2010/7/25 Jonathan Nieder <jrnie...@gmail.com>:
> Frédéric Brière wrote:
>
>> I stumbled on this problem yesterday, working in a sparse checkout of
>> linux-next.
>
> First of all, thanks for reporting this and sorry to take so long
> to get back to you.
>
>> # See how the non-conflicting file is still in the index?
>> git status
>>
>> # And see how reset --hard does not touch it, but --mixed does?
>> git reset --hard
>> git status
>> git reset --mixed
>> git status
>
> It seems that "read-tree --reset -u" should be somehow taught
> to imply the effect of "read-tree --reset" (without -u). Duy: does
> that seem reasonable? Are there any subtleties I should expect to
> run into?
Still studying the code. But does this help?
diff --git a/unpack-trees.c b/unpack-trees.c
index 8cf0da3..32764c2 100644
--- a/unpack-trees.c
+++ b/unpack-trees.c
@@ -1439,7 +1439,7 @@ int oneway_merge(struct cache_entry **src,
struct unpack_trees_options *o)
if (old && same(old, a)) {
int update = 0;
- if (o->reset && !ce_uptodate(old) && !ce_skip_worktree(old)) {
+ if (o->reset && !ce_uptodate(old)) {
struct stat st;
if (lstat(old->name, &st) ||
ie_match_stat(o->src_index, old, &st,
CE_MATCH_IGNORE_VALID|CE_MATCH_IGNORE_SKIP_WORKTREE))
--
Duy
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
