bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

Hi,
There is a bug when using conjointly these two features. Here are the
relevant portions of configuration :
;;make ibuffer resize itself to its buffer size
(setq ibuffer-use-other-window t)
(setq ibuffer-default-shrink-to-minimum-size t)

No specific configuration for erc, just enabling tracker is enough.
Then, run erc, and, in any erc buffer, call ibuffer. The screen briefly
flashes as an ERC window opens then disappears, and the ibuffer window is
reduced to a much smaller size than what it would be outside ERC. I think
the issue is erc's hook to window-configuration-change somehow messing with
ibuffer, but I'm not experienced enough to find where the bug is.

Antoine Levitt

[-- Attachment #2: Type: text/html, Size: 750 bytes --]

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[Chong Yidong]

Hi guys,

Could you take a look at this bug report?  I'm not sure whether ERC or
ibuffer needs fixing.

> There is a bug when using conjointly these two features. Here are the
> relevant portions of configuration :
>
> ;;make ibuffer resize itself to its buffer size
> (setq ibuffer-use-other-window t)
> (setq ibuffer-default-shrink-to-minimum-size t)
>
> No specific configuration for erc, just enabling tracker is enough.
> Then, run erc, and, in any erc buffer, call ibuffer. The screen
> briefly flashes as an ERC window opens then disappears, and the
> ibuffer window is reduced to a much smaller size than what it would be
> outside ERC. I think the issue is erc's hook to
> window-configuration-change somehow messing with ibuffer, but I'm not
> experienced enough to find where the bug is.

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 1045 bytes --]

Another related problem (fixing one will probably fix the other, but I'm
posting just in case) : doing the same manipulation resets the count of
unseen messages
2008/9/2 Chong Yidong <cyd@stupidchicken.com>

> Hi guys,
>
> Could you take a look at this bug report?  I'm not sure whether ERC or
> ibuffer needs fixing.
>
> > There is a bug when using conjointly these two features. Here are the
> > relevant portions of configuration :
> >
> > ;;make ibuffer resize itself to its buffer size
> > (setq ibuffer-use-other-window t)
> > (setq ibuffer-default-shrink-to-minimum-size t)
> >
> > No specific configuration for erc, just enabling tracker is enough.
> > Then, run erc, and, in any erc buffer, call ibuffer. The screen
> > briefly flashes as an ERC window opens then disappears, and the
> > ibuffer window is reduced to a much smaller size than what it would be
> > outside ERC. I think the issue is erc's hook to
> > window-configuration-change somehow messing with ibuffer, but I'm not
> > experienced enough to find where the bug is.
>

[-- Attachment #2: Type: text/html, Size: 1423 bytes --]

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[martin rudalics]

CC-ing Keegan because Bug #466 describes another problem with ERC and
windows.

 >> There is a bug when using conjointly these two features. Here are the
 >> relevant portions of configuration :
 >>
 >> ;;make ibuffer resize itself to its buffer size
 >> (setq ibuffer-use-other-window t)
 >> (setq ibuffer-default-shrink-to-minimum-size t)
 >>
 >> No specific configuration for erc, just enabling tracker is enough.
 >> Then, run erc, and, in any erc buffer, call ibuffer. The screen
 >> briefly flashes as an ERC window opens then disappears, and the
 >> ibuffer window is reduced to a much smaller size than what it would be
 >> outside ERC. I think the issue is erc's hook to
 >> window-configuration-change somehow messing with ibuffer, but I'm not
 >> experienced enough to find where the bug is.

Can't offer much insight but processing window-configuration-change
hooks has changed with Emacs 23.  So this

(defun erc-track-modified-channels ()
   "Hook function for `erc-insert-post-hook' to check if the current
buffer should be added to the modeline as a hidden, modified
channel.  Assumes it will only be called when current-buffer
is in `erc-mode'."
   (let ((this-channel (or (erc-default-target)
			  (buffer-name (current-buffer)))))
     ...			

might have been broken for some reason.  To verify, someone would have
to try with Emacs 22 and 23 inserting just before the `let' something
like

(unless (eq major-mode 'erc-mode)
   (message "Not in erc-mode ..."))

and look whether it triggers.

martin

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 2293 bytes --]

Hi,
I did some testing, and the bug was not in tracker as I suspected, it's in
scrolltobottom, in erc-goodies.el. This should be much easier for you to
debug: it's only a few lines long. I strongly suspect the part following
this comment is responsible :
      ;; Temporarily bind resize-mini-windows to nil so that users who have
it
      ;; set to a non-nil value will not suffer from premature minibuffer
      ;; shrinkage due to the below recenter call.  I have no idea why this
      ;; works, but it solves the problem, and has no negative side effects.
      ;; (Fran Litterio, 2003/01/07)

Looks like it does have side effects after all :-)
2008/9/3 martin rudalics <rudalics@gmx.at>

