[darcs-users] EOL for darcs-1 repository format [was: soc final report]
Trent W. Buck
trentbuck at gmail.com
Sun Aug 16 11:04:56 UTC 2009
Petr Rockai <me at mornfall.net> writes:
> Recently, it has been noted, that for old-fashioned repositories, darcs
> whatsnew has regressed in performance by a fair amount. This is true, but it's
> not clear if it is worth addressing properly. For 2.3.1, the easy fix is to
> restore the old code path when the repository is not hashed. However, I have
> already removed unsafeDiff in darcs-hs and I want to keep it that way, so we
> either need to start treating old-fashioned as second-class, or either come up
> with a more systematic fix.
Eric, I think we need a policy decision on this.
When can/should we EOL the darcs-1 repository format?
More information about the darcs-users