[hypermail] naming files

 jose.kahan_at_w3.org (Jose Kahan) writes:
>I also want to see if I find an alternate algorithm for naming files so
>that they don't grow up sequentially. Sometimes when rebuilding an
>archive, if you have deleted or skipped a message, all the URLs change
>and it would be nice to avoid this dependency.

 Note that it is now possible to delete messages (in the sense of making them invisible to normal browsing) via the "deleted" option without changing URLs.

>I think I'll also need to store in a dbm file some relationship between
>the filename and the number we were using before, so that we can search it
>easily. This is not yet very clear, because I am just starting work on it.

 I'm uncertain whether I would want dependence on a dbm file to be the default mode of operation. I presume you would make this optional, at least when it is first introduced?

>I tried my old hypermail CVS access and not surprisingly it doesn't work
>anymore. I would like to have a write access again as it would simplify
>my work when I have to make a commit (rather than having two parallel bases).

 Ashley controlled cvs access last I heard, not Kent.

-- 
------------------------------------------------------------------------------
Peter McCluskey          | Free Dmitry Sklyarov! http://www.freesklyarov.org/
http://www.rahul.net/pcm |