Re: Make --help output fit within 80 columns per line - Mailing list pgsql-hackers

From torikoshia
Subject Re: Make --help output fit within 80 columns per line
Date
Msg-id eba6645a0e228958d8541f213d66da0f@oss.nttdata.com
Whole thread Raw
In response to Re: Make --help output fit within 80 columns per line  (Peter Eisentraut <peter@eisentraut.org>)
Responses Re: Make --help output fit within 80 columns per line
Re: Make --help output fit within 80 columns per line
List pgsql-hackers
On Tue, Sep 19, 2023 at 3:23 AM Greg Sabino Mullane <htamfids@gmail.com> 
wrote:
Thanks for your investigation!

> On Fri, Sep 15, 2023 at 11:11 AM torikoshia 
> <torikoshia@oss.nttdata.com> wrote:
>> I do not intend to adhere to this rule(my terminals are usually bigger
>> than 80 chars per line), but wouldn't it be a not bad direction to use
>> 80 characters for all commands?
> 
> Well, that's the question du jour, isn't it? The 80 character limit is 
> based on punch cards, and really has no place in modern systems. While 
> gnu systems are stuck in the past, many other ones have moved on to 
> more sensible defaults:
> 
> $ wget --help | wc -L
> 110
> 
> $ gcloud --help | wc -L
> 122
> 
> $ yum --help | wc -L
> 122
> 
> git is an interesting one, as they force things through a pager for 
> their help, but if you look at their raw help text files, they have 
> plenty of times they go past 80 when needed:
> 
> $ wc -L git/Documentation/git-*.txt | sort -g | tail -20
>     109 git-filter-branch.txt
>     109 git-rebase.txt
>     116 git-diff-index.txt
>     116 git-http-fetch.txt
>     117 git-restore.txt
>     122 git-checkout.txt
>     122 git-ls-tree.txt
>     129 git-init-db.txt
>     131 git-push.txt
>     132 git-update-ref.txt
>     142 git-maintenance.txt
>     144 git-interpret-trailers.txt
>     146 git-cat-file.txt
>     148 git-repack.txt
>     161 git-config.txt
>     162 git-notes.txt
>     205 git-stash.txt
>     251 git-submodule.txt
> 
> So in summary, I think 80 is a decent soft limit, but let's not stress 
> out about some lines going over that, and make a hard limit of perhaps 
> 120.

+1. It may be a good compromise.
For enforcing the hard limit, is it better to add a regression test like 
patch 0001?


On 2023-09-21 16:45, Peter Eisentraut wrote:
> On 31.08.23 09:47, torikoshia wrote:
>> BTW, psql --help outputs the content of PGHOST, which caused a failure 
>> in the test:
>> 
>> ```
>> -h, --host=HOSTNAME      database server host or socket directory
>>                           (default: 
>> "/var/folders/m7/9snkd5b54cx_b4lxkl9ljlcc0000gn/T/LobrmSUf7t")
>> ```
>> 
>> It may be overkill, added a logic for removing the content of PGHOST.
> 
> I wonder, should we remove this?  We display the
> environment-variable-based defaults in psql --help, but not for any
> other programs.  This is potentially misleading.

Agreed. It seems inconsistent with other commands.
Patch 0002 removed environment-variable-based defaults in psql --help.

-- 
Regards,

--
Atsushi Torikoshi
NTT DATA Group Corporation
Attachment

pgsql-hackers by date:

Previous
From: Kuwamura Masaki
Date:
Subject: Clarify where the severity level is defined
Next
From: Yugo NAGATA
Date:
Subject: Doc: vcregress .bat commands list lacks "taptest"