On Sun, Apr 08, 2012 at 12:36:05PM +0100, Phil Holmes wrote:
OK. I'm in the final throes of updating the LSR and have found 2
issues exemplified by a single file. In snippets/new we used to
have a file called screech-boink.ly, with a doctitle of "Screech and
Boink" and an LSR tag of "headwords".
First problem. If this is added to the LSR with that Title, the
filename produced in the tarball is screech-and-boink.ly, which is
I approve of changing the filename.
So it would seem the docs would need updating. But if I
git grep screech-boink I get 8 itely files (all the translations),
10 po/pot files plus a lytex file. Do all these need updating? Is
there a quick grep-py way of doing it if so?
- add the new filename
- change it in the English docs
- tell the translators to update their stuff
- add a Critical issue for this update
- when the translators are done, remove the old filename.
Second problem. makelsr.py uses an explicit list of directories
where it takes snippets from to add them to Doc/snippets. It's
So it doesn't even find screech-and-boink.ly, since it's only in the
headwords directory. Does anyone know why makelsr doesn't just
iterate over all the subdirectories in the tarball, rather that
restricting itself to a specific list?
I didn't know about glob.glob() when I first wrote makelsr.py ?
I can't think of any other reason.
This issue could be fixed 3 ways: add a further tag to the snippet
in the LSR; extend the list of tags; or use all the files in the
directory. I'd prefer the latter, since the current system is
confusing - it took me about an hour to work out why screech and
boink wasn't in the snippet listing.
I don't like the first one. I'm fine with having the list of tags
extended further or looking at all subdirectories of the lsr
directory. I think I have a slight preference for just extending
the list of tags, since that's a less intruisive change and thus
has less chance of blowing up in our faces. But since you prefer
looking at all subdirectories, that's fine with me.