GNU bug report logs - #9209
bidi slowdown

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 9209 in the body.
You can then email your comments to 9209 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: emacs user <user.emacs <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: bidi slowdown
Date: Sat, 30 Jul 2011 22:35:25 -0400

when using vm, I still see significant slowdowns in cursor motion in
the vm-summary buffer when bidi is on.  this is with a 3000+ lines
long buffer with the summary of the messages in the inbox.  please let
me know if there is any additional information I can provide.

2011-07-22  Eli Zaretskii

	Fix a significant slow-down of cursor motion with C-n, C-p,
	C-f/C-b, and C-v/M-v that couldn't keep up with keyboard
	auto-repeat under bidi redisplay in fontified buffers.

2011-07-28  Eli Zaretskii

	* buffer.c (init_buffer_once): Set bidi-display-reordering to t by
	default.

In GNU Emacs 24.0.50.1 (i386-apple-darwin11.0.0, NS apple-appkit-1138.00)
 of 2011-07-29 on dhcp-10-249-70-250.xx.edu
Windowing system distributor `Apple', version 10.3.1138
configured using `configure  '--with-ns' '--without-x''

Message #8 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: emacs user <user.emacs <at> gmail.com>
Cc: 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Sun, 31 Jul 2011 06:03:43 +0300

> Date: Sat, 30 Jul 2011 22:35:25 -0400
> From: emacs user <user.emacs <at> gmail.com>
> 
> when using vm, I still see significant slowdowns in cursor motion in
> the vm-summary buffer when bidi is on.  this is with a 3000+ lines
> long buffer with the summary of the messages in the inbox.  please let
> me know if there is any additional information I can provide.

Yes, I need an example of the buffer text.  Also, does it matter
whether point is near the beginning or the end of the buffer (or in
any other particular place)?

In general, the bidi display _is_ slower, because it does more, but I
hope to get it to be tolerable.

Thanks.

Message #11 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: emacs user <user.emacs <at> gmail.com>
Cc: 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Sun, 31 Jul 2011 02:22:58 -0400

Please keep 9209 <at> debbugs.gnu.org on the CC list, so that this
discussion gets recorded in the bug tracker.

> Date: Sun, 31 Jul 2011 00:00:26 -0400
> From: emacs user <user.emacs <at> gmail.com>
> 
> I checked and it's perfectly OK near the beginning of the buffer, and
> very slow near the end.

OK.  Next, please tell which cursor keys are slow to respond.  Are C-f
and C-b slow?  Are left and right arrow significantly slower than C-f
and C-b?  What about C-n/C-p resp up- and down-arrows?  Finally, what
about PageUp/PageDown and C-v/M-v?

Please also try setting the variable bidi-paragraph-direction to
left-to-right, and see if that solves the problem.

Thanks.

Message #14 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: emacs user <user.emacs <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Sun, 31 Jul 2011 05:50:05 -0400

it seems that c-p is slower than c-f, c-b and c-n, M-v and c-v which
are all also slow.  setting bidi-paragraph-direction to left-to-right
does seem to fix the problem.

On Sun, Jul 31, 2011 at 2:22 AM, Eli Zaretskii <eliz <at> gnu.org> wrote:
> Please keep 9209 <at> debbugs.gnu.org on the CC list, so that this
> discussion gets recorded in the bug tracker.
>
>> Date: Sun, 31 Jul 2011 00:00:26 -0400
>> From: emacs user <user.emacs <at> gmail.com>
>>
>> I checked and it's perfectly OK near the beginning of the buffer, and
>> very slow near the end.
>
> OK.  Next, please tell which cursor keys are slow to respond.  Are C-f
> and C-b slow?  Are left and right arrow significantly slower than C-f
> and C-b?  What about C-n/C-p resp up- and down-arrows?  Finally, what
> about PageUp/PageDown and C-v/M-v?
>
> Please also try setting the variable bidi-paragraph-direction to
> left-to-right, and see if that solves the problem.
>
> Thanks.
>

Message #17 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: emacs user <user.emacs <at> gmail.com>
Cc: 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Sun, 31 Jul 2011 06:04:08 -0400

> Date: Sun, 31 Jul 2011 05:50:05 -0400
> From: emacs user <user.emacs <at> gmail.com>
> Cc: 9209 <at> debbugs.gnu.org
> 
> it seems that c-p is slower than c-f, c-b and c-n, M-v and c-v which
> are all also slow.  setting bidi-paragraph-direction to left-to-right
> does seem to fix the problem.

