Looking very nice. midori_browser_download_status_cb still should be more robust. It needs to check if the download failed, otherwise it could be waiting forever. And it should show a UI error rather than g_error, which would be an unexpected crash in the perception of the user.
I like user-specific temporary folders here. Nice idea from a data privacy point of view and reducing noise in /tmp.
I suggest these follow-ups after getting the initial patch in, they can be separate bugs:
- Unit tests for midori_browser_download_prepare_filename
- GTK+ 3 app choice fallback
- Use new functions in existing download opening code
We may be having a release this week and I would wait with this feature so we don't get a new very visible string just before release.
Looking very nice. midori_ browser_ download_ status_ cb still should be more robust. It needs to check if the download failed, otherwise it could be waiting forever. And it should show a UI error rather than g_error, which would be an unexpected crash in the perception of the user.
I like user-specific temporary folders here. Nice idea from a data privacy point of view and reducing noise in /tmp.
I suggest these follow-ups after getting the initial patch in, they can be separate bugs: browser_ download_ prepare_ filename
- Unit tests for midori_
- GTK+ 3 app choice fallback
- Use new functions in existing download opening code
We may be having a release this week and I would wait with this feature so we don't get a new very visible string just before release.