diff options
author | Frechdachs <Frechdachs@users.noreply.github.com> | 2016-02-09 22:53:52 +0100 |
---|---|---|
committer | Frechdachs <Frechdachs@users.noreply.github.com> | 2016-02-09 22:53:52 +0100 |
commit | ec91bf3e57c19913cc93ac25587e4c44a74b588e (patch) | |
tree | abe657510817d3b9740fa478beac8bcc4f8cc7f9 /LICENSE | |
parent | 694a2c2c622b9772fadca8fd082f97419819f54c (diff) | |
download | python-mpv-ec91bf3e57c19913cc93ac25587e4c44a74b588e.tar.gz python-mpv-ec91bf3e57c19913cc93ac25587e4c44a74b588e.tar.bz2 python-mpv-ec91bf3e57c19913cc93ac25587e4c44a74b588e.zip |
Fix unobserve_property
There were three problems:
1. MPV.unobserve_property called _mpv_observe_property instead of _mpv_unobserve_property.
2. _mpv_unobserve_property returns the number of properties that were assigned to the handler that is being removed. Because the return value is not 0 in such a case, ErrorCode.raise_for_ec tries to raise an error. To fix that, I changed the ErrorCode.raise_for_ec functon not to raise an error if ec is larger than 0. (If there is a positive return value, there should not have been an error anyway, if I'm not mistaken.)
3. Calling MPV.unobserve_property for a handle that is not currently been used, should not result in a KeyError being raised.
An alternative to 2. would be to add a restype to the _handle_func of 'mpv_unobserve_property' and returning that value in MPV.unobserve_property. (That way, raise_for_ec is never called.) But I don't think this value is useful in any way. Even the built in lua interface does not return that value.
Diffstat (limited to 'LICENSE')
0 files changed, 0 insertions, 0 deletions