On 7/11/2013 11:26 PM, Magnus Ihse Bursie wrote:
On 2013-11-06 03:01, Jonathan Gibbons wrote:
Pwrhaps we should only have one copy of the webrev script around.
... and perhaps it should not be tied to a specific source branch. That
has always seemed weird to me.
I would suggest creating a code-tools/webrev repository, and moving
webrev.ksh there.
Create it where exactly?
Having a copy of webrev in the openjdk forest is useful for people to
get their hands on it. But webrev is not specific to openjdk or a
particular forest, it is just a tool. I keep a copy of webrev in ~/bin
because I use it over numerous forests and I always want the same
version of webrev used (latest one or at least late enough to have the
features I like). I then have to update it periodically.
David
-----
To keep backwards compatibility, it's relatively simple to replace
make/script/webrev.ksh with a thin wrapper that does:
1) check for the existence of a hg repo in e.g.
common/code-tools/webrev, and if not, clones it, otherwise, does a hg
pull -u.
2) sanity checks that common/code-tools/webrev/webrev.ksh exists
3) calls common/code-tools/webrev/webrev.ksh with $@.
/Magnus
