[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: bash man page and FreeBSD mandoc
From: |
David Ongaro |
Subject: |
Re: bash man page and FreeBSD mandoc |
Date: |
Tue, 30 Apr 2024 17:07:28 -0700 |
User-agent: |
Gnus/5.13 (Gnus v5.13) |
Chet Ramey <chet.ramey@case.edu> writes:
> On 4/29/24 12:43 PM, David Ongaro wrote:
>> Chet Ramey writes:
>>
>>> On 12/22/18 12:49 AM, Derek Schrock wrote:
>>>> On Fri, Dec 21, 2018 at 09:59:16AM EST, Chet Ramey wrote:
>>>>> This is a macro redefinition fix for an old bug in the BSD and Ultrix man
>>>>> macros. We can try removing it, since there probably aren't any remaining
>>>>> systems using them.
>>>>
>>>> Ok, I see that whole block is related to the comment. So the fix here
>>>> would be to remove lines 14 to 38 not just removing the characters?
>>>
>>> Yes. The fixed version of that macro still lives on in the Solaris 10
>>> man macros, but the broken ones are probably all gone.
>> As of now I can still see these control characters included in the
>> man
>> page of a current bash version (GNU bash, version 5.2.26(1)-release
>> (aarch64-apple-darwin23.2.0)).
>
> They're commented out, and troff will ignore them. It looks like mandoc
> has a problem with characters that should be ignored because they
> appear in a commented line. I'd say that's a bug in mandoc.
I don't know if it's a bug in mandoc or intended behavior. The end
result is that for endusers on BSD based systems the bash manpage
appears to potentially be broken.
But as you mention yourself above, this macro redefinition shouldn't
even be relevant anymore, even for these old BSD and Utrix macros. And
even if, their Users have to be aware of the bug and explicitly comment
out the relevant code in the manpage. If we worry about that we could
just replace the comment with a link to an explanation and the relevant
macro redefinitions.
> I suppose we can work around it.
And no matter how we spin it, it doesn't make sense to keep around a
workaround for a non-bash bug which probably doesn't even affect anyone
anymore because we argue we don't have to do a "workaround" for another
non-bash bug which as of now affects millions.
And since the "workaround" is just a cleanup I wouldn't even call it as
such.