negative-refspec: fix segfault on : refspec
The logic added to check for negative pathspec match by c0192df630 (refspec: add support for negative refspecs, 2020-09-30) looks at refspec->src assuming it is never NULL, however when remote.origin.push is set to ":", then refspec->src is NULL, causing a segfault within strcmp. Tell git to handle matching refspec by adding the needle to the set of positively matched refspecs, since matching ":" refspecs match anything as src. Add test for matching refspec pushes fetch-negative-refspec both individually and in combination with a negative refspec. Signed-off-by: Nipunn Koorapati <nipunn@dropbox.com> Signed-off-by: Junio C Hamano <gitster@pobox.com>
Nipunn Koorapati committed
Dec 22, 2020 at 03:58 UTC
18f9c9884582c743d8ba04ef5cbbe647947d2578
2 files changed
+58
-3
remote.c
+7
-3
@@ -755,9 +755,13 @@ static int query_matches_negative_refspec(struct refspec *rs, struct refspec_ite
755
756
if (match_name_with_pattern(key, needle, value, &expn_name))
757
string_list_append_nodup(&reversed, expn_name);
758
- } else {
759
- if (!strcmp(needle, refspec->src))
760
- string_list_append(&reversed, refspec->src);
758
+ } else if (refspec->matching) {
759
+ /* For the special matching refspec, any query should match */
760
+ string_list_append(&reversed, needle);
761
+ } else if (!refspec->src) {
762
+ BUG("refspec->src should not be null here");
763
+ } else if (!strcmp(needle, refspec->src)) {
764
+ string_list_append(&reversed, refspec->src);
765
}
766
}
767
t/t5582-fetch-negative-refspec.sh
+51
@@ -186,4 +186,55 @@ test_expect_success "fetch --prune with negative refspec" '
186
)
187
'
188
189
+test_expect_success "push with matching : and negative refspec" '
190
+ # Manually handle cleanup, since test_config is not
191
+ # prepared to take arbitrary options like --add
192
+ test_when_finished "test_unconfig -C two remote.one.push" &&
193
+
194
+ # For convenience, we use "master" to refer to the name of
195
+ # the branch created by default in the following.
196
+ #
197
+ # Repositories two and one have branches other than "master"
198
+ # but they have no overlap---"master" is the only one that
199
+ # is shared between them. And the master branch at two is
200
+ # behind the master branch at one by one commit.
201
+ git -C two config --add remote.one.push : &&
202
+
203
+ # A matching push tries to update master, fails due to non-ff
204
+ test_must_fail git -C two push one &&
205
+
206
+ # "master" may actually not be "master"---find it out.
207
+ current=$(git symbolic-ref HEAD) &&
208
+
209
+ # If master is in negative refspec, then the command will not attempt
210
+ # to push and succeed.
211
+ git -C two config --add remote.one.push "^$current" &&
212
+
213
+ # With "master" excluded, this push is a no-op. Nothing gets
214
+ # pushed and it succeeds.
215
+ git -C two push -v one
216
+'
217
+
218
+test_expect_success "push with matching +: and negative refspec" '
219
+ test_when_finished "test_unconfig -C two remote.one.push" &&
220
+
221
+ # The same set-up as above, whose side-effect was a no-op.
222
+ git -C two config --add remote.one.push +: &&
223
+
224
+ # The push refuses to update the "master" branch that is checked
225
+ # out in the "one" repository, even when it is forced with +:
226
+ test_must_fail git -C two push one &&
227
+
228
+ # "master" may actually not be "master"---find it out.
229
+ current=$(git symbolic-ref HEAD) &&
230
+
231
+ # If master is in negative refspec, then the command will not attempt
232
+ # to push and succeed
233
+ git -C two config --add remote.one.push "^$current" &&
234
+
235
+ # With "master" excluded, this push is a no-op. Nothing gets
236
+ # pushed and it succeeds.
237
+ git -C two push -v one
238
+'
239
+
240
test_done