obsolete functions found during the build process

obsolete functions found during the build process

[Jean-Christophe Helary]

There are plenty of warnings during the build process about obsolete functions "as of 2n.n".

For ex:

In nnmh-request-list-1:
gnus/nnmh.el:245:14:Warning: ‘string-to-multibyte’ is an obsolete function (as
    of 26.1); use ‘decode-coding-string’.

Is there anything to be done about that ?

Jean-Christophe

Re: obsolete functions found during the build process

[Eli Zaretskii]

> From: Jean-Christophe Helary <jean.christophe.helary@gmail.com>
> Date: Sat, 12 Aug 2017 09:42:33 +0200
> 
> In nnmh-request-list-1:
> gnus/nnmh.el:245:14:Warning: ‘string-to-multibyte’ is an obsolete function (as
>     of 26.1); use ‘decode-coding-string’.
> 
> Is there anything to be done about that ?

In general, replace the obsolete functions as the warning suggests.
In some cases, the use of the obsolete function is justified, though.

Re: obsolete functions found during the build process

[Jean-Christophe Helary]

Thank you Eli.

Jean-Christophe 

> On Aug 12, 2017, at 10:45, Eli Zaretskii <eliz@gnu.org> wrote:
> 
>> From: Jean-Christophe Helary <jean.christophe.helary@gmail.com>
>> Date: Sat, 12 Aug 2017 09:42:33 +0200
>> 
>> In nnmh-request-list-1:
>> gnus/nnmh.el:245:14:Warning: ‘string-to-multibyte’ is an obsolete function (as
>>    of 26.1); use ‘decode-coding-string’.
>> 
>> Is there anything to be done about that ?
> 
> In general, replace the obsolete functions as the warning suggests.
> In some cases, the use of the obsolete function is justified, though.

Re: obsolete functions found during the build process

[Paul Eggert]

Eli Zaretskii wrote:
> In some cases, the use of the obsolete function is justified, though.


Is there some way for code to say "I know this function is obsolete, but the 
call is justified so don't warn about it"? That would be helpful, so that 
compiles could be clean and so that people would be more likely to pay attention 
to any stray warnings that do crop up.

Re: obsolete functions found during the build process

[Eric Abrahamsen]

Paul Eggert <eggert@cs.ucla.edu> writes:

> Eli Zaretskii wrote:
>> In some cases, the use of the obsolete function is justified, though.
>
>
> Is there some way for code to say "I know this function is obsolete,
> but the call is justified so don't warn about it"? That would be
> helpful, so that compiles could be clean and so that people would be
> more likely to pay attention to any stray warnings that do crop up.

Maybe the `with-no-warnings' macro?

Eric

Re: obsolete functions found during the build process

[Stefan Monnier]

> In general, replace the obsolete functions as the warning suggests.
> In some cases, the use of the obsolete function is justified, though.

AFAICT, there are indeed cases where using decode-coding-string instead
of string-to-multibyte is not really an option, but that means
"obsolete" is not really a good description.

Using with-no-warnings to mask obsolescence warnings is OK when we know
that the code is only kept for backward compatibility, but when there is
no good replacement, it's not a good solution.

So we should strive to provide valid alternatives for every use of
string-to-multibyte (or de-obsolete it, tho it's not my favorite solution).


        Stefan