GNU bug report logs - #77421
[PATCH] [etags] Fix module definition with ::

Previous Next

To reply to this bug, email your comments to 77421 AT debbugs.gnu.org.
There is no need to reopen the bug first.

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

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

From: Laurent Stacul <laurent.stacul <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: [PATCH] [etags] Fix module definition with ::
Date: Mon, 31 Mar 2025 20:45:12 +0200

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

Tags: patch

Tags: patch

Hello dear Emacs maintainers,

I have a patch I would like to submit. I tried to follow all the docs
but if I did something wrong, don't hesitate to tell me.

Problem: In ruby we can define a nested module/class the safe way (in
the sense, if the parent module does not exist, it will define it):

  module M
    module N
    end
  end

Or, if M already exists, we can also write:

  module M::N; end

With the later notation, the tag generated by etags will be M::N. When
browsing the code, using xref-find-definitions when the point is on N,
will not be able to find the definition of N because the implicit tag
name is M::N.

This is the same problem with nested classes or even some rare
allowed definitions like explicitely defining a module/class from the
global namespace:

  class ::A; end

Solution: We need to give an explicit tag name. To achieve this, on
module/class definition we truncate the name to the last found column.

Best Regards,
Laurent

In GNU Emacs 31.0.50 (build 1, x86_64-pc-linux-gnu, GTK+ Version
 3.24.43, cairo version 1.18.2) of 2025-03-29 built on Host-002
Repository revision: 04bd6497300789cd90b365299885517d92292648
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12101016
System Description: Fedora Linux 41 (Workstation Edition)

Configured using:
 'configure --prefix=/home/stac/.local --disable-gc-mark-trace
 --with-cairo --with-dbus --with-gif --with-gpm=no --with-gnutls
 --with-harfbuzz --with-imagemagick --with-jpeg --with-mailutils
 --with-modules --with-native-compilation=aot --with-png --with-rsvg
 --with-sqlite3 --with-tiff --with-tree-sitter --with-webp
 --with-wide-int --with-x-toolkit=gtk3 --with-xinput2 --with-xpm
 --without-pop 'CFLAGS=-O2 -pipe -mtune=native -march=native
 -fomit-frame-pointer''

[0001-etags-Fix-module-definition-with.patch (text/patch, attachment)]

Message #8 received at 77421 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Laurent Stacul <laurent.stacul <at> gmail.com>, Dmitry Gutov <dmitry <at> gutov.dev>
Cc: 77421 <at> debbugs.gnu.org
Subject: Re: bug#77421: [PATCH] [etags] Fix module definition with ::
Date: Tue, 01 Apr 2025 14:53:30 +0300

> From: Laurent Stacul <laurent.stacul <at> gmail.com>
> Date: Mon, 31 Mar 2025 20:45:12 +0200
> 
> I have a patch I would like to submit. I tried to follow all the docs
> but if I did something wrong, don't hesitate to tell me.
> 
> Problem: In ruby we can define a nested module/class the safe way (in
> the sense, if the parent module does not exist, it will define it):
> 
>   module M
>     module N
>     end
>   end
> 
> Or, if M already exists, we can also write:
> 
>   module M::N; end
> 
> With the later notation, the tag generated by etags will be M::N. When
> browsing the code, using xref-find-definitions when the point is on N,
> will not be able to find the definition of N because the implicit tag
> name is M::N.
> 
> This is the same problem with nested classes or even some rare
> allowed definitions like explicitely defining a module/class from the
> global namespace:
> 
>   class ::A; end
> 
> Solution: We need to give an explicit tag name. To achieve this, on
> module/class definition we truncate the name to the last found column.

Thanks.

Dmitry, any comments?

Message #11 received at 77421 <at> debbugs.gnu.org (full text, mbox):

From: Dmitry Gutov <dmitry <at> gutov.dev>
To: Eli Zaretskii <eliz <at> gnu.org>, Laurent Stacul <laurent.stacul <at> gmail.com>
Cc: 77421 <at> debbugs.gnu.org
Subject: Re: bug#77421: [PATCH] [etags] Fix module definition with ::
Date: Wed, 2 Apr 2025 00:52:18 +0300

Hi!

On 01/04/2025 14:53, Eli Zaretskii wrote:
>> Or, if M already exists, we can also write:
>>
>>    module M::N; end
>>
>> With the later notation, the tag generated by etags will be M::N. When
>> browsing the code, using xref-find-definitions when the point is on N,
>> will not be able to find the definition of N because the implicit tag
>> name is M::N.
>>
>> This is the same problem with nested classes or even some rare
>> allowed definitions like explicitely defining a module/class from the
>> global namespace:
>>
>>    class ::A; end
>>
>> Solution: We need to give an explicit tag name. To achieve this, on
>> module/class definition we truncate the name to the last found column.
> Thanks.
> 
> Dmitry, any comments?

Yes, this all seems right.

This might have not come up before because this convention is relatively 
less popular, but the proposed change should be an improvement.

Message #14 received at 77421 <at> debbugs.gnu.org (full text, mbox):

From: Laurent Stacul <laurent.stacul <at> gmail.com>
To: Dmitry Gutov <dmitry <at> gutov.dev>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 77421 <at> debbugs.gnu.org,
 Laurent Stacul <laurent.stacul <at> gmail.com>
Subject: Re: bug#77421: [PATCH] [etags] Fix module definition with ::
Date: Wed, 02 Apr 2025 15:55:09 +0200

Hello,

Thanks for your review.

> This might have not come up before because this convention is
> relatively less popular, but the proposed change should be an
> improvement.

I agree this is not popular and I would not recommend this form of
definition but some codebases use it.

I just wanted to raise awareness on the changes on the two files
CTAGS.good_crlf and CTAGS.good_update which look huge (I was expecting 3
lines changed as in the other files). What I did is described in my
update of the README but I am not sure I did the right thing.

Regards,
Laurent

Message #19 received at 77421-done <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Dmitry Gutov <dmitry <at> gutov.dev>
Cc: 77421-done <at> debbugs.gnu.org, laurent.stacul <at> gmail.com
Subject: Re: bug#77421: [PATCH] [etags] Fix module definition with ::
Date: Thu, 03 Apr 2025 09:13:25 +0300

> Date: Wed, 2 Apr 2025 00:52:18 +0300
> Cc: 77421 <at> debbugs.gnu.org
> From: Dmitry Gutov <dmitry <at> gutov.dev>
> 
> Hi!
> 
> On 01/04/2025 14:53, Eli Zaretskii wrote:
> >> Or, if M already exists, we can also write:
> >>
> >>    module M::N; end
> >>
> >> With the later notation, the tag generated by etags will be M::N. When
> >> browsing the code, using xref-find-definitions when the point is on N,
> >> will not be able to find the definition of N because the implicit tag
> >> name is M::N.
> >>
> >> This is the same problem with nested classes or even some rare
> >> allowed definitions like explicitely defining a module/class from the
> >> global namespace:
> >>
> >>    class ::A; end
> >>
> >> Solution: We need to give an explicit tag name. To achieve this, on
> >> module/class definition we truncate the name to the last found column.
> > Thanks.
> > 
> > Dmitry, any comments?
> 
> Yes, this all seems right.
> 
> This might have not come up before because this convention is relatively 
> less popular, but the proposed change should be an improvement.

Thanks, now installed on master, and closing the bug.

Previous Next

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