Hi, Yann Dirson wrote:
> In many cases one can get the index statinfo out of sync with > worktree, resulting in eg. gitk listing all files as modified, > although it shows most of them with no real change - imaging what that > gives on eg the kernel tree: gitk effectively unusable. Of course "git add --refresh ." can be used as a workaround just like "git status" can. Do you remember what particular actions are resulting in the stat information going awry? You may also be interested in this discussion: [1]. > Usually just running "git status" fixes the problem, but it happens > 1.7.1 has a serious regression and does not run it. This is fixed in > commit 4bb6644d03f6932b94c24c3825e28865f493b692 "git status: refresh > the index if possible". We should at least cherry-pick that one. That would not be so bad as long as we take v1.7.2~25^2 (do not write out the index when status does not have to, 2010-07-06) at the same time. Anyway, my current hope is to go with 1.7.2.something. If that does not work out, a list like this of patches to use on top of v1.7.1 would be helpful indeed. Thanks for the report (and thanks for all your work on git!), Jonathan [1] http://thread.gmane.org/gmane.comp.version-control.git/144176/focus=144189 -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
