Re: Change atoi to strtol in same place
От | Michael Paquier |
---|---|
Тема | Re: Change atoi to strtol in same place |
Дата | |
Msg-id | 20190910053553.GB11737@paquier.xyz обсуждение исходный текст |
Ответ на | Re: Change atoi to strtol in same place (Joe Nelson <joe@begriffs.com>) |
Ответы |
Re: Change atoi to strtol in same place
Re: Change atoi to strtol in same place |
Список | pgsql-hackers |
On Mon, Sep 09, 2019 at 11:02:49PM -0500, Joe Nelson wrote: > Alvaro Herrera from 2ndQuadrant wrote: > > Okay, so who is submitting a new version here? Surafel, Joe? > > I've attached a revision that uses pg_strtoint64 from str2int-10.patch > (posted in <alpine.DEB.2.21.1909032007230.21690@lancre>). My patch is > based off that one and c5bc7050af. > > It covers the same front-end utilities as atoi-to-strtol-v2.patch. Other > utilities in the pg codebase still have atoi inside, but I thought I'd > share my progress to see if people like the direction. If so, I can > update the rest of the utils. > > I added a helper function to a new file in src/fe_utils, but had to > modify Makefiles in ways that might not be too clean. Maybe there's a > better place for the function. Using a wrapper in src/fe_utils makes sense. I would have used options.c for the new file, or would that be too much generic? The code indentation is weird, particularly the switch/case in pg_strtoint64_range(). The error handling is awkward. I think that you should just call pg_log_error in pg_strtoint64_range instead of returning an error string as you do. You could do that by passing down the option name to the routine, and generate a new set of error messages using that. Why do you need to update common/Makefile? The builds on Windows are broken. You are adding one file into fe_utils, but forgot to update @pgfeutilsfiles in Mkvcbuild.pm. -- Michael
Вложения
В списке pgsql-hackers по дате отправления: