A watchlist is a mechanism that allows a developer (a "watcher") to watch over portions of code that the watcher is interested in. A watcher will be cc-ed on changes that modify that portion of code, thereby giving that watcher an opportunity to make comments on chromium-review.googlesource.com even before the change is committed.
Watchlists are defined using a WATCHLISTS
file, which resides at the root of a
repository. A typical WATCHLISTS
file looks like:
{
'WATCHLIST_DEFINITIONS': {
'valgrind': {
'filepath': 'tools/valgrind/',
},
'mac': {
'filepath': 'cocoa|\.mm$|(_mac|_posix)\.(cc|h)$',
},
},
'WATCHLISTS': {
'valgrind': ['[email protected]', '[email protected]'],
},
}
In this case, watchlists named valgrind
and mac
are defined in
WATCHLIST_DEFINITIONS
and their corresponding watchers declared in
WATCHLISTS
.
In the example above, whenever a new changeset is created that refers to any
file in tools/valgrind/
, the 'valgrind'
watchlist will be triggered and
[email protected]
& [email protected]
will be cc-ed to the changeset
for review. A regular expression can be used as the matching pattern. Matches
are determined using python's re.search()
function call, so matching A_WORD
is the same as matching .*A_WORD.*
.
Each name in WATCHLISTS
must be defined first in WATCHLIST_DEFINITIONS
.
Watchlist processing takes place on Gerrit with the "Watchlists" analyzer and is non-binding; that is, an approval from that watcher is not needed for commit. It merely gives the watcher an opportunity to make comments, if any.
You create new watchlists or add yourself to existing watchlists by editing the WATCHLISTS file at the base of the repository.
It's advisable to run watchlists.py
to verify that your new rules work.
Example (from src):
python third_party/depot_tools/watchlists.py PATH/TO/FILE1 PATH/TO/FILE2 ....