GNU bug report logs - #77444
Allow install package from git repo when autocrlf is true

Previous Next

To reply to this bug, email your comments to 77444 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: "Yue Yi" <include_yy <at> qq.com>
To: "bug-gnu-emacs" <bug-gnu-emacs <at> gnu.org>
Subject: Allow install package from git repo when autocrlf is true
Date: Wed, 2 Apr 2025 09:17:16 +0800

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

Hello Emacs, As is well known, when compiling Emacs on Windows, it is best to set autocrlf to false before cloning the code to avoid issues with the autogen.sh script failing to run properly: git config --global core.autocrlf false Similarly, when autocrlf is set to true, installing a package via `package-vc-install' will result in the following error: Debugger entered--Lisp error: (error "Invalid version syntax: ¡®0.2\15¡¯")   error("Invalid version syntax: `%s'" "0.2\15")   version-to-list("0.2\15")   package-strip-rcs-id("0.2\15")   package-vc--version(#s(package-desc ...))   package-vc--generate-description-file(#s(package-desc ...))   package-vc--unpack-1(#s(package-desc ...))   package-vc--unpack(#s(package-desc ...))   package-vc-install("https://github.com/someone/some-package") In the error message above, \15 corresponds to decimal 13, which is the ASCII CR (displayed as ^M in Emacs). A simple inspection reveals that `lm-header', which is called within `package-vc--version', does not remove the trailing CR character. To address this, one possible solution is to modify the regular expression in lm-header to match and remove the CR character: diff --git a/lisp/emacs-lisp/lisp-mnt.el b/lisp/emacs-lisp/lisp-mnt.el index 111d512ef59..9cac77eb524 100644 --- a/lisp/emacs-lisp/lisp-mnt.el +++ b/lisp/emacs-lisp/lisp-mnt.el @@ -261,7 +261,7 @@ lm-header  		(if (save-excursion  		      (skip-chars-backward "^$" (match-beginning 0))  		      (= (point) (match-beginning 0))) -		    "[^\n]+" "[^$\n]+"))) +		    "[^\r\n]+" "[^$\r\n]+")))        (match-string-no-properties 0))))    (defun lm-header-multiline (header) However, at this point, the ^M characters in the buffer will affect the docstring, leading to numerous errors during byte compilation. Warning: docstring contains control char #x0d (position xxx) For this issue, is it a better choice to set autocrlf to input or false rather than modifying the code in package-vc?

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

Message #8 received at 77444 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: "Yue Yi" <include_yy <at> qq.com>
Cc: 77444 <at> debbugs.gnu.org
Subject: Re: bug#77444: Allow install package from git repo when autocrlf is
 true
Date: Wed, 02 Apr 2025 15:12:09 +0300

> Date: Wed, 2 Apr 2025 09:17:16 +0800
> From:  "Yue Yi" via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs <at> gnu.org>
> 
> diff --git a/lisp/emacs-lisp/lisp-mnt.el b/lisp/emacs-lisp/lisp-mnt.el
> index 111d512ef59..9cac77eb524 100644
> --- a/lisp/emacs-lisp/lisp-mnt.el
> +++ b/lisp/emacs-lisp/lisp-mnt.el
> @@ -261,7 +261,7 @@ lm-header
>  		(if (save-excursion
>  		      (skip-chars-backward "^$" (match-beginning 0))
>  		      (= (point) (match-beginning 0)))
> -		    "[^\n]+" "[^$\n]+")))
> +		    "[^\r\n]+" "[^$\r\n]+")))
>        (match-string-no-properties 0))))
>  
>  (defun lm-header-multiline (header)
> 
> However, at this point, the ^M characters in the buffer will affect the
> docstring, leading to numerous errors during byte compilation.
> 
> Warning: docstring contains control char #x0d (position xxx)

Also, any lone ^M characters will be taken as EOL.  So only a *M
before a newline should be considered EOL.

> For this issue, is it a better choice to set autocrlf to input or false
> rather than modifying the code in package-vc?

Personally, I think using Git with autocrlf set to anything but false
is asking for trouble.  EOL conversions is a Git feature one should
never use if one wants to stay sane.  So my recommendation would be to
force autocrlf = false.

Message #11 received at 77444 <at> debbugs.gnu.org (full text, mbox):

From: "Yue Yi" <include_yy <at> qq.com>
To: "Eli Zaretskii" <eliz <at> gnu.org>
Cc: 77444 <77444 <at> debbugs.gnu.org>
Subject: Re: bug#77444: Allow install package from git repo when autocrlf is
 true
Date: Thu, 3 Apr 2025 00:04:48 +0800

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

> Personally, I think using Git with autocrlf set to anything but false > is asking for trouble.  EOL conversions is a Git feature one should > never use if one wants to stay sane.  So my recommendation would be to > force autocrlf = false. Thanks, I think that makes sense. autocrlf should always be false.

[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.