[vim/vim] Macros don't correctly record mapped Ctrl keys (Issue #11270)
David Briscoe
Steps to reproduce
Mapping C-L to Esc doesn't work correctly in macros:
- gvim --clean
:inoremap <C-L> <Esc>- populate buffer:
ivim<CR>devs<C-L> - record a macro using C-L:
ggqqIhi <C-L> - replay macro:
j@q
Result:
hi vim
hi �üLdevs
Expected behaviour
When playing a macro, mappings should work the same as they while recording the macro. Using Ctrl in a mapping shouldn't replay as special characters. This was the behaviour on vim 8.2. I can verify that it works this way on 8.1.2269 (in a terminal) and I think it was also correct on 8.2.3399.
Expected:
hi vim
hi devs
Version of Vim
9.0.636
Environment
Win 10 21H2
gvim
Using build of vim from scoop (installer from https://github.com/vim/vim-win32-installer)
Logs and stack traces
No response
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
zeertzjq
This fixes the bug:
diff --git a/src/map.c b/src/map.c index d063e053a..3463adcf1 100644 --- a/src/map.c +++ b/src/map.c @@ -1753,7 +1753,11 @@ vim_strsave_escape_csi(char_u *p) d = res; for (s = p; *s != NUL; ) { - if (s[0] == K_SPECIAL && s[1] != NUL && s[2] != NUL) + if ((s[0] == K_SPECIAL +#ifdef FEAT_GUI + || s[0] == CSI +#endif + ) && s[1] != NUL && s[2] != NUL) { // Copy special key unmodified. *d++ = *s++;
Not sure if it is possible to write a test.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
David Briscoe
Wow, fast find! And I didn't realize it was gui-only, but testing in terminal vim9 I don't can't repro.
For tests, https://github.com/vim/vim/blob/master/src/testdir/test_gui.vim doesn't seem to have any macro tests. I was using let @q = to convert my broken macros into working ones, so that's not a valid way to test. feedkeys("qqIhi\<C-L>q@q") doesn't work and feedkeys("qqIhi\<C-L>q@q", "L") (or "t") don't exhibit the bug.
Not sure how you can test without a script that sends keys to vim.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
zeertzjq
Note: unlike a previous issue, the #ifdef FEAT_GUI in the patch above probably cannot be omitted.
With FEAT_GUI, CSI is just an ordinary character, so removing the #ifdef FEAT_GUI can cause problems like treating a part of a UTF-8 sequence as a part of a GUI termcap when FEAT_GUI is not defined.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
zeertzjq
Note: unlike a previous issue, the #ifdef FEAT_GUI in the patch above probably cannot be omitted.
Without FEAT_GUI, CSI is just an ordinary character, so removing the #ifdef FEAT_GUI can cause problems like treating a part of a UTF-8 sequence as a part of a GUI termcap when FEAT_GUI is not defined.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
zeertzjq
I guess the easiest way to test this is to just hardcode the register content. I've create a PR #11275.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
Bram Moolenaar
Closed #11270 as completed via 2cd0f27.
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.
David Briscoe
Fix works for me on 9.0.662. It even correctly plays macros recorded on 9.0.636.
Thanks @zeertzjq!
—
Reply to this email directly, view it on GitHub.
You are receiving this because you are subscribed to this thread.