> CC-ing Keegan because Bug #466 describes another problem with ERC and
> windows.
>
> >> There is a bug when using conjointly these two features. Here are the
> >> relevant portions of configuration :
> >>
> >> ;;make ibuffer resize itself to its buffer size
> >> (setq ibuffer-use-other-window t)
> >> (setq ibuffer-default-shrink-to-minimum-size t)
> >>
> >> No specific configuration for erc, just enabling tracker is enough.
> >> Then, run erc, and, in any erc buffer, call ibuffer. The screen
> >> briefly flashes as an ERC window opens then disappears, and the
> >> ibuffer window is reduced to a much smaller size than what it would be
> >> outside ERC. I think the issue is erc's hook to
> >> window-configuration-change somehow messing with ibuffer, but I'm not
> >> experienced enough to find where the bug is.
>
> Can't offer much insight but processing window-configuration-change
> hooks has changed with Emacs 23.  So this
>
> (defun erc-track-modified-channels ()
>  "Hook function for `erc-insert-post-hook' to check if the current
> buffer should be added to the modeline as a hidden, modified
> channel.  Assumes it will only be called when current-buffer
> is in `erc-mode'."
>  (let ((this-channel (or (erc-default-target)
>                          (buffer-name (current-buffer)))))
>    ...
>
> might have been broken for some reason.  To verify, someone would have
> to try with Emacs 22 and 23 inserting just before the `let' something
> like
>
> (unless (eq major-mode 'erc-mode)
>  (message "Not in erc-mode ..."))
>
> and look whether it triggers.
>
> martin
>
>
>
>
>
>

[-- Attachment #2: Type: text/html, Size: 3182 bytes --]

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[John Paul Wallington]

> Hi,

Hi Antoine,

> I did some testing, and the bug was not in tracker as I suspected, it's in
> scrolltobottom, in erc-goodies.el. This should be much easier for you to
> debug: it's only a few lines long. I strongly suspect the part following
> this comment is responsible :
>       ;; Temporarily bind resize-mini-windows to nil so that users who have
> it
>       ;; set to a non-nil value will not suffer from premature minibuffer
>       ;; shrinkage due to the below recenter call.  I have no idea why this
>       ;; works, but it solves the problem, and has no negative side effects.
>       ;; (Fran Litterio, 2003/01/07)
> 
> Looks like it does have side effects after all :-)

Sorry, I'm having difficulties reproducing the bug.

Could you try getting rid of the code that binds
`resize-mini-windows', perhaps using the function below, and test
whether your suspicion is correct?


(defun erc-scroll-to-bottom (window display-start)
  "Changed to see whether not let binding `resize-mini-windows' to nil
doesn't cause the bug."
  (if (window-live-p window)
      (erc-with-selected-window window
        (save-restriction
          (widen)
          (when (and erc-insert-marker
                     ;; we're editing a line. Scroll.
                     (> (point) erc-insert-marker))
            (save-excursion
              (goto-char (point-max))
              (recenter (or erc-input-line-position -1))
              (sit-for 0)))))))

bug#858: Conflict between ERC mode tracking and ibuffer window shrinking

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 2017 bytes --]

My suspicion was wrong, the binding doesn't change anything. I just said
that by looking at the comment, but didn't try. Sorry about that.
However, (defun erc-scroll-to-bottom (window display-start) nil) removes the
bug.
The bug should be reproducable with emacs -Q, the ibuffer configuration
lines I gave in the first bug report, and scroll-to-bottom mode on. I'm
using latest CVS, the bug also happens with a build from july.
Best of luck
Antoine
2008/9/5 John Paul Wallington <jpw@pobox.com>

> > Hi,
>
> Hi Antoine,
>
> > I did some testing, and the bug was not in tracker as I suspected, it's
> in
> > scrolltobottom, in erc-goodies.el. This should be much easier for you to
> > debug: it's only a few lines long. I strongly suspect the part following
> > this comment is responsible :
> >       ;; Temporarily bind resize-mini-windows to nil so that users who
> have
> > it
> >       ;; set to a non-nil value will not suffer from premature minibuffer
> >       ;; shrinkage due to the below recenter call.  I have no idea why
> this
> >       ;; works, but it solves the problem, and has no negative side
> effects.
> >       ;; (Fran Litterio, 2003/01/07)
> >
> > Looks like it does have side effects after all :-)
>
> Sorry, I'm having difficulties reproducing the bug.
>
> Could you try getting rid of the code that binds
> `resize-mini-windows', perhaps using the function below, and test
> whether your suspicion is correct?
>
>
> (defun erc-scroll-to-bottom (window display-start)
>  "Changed to see whether not let binding `resize-mini-windows' to nil
> doesn't cause the bug."
>  (if (window-live-p window)
>      (erc-with-selected-window window
>        (save-restriction
>          (widen)
>          (when (and erc-insert-marker
>                     ;; we're editing a line. Scroll.
>                     (> (point) erc-insert-marker))
>            (save-excursion
>              (goto-char (point-max))
>              (recenter (or erc-input-line-position -1))
>              (sit-for 0)))))))
>

[-- Attachment #2: Type: text/html, Size: 2912 bytes --]