-
-
Notifications
You must be signed in to change notification settings - Fork 31.6k
IDLE : Bug in keybinding validity check #65718
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
Steps to reproduce the bug:
Now the <Control-Key-C> binding will be assigned to two actions. This bug only applies if the other action has more than one binding. Attaching a patch to fix this issue. |
I do not see this. I only see <Control-Key-c> duplicated, which is bad enough to be patched. What did you leave out ;-). Also see bpo-12387, which we need to finish, and Ned's link to http://wiki.tcl.tk/28331 A patch for KeysOK should include test_bindings.py with class KeysOK_Test. As with the test of name_ok in test_configname.py (sp?), the method should be embedded in a dummy class and use a dummy message box. The problem with the method goes deeper. It is not called if the 'advanced' (hand entry, therefore primitive) pane is used. As near as I can tell (and if I am wrong, please tell me), multiple bindings, which are required to ignore lower-upper case issues, can only be entered with this 'advanced' method. Hence, most new entries are not checked at all. The basic pane, which is advanced in that one can only enter possibly valid combination, should be modified to allow multiple entries. (The two panes should be called 'Easy' and 'Error-prone ;-) As it is, 'keys' is a misnomer[note], "keys.strip()" is a no-op in that 'keys' is auto-generated, and "keySequence = keys.split()" is the same as "keySequence = [keys]"[note]. Many of the tests only test mistakes that can happen with hand-entry, which is not tested. I believe that the duplicate check is the only one needed (except possible for the modifier check, which I have not looked at enough to tell). [note] If the easy pane were modified to define multiple keys, and all were passed at once, then the two noted lines would become valid, but then each key, would have to be tested individually against the flat list. This suggests to me that the flattening should be done just once and that KeysOK should become KeyOK, or rather key_ok. It might even be replaced by "newkey in flat_list" or something nearly that simple. |
A small bug in line 185 in keybindingDialog.py. 'F2' appears twice and 'F3' is missing. Since this is a typo, I did not create a new issue. |
bpo-6739 may also be relevant |
bpo-6739 also has a patch to refuse invalid key bindings. |
I believe patch affects same area of file as patch for bpo-6739. |
bpo-6739 is about rejecting *invalid* sequences, this is about rejecting a *duplicate* valid sequence. Both fixes are needed. |
configdialog.ConfigDialog.getNewKeys() calls config_key.GetKeysDialog with a list of lists of one or more sequences (currentKeySequences). GetKeysDialog.KeysOK looks for keys.split() in currentKeySequences. Since KeysOK on only called for the no-space 'keys' produced by the basic dialog, keys.split is always [keys]. This can only match one of the length-1 lists in currentKeySequences. The patch semi-flattens the latter to a list of lists of 1 sequence each A separate improvement would be to have a reverse map of sequences to pseudoevents, so the error message can specify the exact conflict instead of just saying 'some conflict'. I have separately thought that the [keys] tab of ConfigDialog should be able to display such a mapping to help one plan a new key mapping. Off topic for this issue: the keys tab could have a 'load key-def file' function that would check a definition and if ok, load it. |
Thanks for the patch. The F3-missing bug has been fixed. |
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
bugs.python.org fields:
The text was updated successfully, but these errors were encountered: