Re: can we use different function in place of atoi in vacuumdb.c file - Mailing list pgsql-hackers

From Surafel Temesgen
Subject Re: can we use different function in place of atoi in vacuumdb.c file
Date
Msg-id CALAY4q_-Vq1H=L3s__OhNFhQ5+xCPzK6D7HgrwWWnwTAPsXsSw@mail.gmail.com
Whole thread Raw
In response to can we use different function in place of atoi in vacuumdb.c file  (Mahendra Singh Thalor <mahi6run@gmail.com>)
List pgsql-hackers

Hi,
On Thu, Jan 23, 2020 at 3:56 PM Mahendra Singh Thalor <mahi6run@gmail.com> wrote:
Hi all,
While reviewing one patch, I found that if we give any non-integer string to atoi (say aa), then it is returning zero(0) as output so we are not giving any error(assuming 0 as valid argument) and continuing our operations.

Ex:
Let say, we gave "-P aa" (patch is in review[1]), then it will disable parallel vacuum because atoi is returning zero as parallel degree but ideally it should give error or at least it should not disable parallel vacuum.

I think, in-place of atoi, we should use different function ( defGetInt32, strtoint) or we can write own function.

Thoughts?

[1]: https://www.postgresql.org/message-id/CA%2Bfd4k6DgwtQSr4%3DUeY%2BWbGuF7-oD%3Dm-ypHPy%2BsYHiXZc%2BhTUQ%40mail.gmail.com
-- 


For server side there are also scanint8 function for string parsing and for Clint side we can use strtoint function that didn't have the above issue and it accept char pointer which is I think suitable for [1] usage.

regards

Surafel

 

pgsql-hackers by date:

Previous
From: Takashi Menjo
Date:
Subject: RE: [PoC] Non-volatile WAL buffer
Next
From: Michael Paquier
Date:
Subject: Re: [PATCH] /src/backend/access/transam/xlog.c, tiny improvements