Re: run pgindent on a regular basis / scripted manner
От | Andrew Dunstan |
---|---|
Тема | Re: run pgindent on a regular basis / scripted manner |
Дата | |
Msg-id | 5b9dede4-3c70-b8c1-c53e-26cdf4125b35@dunslane.net обсуждение исходный текст |
Ответ на | Re: run pgindent on a regular basis / scripted manner (Alvaro Herrera <alvherre@alvh.no-ip.org>) |
Список | pgsql-hackers |
On 2023-04-28 Fr 05:25, Alvaro Herrera wrote:
On 2023-Feb-05, Andrew Dunstan wrote:So here's a diff made from running perltidy v20221112 with the additional setting --valign-exclusion-list=", = => || && if unless"I ran this experimentally with perltidy 20230309, and compared that with the --novalign behavior (not to propose the latter -- just to be aware of what else is vertical alignment doing.)
Thanks for looking.
Based on the differences between both, I think we'll definitely want to include =~ and |= in this list, and I think we should discuss whether to also include "or" (for "do_stuff or die()" type of constructs) and "qw" (mainly used in 'use Foo qw(one two)' import lists). All these have effects (albeit smaller than the list you gave) on our existing code.
I'm good with all of these I think
If you change from an exclusion list to --novalign then you lose alignment of trailing # comments, which personally I find a loss, even though they're still a multi-line effect. Another change would be that it ditches alignment of "{" but that only changes msvc/Install.pm, so I think we shouldn't worry; and then there's this one: -use PostgreSQL::Test::Utils (); +use PostgreSQL::Test::Utils (); use PostgreSQL::Test::BackgroundPsql (); which I think we could just change to qw() if we cared enough (but I bet we don't).
Yeah, me too.
All in all, I think sticking to --valign-exclusion-list=", = => =~ |= || && if or qw unless" is a good deal.
wfm
cheers
andrew
-- Andrew Dunstan EDB: https://www.enterprisedb.com
В списке pgsql-hackers по дате отправления: