July 2010 - implementations-list - lists.ourproject.org

Bug: dired bindings break Ex commands.
by Serguei Jidkov 24 Jul '10

24 Jul '10

Hi everyone! I'm yet another long time vim user trying to make a transition to emacs. Recently I've discovered vimpulse and, suddenly, emacs started to look much more usable. :) Thanks for making this great package! Sure, it's not without its quirks but I still find it life-saving (or fingers-saving, at the very least). Here is a little bug in the current git master I've discovered just before switching vimpulse-wants-vi-keys-in-* variables off. In vimpulse-compatibility.el: > (eval-after-load 'dired > '(when vimpulse-want-vi-keys-in-dired > ;; [... snip ...] > (add-to-list 'ex-token-alist '("e" (epa-dired-do-encrypt))) > (add-to-list 'ex-token-alist '("s" (epa-dired-do-sign))) > (add-to-list 'ex-token-alist '("v" (epa-dired-do-verify))) > (add-to-list 'ex-token-alist '("d" (epa-dired-do-decrypt))) > (viper-modify-major-mode 'dired-mode 'vi-state map)))) I guess the idea was to add buffer local Ex commands, but it doesn't work. Once dired mode is entered, commands get overwritten globally. After that every attempt to use :e fails with "dired-move-to-filename: No file on this line" message. Of course unsetting `vimpulse-wants-vi-keys-in-dired' "solves" the problem, as does removing of the `add-to-list' expressions. But maybe you will want to do something more creative about it. :)

2 1

XEmacs has no [remap].
by Štěpán Němec 23 Jul '10

23 Jul '10

Subject line says it all. I learned about that only today accidentally from this mail on emacs-devel: <http://permalink.gmane.org/gmane.emacs.devel/127667> Vimpulse uses remapping at a few places so I thought I'd mention that, as AFAIK Vegard tries to keep things XEmacs-compatible (and has my full moral support ;-D). I'm not suggesting we do anything about that, at least until some complaints occur. (:rant It's so sad, really -- I'd love to be using XEmacs (see the recent copyright thread on this list), but it seems to only work reasonably well out-of-the-box in the terminal for me. I'm sure you can get it working in GUI, too, but my hostility towards the FSF probably just isn't big enough yet. :-|) Štěpán

2 4

[PATCH] Bugfix: no longer selects next word when invoking 'viw' at end of a word
by Tim Harper 23 Jul '10

23 Jul '10

--- vimpulse-text-object-system.el | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/vimpulse-text-object-system.el b/vimpulse-text-object-system.el index bdfea28..33acc59 100644 --- a/vimpulse-text-object-system.el +++ b/vimpulse-text-object-system.el @@ -410,6 +410,7 @@ specifies whether to include the quote marks in the range." (viper-backward-word (cons arg ?r)))) (lambda (arg) (vimpulse-limit (line-beginning-position) (line-end-position) + (backward-char) (viper-end-of-word (cons arg ?r)))))) (vimpulse-define-text-object vimpulse-a-Word (arg) -- 1.7.1.1

2 1

[PATCH] ignore elc files
by Tim Harper 23 Jul '10

23 Jul '10

--- .gitignore | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) create mode 100644 .gitignore diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..c531d98 --- /dev/null +++ b/.gitignore @@ -0,0 +1 @@ +*.elc -- 1.7.1.1

2 1

[PATCH] Fixes deactivate-mark issue when expanding selection to text-objects
by Tim Harper 23 Jul '10

23 Jul '10

When CUA mode is active, commands such as viw, viW, vis, etc. would exit visual mode because cua-mode, on activation, sets deactivate-mark to t, and deactivate-mark isn't protected by save-excursion. This change prevents that, and therefore fixes this issue. --- vimpulse-visual-mode.el | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/vimpulse-visual-mode.el b/vimpulse-visual-mode.el index a9c0240..2430d0a 100644 --- a/vimpulse-visual-mode.el +++ b/vimpulse-visual-mode.el @@ -281,9 +281,10 @@ Saves the previous state of Transient Mark mode in (transient-mark-mode 1) (transient-mark-mode -1))) (when (boundp 'cua-mode) - (if (vimpulse-visual-before cua-mode) - (cua-mode 1) - (cua-mode -1))) + (let (deactivate-mark) ;; prevent cua-mode from setting deactivate-mark to t + (if (vimpulse-visual-before cua-mode) + (cua-mode 1) + (cua-mode -1)))) (when (boundp 'zmacs-regions) (let ((oldval (vimpulse-visual-before zmacs-regions))) (setq zmacs-regions oldval))))) -- 1.7.1.1

2 2

Strange visual-line bug
by Tim Harper 23 Jul '10

23 Jul '10

After a while of using vimpulse mode in a buffer, it gets into a strange state where if I go from vi-mode to visual-line-select mode by pressing "V", it "remembers" my previous visual-line-selection (cursor position, number lines selected and all) and takes me back to it. This is undesirable, naturally. I don't know exactly how to reproduce the bug, but have managed to record a video of it in action: http://screencast.com/t/YjBlMmVkMjQ The buggy-state is buffer-local. Killing a buffer, and then re-opening the file, restores sanity, as demonstrated in the video. Thanks for vimpulse mode! Tim

2 2

Re: Strange visual-line bug
by Tim Harper 22 Jul '10

22 Jul '10

Hah, that's the coolest report so far? Awesome! seriously I use video because I'm lazy... its a ton of work to serialize an experience into text. I'm using GNU emacs, 23.2. (Cocoa integration was merged in for 23) Interesting that its caused by visual line mode being deactivated properly... before I apply your patch, ill play around a little more and see if I can see what leads to it, but being fault tolerant never hurts... it's a jungle in emacs. Tim On Jul 21, 2010 3:40 AM, "Vegard Øye" <vegard_oye(a)hotmail.com> wrote: On 2010-07-20 16:12, Tim Harper wrote: > I don't know exactly how to reproduce the bug, but have ma... Hi Tim! Welcome to the list, and thanks for the coolest bug report so far :) Let's get the basics out of the way first. Is that Aquamacs (what version?), and are you using the latest developmental commit of Vimpulse? I don't know exactly why this happens, but I've been able to reproduce it by injecting a "bad value" into one of Visual mode's buffer-local variables. It seems that somehow Visual mode isn't deactivated properly and that a faulty variable causes problems upon reactivation. The variable is remembered across Visual sessions, giving rise to a persistent "buggy state". Furthermore, as the restoration of the previous selection occurs in `post-command-hook' -- after hl-line-mode et al. have refreshed their own highlighting -- the other modes aren't able to pick up on it, and so their highlighting doesn't coincide with Visual mode's. Commit [9d1af90898] should be more robust with regard to such values. Let me know if the bug persists. -- Vegard _______________________________________________ implementations-list mailing list implementations-list(a)lists.ourproject.org https://lists.ourproject.org/cgi-bin/mailman/listinfo/implementations-list

3 5

Thanks!
by Marc Weber 21 Jul '10

21 Jul '10

Hi list, I've been using Vim for years. However I'm at a point I think Vim will no longer serve (all) my needs - that's because the core is lacking functionality - which can be implemented - but would take a lot of time. I want to thank you - I was very pleased seeing that these features just work: - ctrl-vI..<esc> (instert into multiple lines) - text objects - all visual modes - % in combination with visual mode. So Emacs comes pretty close to what I'm used to (which I use most) There is one last thing which I'm missing: Opening files by glob patterns: noremap :e :e<space>**/* In #emacs I was told use ido-find-file .. However patterns like this **/*Main.scala doen't work for me. creating a buffer with output of find then using gf (which works!) is not a perfect solution. Having gF implemented properly would be fine, because I use it often when reading Exception logs - I can still fallback to Vim. Do you have some more hacks in you ~/.emacs making using Emacs easier for Vim addicts? Eg is there a :qa! like command (without the nasty: You have unsaved open files..) Marc Weber

3 7

[Vimpulse] jumping to marks with `` and ''
by Stephen Bach 18 Jul '10

18 Jul '10

Hi, I notice that jumping to the previous mark with `` or '' turns on some kind of range highlighting (which it doesn't when jumping to a specifically-named mark, e.g. 'a or `b). I think this is a long standing bug, not a new regression.

2 2

[Vimpulse] region-active-p
by Stephen Bach 17 Jul '10

17 Jul '10

Hi, Looks like somewhere along the line Vimpulse broke compatibility with GNU Emacs 22.3 (for which it claims support in the Commentary). Fixing it may be as simple as defining a replacement for region-active-p, which only exists in GNU Emacs 23.

2 1