Tangible* progress.

Okay, I've gotta sleep now and work tomorrow, so here's what's done:

I commented out the debugging stuff (but left it there as comments for
the moment), and the keywords should not get HASH in them anymore,
unless they come that way before Metadata.pm gets its hands on them.
Additionally, I've placed a comment in Metadata.pm explaining why
files with id attributes on all of their elements don't have their
metadata parsed correctly, and how it can be fixed, but I haven't
actually done that yet and probably won't get to it until after the
release -- which means some images won't be fully fixed until after
this release.  But some of the files *can* be fixed with things as
they stand now, so I checked in Metadata.pm like that.

> There's still stuff to do.  The keywords, for instance, are not the
> only (just the most obvious) place where the HASH issue was popping
> up, so I need to fix a couple of other places too.  That shouldn't
> be a problem, now that I have a quite clear understanding of what
> was going wrong.

Yet somehow it didn't all get done today.

> And then I need to take the log excerpts I've harvested and apply
> them to the dammaged files, from 0.13 through the present.  (I have
> a repair script that does this, but so far it only fixes the
> keywords; I need to make it fix everything that was broken.)

Part of this is done.  I think it is usable enough to give us a
significant improvement over the last release.

-- 
$;=sub{$/};@;=map{my($a,$b)=($_,$;);$;=sub{$a.$b->()}}
split//,"ten.thgirb\@badanoj$/ --";$\=$ ;-> ();print$/