Daniel Rall wrote:
> On Thu, 31 May 2007, email@example.com wrote:
>> When searching for child mergeinfo, look for proper children of the given path,
>> not just paths with a common prefix. Followup to r25138.
>> * subversion/libsvn_fs_util/mergeinfo-sqlite-index.c
>> (get_mergeinfo_for_children): Select only paths which are proper children of
>> the requested path.
>> --- trunk/subversion/libsvn_fs_util/mergeinfo-sqlite-index.c (original)
>> +++ trunk/subversion/libsvn_fs_util/mergeinfo-sqlite-index.c Thu May 31 07:50:12 2007
>> @@ -619,7 +619,7 @@
>> "WHERE path LIKE ? AND revision <= ?;",
>> -1, &stmt, NULL), db);
>> - like_path = apr_psprintf(pool, "%s%%", path);
>> + like_path = apr_psprintf(pool, "%s/%%", path);
>> SQLITE_ERR(sqlite3_bind_text(stmt, 1, like_path, -1, SQLITE_TRANSIENT), db);
>> SQLITE_ERR(sqlite3_bind_int64(stmt, 2, rev), db);
> Looks good, Hyrum. Is our test case exercising this yet? (I think it
> should, to prevent regression.)
Not directly, but the merge sensitive log functionality depends on this
function, so the test(s) are hitting this function indirectly. Do we
have a simple way to directly test this function?
Received on Thu May 31 20:16:05 2007