Thanks.

What is the name of the vm summary buffer?  Is that name fixed when vm
is used?

I think bidi-paragraph-direction should be automatically set to
left-to-right.  The question (being discussed on emacs-devel) is how
to do that, given that vm is not bundled with Emacs.

Message #20 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: emacs user <user.emacs <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: "Uday S. Reddy" <u.s.reddy <at> cs.bham.ac.uk>, 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Sun, 31 Jul 2011 17:00:46 -0400

On Sun, Jul 31, 2011 at 6:04 AM, Eli Zaretskii <eliz <at> gnu.org> wrote:
>> Date: Sun, 31 Jul 2011 05:50:05 -0400
>> From: emacs user <user.emacs <at> gmail.com>
>> Cc: 9209 <at> debbugs.gnu.org
>>
>> it seems that c-p is slower than c-f, c-b and c-n, M-v and c-v which
>> are all also slow.  setting bidi-paragraph-direction to left-to-right
>> does seem to fix the problem.
>
> Thanks.
>
> What is the name of the vm summary buffer?  Is that name fixed when vm
> is used?

it is INBOX-summary where INBOX is the inbox file name, so it varies among users

>
> I think bidi-paragraph-direction should be automatically set to
> left-to-right.  The question (being discussed on emacs-devel) is how
> to do that, given that vm is not bundled with Emacs.

perhaps Uday can help...?

Message #23 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Uday Reddy <usr.vm.rocks <at> gmail.com>
To: emacs user <user.emacs <at> gmail.com>
Cc: "Uday S. Reddy" <u.s.reddy <at> cs.bham.ac.uk>, Eli Zaretskii <eliz <at> gnu.org>,
	9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Mon, 1 Aug 2011 22:01:34 +0100

emacs user writes:

> > What is the name of the vm summary buffer?  Is that name fixed when vm
> > is used?
> 
> it is INBOX-summary where INBOX is the inbox file name, so it varies among
> users 

The default name is "<folder-name> Summary"

> >
> > I think bidi-paragraph-direction should be automatically set to
> > left-to-right.  The question (being discussed on emacs-devel) is how
> > to do that, given that vm is not bundled with Emacs.
> 
> perhaps Uday can help...?
> 

There is of course the `vm-summary-mode-hook'.  You can set various things
there.

Cheers,
Uday

Message #26 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Uday Reddy <usr.vm.rocks <at> gmail.com>
Cc: u.s.reddy <at> cs.bham.ac.uk, user.emacs <at> gmail.com, 9209 <at> debbugs.gnu.org
Subject: Re: bug#9209: bidi slowdown
Date: Tue, 02 Aug 2011 01:42:38 -0400

> Date: Mon, 1 Aug 2011 22:01:34 +0100
> From: Uday Reddy <usr.vm.rocks <at> gmail.com>
> Cc: Eli Zaretskii <eliz <at> gnu.org>,
>     9209 <at> debbugs.gnu.org,
>     "Uday S. Reddy" <u.s.reddy <at> cs.bham.ac.uk>
> 
> > > I think bidi-paragraph-direction should be automatically set to
> > > left-to-right.  The question (being discussed on emacs-devel) is how
> > > to do that, given that vm is not bundled with Emacs.
> > 
> > perhaps Uday can help...?
> > 
> 
> There is of course the `vm-summary-mode-hook'.  You can set various things
> there.

Would it make sense for VM to do that automatically as part of
preparing the summary buffer, i.e. as part of vm-summary-mode?

Message #29 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: u.s.reddy <at> cs.bham.ac.uk, 9209 <at> debbugs.gnu.org, user.emacs <at> gmail.com,
	Uday Reddy <usr.vm.rocks <at> gmail.com>
Subject: Re: bug#9209: bidi slowdown
Date: Tue, 2 Aug 2011 19:37:55 +0100

Eli Zaretskii writes:

> Would it make sense for VM to do that automatically as part of
> preparing the summary buffer, i.e. as part of vm-summary-mode?

Sure.  At the moment, unibyte/multibyte settings are done when a buffer is
created.  We could do set other things similarly.

What needs to be done? set bidi-paragraph-direction to 'left-to-right?
Can't the effect be achieved by doing setq-default initially?

