GNU bug report logs - #39763
Incorrect behaviour of Farsi (farsi-isiri-9147) input method

Previous Next

To reply to this bug, email your comments to 39763 AT debbugs.gnu.org.

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

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

From: Ali AlipourR <alipoor90 <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Incorrect behaviour of Farsi (farsi-isiri-9147) input method
Date: Mon, 24 Feb 2020 12:35:04 +0330

[Message part 1 (text/plain, inline)]

Hi,

Farsi keyboard (isiri 9147) has a second "shift" level, which is usually
implemented to be accessible when holding Alt key.
for example when in farsi input method, pressing the "d" key should insert:
U+06CC = ی
but pressing "Shift + d" should insert:
U+064A = ي
and pressing "Alt + d" should insert:
U+0649 = ى
(trust me it's different from the first one! even though they look similar,
it makes a difference when searching for text etc since the unicode code
points are different (and they are also read differently ... one is
pronounced like i the other one like a, though this is not related to the
discussion))

The problem is, in Emacs, when you set the input method to
farsi-isiri-9147, when you press M-d, it deletes a word, which is expected,
but even when you press C-q M-d, instead of inserting U+0649 = ى, it
inserts Latin character U+00E4 = ä)

and I used the d key and letter U+0649 as an example here, there are more
keys and letters like this. in general, when in farsi-isiri-9147, when you
press C-q M-letter instead of inserting appropriate "second shift level"
Farsi letters according to isiri-9147 layout, Emacs inserts Latin
characters with diacritics according to English keyboard layout.

this was tested under GNOME Shell 3.34.3 (with gnome/X input method set to
EN; since setting it to Farsi, means gnome/X will send Farsi chars to Emacs
and thus no Emacs keybinding (e.g. C-f or C-n or C-d or C-x C-s or ...)
will work at all)
(other programs including gedit, firefox and bash in gnome-terminal, behave
correctly, when gnome input method is set to Farsi; even in Emacs Alt + d
works correctly and inserts U+0649 = ى, regardless of what input method is
selected in Emacs itself, if gnome input method is set to Farsi, but again,
as mentioned, this way no Emacs keybinding will work ... since this way,
gnome is directly sending farsi chars to Emacs)

I will attach isiri-9147 keyboard layout standard and some of its tables
(the column marked in red, are the unicode characters that should be
inserted, when pressing the letters on the rightmost column while holding
Alt key.

Thanks

[Message part 2 (text/html, inline)]

[9147.pdf (application/pdf, attachment)]

[1.png (image/png, attachment)]

[2.png (image/png, attachment)]

[3.png (image/png, attachment)]

Message #8 received at 39763 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Schwab <schwab <at> suse.de>
To: Ali AlipourR <alipoor90 <at> gmail.com>
Cc: 39763 <at> debbugs.gnu.org
Subject: Re: bug#39763: Incorrect behaviour of Farsi (farsi-isiri-9147) input
 method
Date: Mon, 24 Feb 2020 10:56:53 +0100

On Feb 24 2020, Ali AlipourR wrote:

> The problem is, in Emacs, when you set the input method to
> farsi-isiri-9147, when you press M-d, it deletes a word, which is expected,
> but even when you press C-q M-d, instead of inserting U+0649 = ى, it
> inserts Latin character U+00E4 = ä)

What happens if you use Alt-d instead of Meta-d?

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab <at> suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Message #11 received at 39763 <at> debbugs.gnu.org (full text, mbox):

From: Ali AlipourR <alipoor90 <at> gmail.com>
To: Andreas Schwab <schwab <at> suse.de>
Cc: 39763 <at> debbugs.gnu.org
Subject: Re: bug#39763: Incorrect behaviour of Farsi (farsi-isiri-9147) input
 method
Date: Mon, 24 Feb 2020 16:34:16 +0330

[Message part 1 (text/plain, inline)]

by M-d I meant Alt-d as my keyboard doesn't have a meta key

On Mon, Feb 24, 2020 at 1:26 PM Andreas Schwab <schwab <at> suse.de> wrote:

> On Feb 24 2020, Ali AlipourR wrote:
>
> > The problem is, in Emacs, when you set the input method to
> > farsi-isiri-9147, when you press M-d, it deletes a word, which is
> expected,
> > but even when you press C-q M-d, instead of inserting U+0649 = ى, it
> > inserts Latin character U+00E4 = ä)
>
> What happens if you use Alt-d instead of Meta-d?
>
> Andreas.
>
> --
> Andreas Schwab, SUSE Labs, schwab <at> suse.de
> GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
> "And now for something completely different."
>

[Message part 2 (text/html, inline)]

Message #14 received at 39763 <at> debbugs.gnu.org (full text, mbox):

From: Ali AlipourR <alipoor90 <at> gmail.com>
To: Andreas Schwab <schwab <at> suse.de>
Cc: 39763 <at> debbugs.gnu.org
Subject: Re: bug#39763: Incorrect behaviour of Farsi (farsi-isiri-9147) input
 method
Date: Mon, 24 Feb 2020 16:57:12 +0330

[Message part 1 (text/plain, inline)]

to illustrate the problem better:

1- under gnome (other desktop environments should work too, if they
implement standard Farsi/Persian keyboard layout), from settings > Region &
Language > add "Persian" input source

2- switch to "Persian" input source

3- open firefox, go to  https://unicode-table.com and click on the search
bar, then press: Alt+d and then Enter to see the code point for the typed
character

4- set the gnome input source back to English

5- open Emacs, run M-x set-input-method RET farsi-isiri-9147 RET

6- press C-q Alt-d and compare the typed character (which is a latin
character with diacritics) with what you've got in firefox (which was a
Farsi/Arabic Character)

again, i used Alt-d as an example here, you can see a complete list in
attached images in the original email. (as another example try Alt-h)

[Message part 2 (text/html, inline)]

Message #17 received at 39763 <at> debbugs.gnu.org (full text, mbox):

From: Andreas Schwab <schwab <at> suse.de>
To: Ali AlipourR <alipoor90 <at> gmail.com>
Cc: 39763 <at> debbugs.gnu.org
Subject: Re: bug#39763: Incorrect behaviour of Farsi (farsi-isiri-9147) input
 method
Date: Mon, 24 Feb 2020 14:27:26 +0100

On Feb 24 2020, Ali AlipourR wrote:

> by M-d I meant Alt-d as my keyboard doesn't have a meta key

But it must have a key that emits ISO_Level3_Shift.

Andreas.

-- 
Andreas Schwab, SUSE Labs, schwab <at> suse.de
GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE  1748 E4D4 88E3 0EEA B9D7
"And now for something completely different."

Message #20 received at 39763 <at> debbugs.gnu.org (full text, mbox):

From: Ali AlipourR <alipoor90 <at> gmail.com>
To: Andreas Schwab <schwab <at> suse.de>
Cc: 39763 <at> debbugs.gnu.org
Subject: Re: bug#39763: Incorrect behaviour of Farsi (farsi-isiri-9147) input
 method
Date: Mon, 24 Feb 2020 17:05:01 +0330

[Message part 1 (text/plain, inline)]

$xmodmap -pke | grep Alt
keycode 108 = Alt_R Meta_R ISO_Level3_Shift

If that's what you mean

[Message part 2 (text/html, inline)]

Previous Next

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