Cheers,
Uday

Message #32 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
Cc: 9209 <at> debbugs.gnu.org, user.emacs <at> gmail.com, usr.vm.rocks <at> gmail.com
Subject: Re: bug#9209: bidi slowdown
Date: Tue, 02 Aug 2011 22:30:56 +0300

> Date: Tue, 2 Aug 2011 19:37:55 +0100
> From: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
> Cc: Uday Reddy <usr.vm.rocks <at> gmail.com>,
>     user.emacs <at> gmail.com,
>     9209 <at> debbugs.gnu.org,
>     u.s.reddy <at> cs.bham.ac.uk
> 
> What needs to be done? set bidi-paragraph-direction to 'left-to-right?

Yes, in the summary buffer.

> Can't the effect be achieved by doing setq-default initially?

You mean, for all of Emacs?  It's possible we will do just that, but
even so VM should not rely on that, because some user could revert the
default setting.

Message #35 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>, 9209 <at> debbugs.gnu.org,
	user.emacs <at> gmail.com, usr.vm.rocks <at> gmail.com
Subject: Re: bug#9209: bidi slowdown
Date: Tue, 2 Aug 2011 21:10:56 +0100

Eli Zaretskii writes:

> > What needs to be done? set bidi-paragraph-direction to 'left-to-right?
> 
> Yes, in the summary buffer.

Ok, Eli, please try the revision 882 in the 8.2.x branch.

Cheers,
Uday

Message #38 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: Uday Reddy <usr.vm.rocks <at> gmail.com>
To: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 9209 <at> debbugs.gnu.org, user.emacs <at> gmail.com,
	usr.vm.rocks <at> gmail.com
Subject: Re: bug#9209: bidi slowdown
Date: Wed, 3 Aug 2011 00:25:36 +0100

Uday S Reddy writes:

> Ok, Eli, please try the revision 882 in the 8.2.x branch.

Oops, I meant "emacs user" above.

Uday

Message #41 received at 9209 <at> debbugs.gnu.org (full text, mbox):

From: emacs user <user.emacs <at> gmail.com>
To: Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 9209 <at> debbugs.gnu.org, usr.vm.rocks <at> gmail.com
Subject: Re: bug#9209: bidi slowdown
Date: Mon, 8 Aug 2011 18:58:57 -0600

sorry for the delay.  when using the version of vm suggested by Uday,
the slowness is gone and things work fine (thanks Uday!).  if I do
describe variable in the summary buffer I get:

bidi-paragraph-direction is a variable defined in `C source code'.
Its value is left-to-right
Local in buffer RMAIL Summary; global value is nil

BTW, bidi in emacs is an absolutely wonderful addition, thanks for
doing this...!  cheers, E


On Tue, Aug 2, 2011 at 2:10 PM, Uday S Reddy <u.s.reddy <at> cs.bham.ac.uk> wrote:
> Eli Zaretskii writes:
>
>> > What needs to be done? set bidi-paragraph-direction to 'left-to-right?
>>
>> Yes, in the summary buffer.
>
> Ok, Eli, please try the revision 882 in the 8.2.x branch.
>
> Cheers,
> Uday
>

Message #46 received at 9209-done <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: emacs user <user.emacs <at> gmail.com>
Cc: u.s.reddy <at> cs.bham.ac.uk, 9209-done <at> debbugs.gnu.org, usr.vm.rocks <at> gmail.com
Subject: Re: bug#9209: bidi slowdown
Date: Tue, 09 Aug 2011 05:56:12 +0300

> Date: Mon, 8 Aug 2011 18:58:57 -0600
> From: emacs user <user.emacs <at> gmail.com>
> Cc: Eli Zaretskii <eliz <at> gnu.org>, usr.vm.rocks <at> gmail.com, 9209 <at> debbugs.gnu.org
> 
> sorry for the delay.  when using the version of vm suggested by Uday,
> the slowness is gone and things work fine (thanks Uday!).  if I do
> describe variable in the summary buffer I get:
> 
> bidi-paragraph-direction is a variable defined in `C source code'.
> Its value is left-to-right
> Local in buffer RMAIL Summary; global value is nil

Thanks.  So I'm hereby closing this bug report.

> BTW, bidi in emacs is an absolutely wonderful addition, thanks for
> doing this...!  cheers, E

Thank you.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.