Re: Lexer issues - Mailing list pgsql-hackers

From Patrick REED
Subject Re: Lexer issues
Date
Msg-id CAJHg7sdaU3qfJVkYpS-8+Nnd63OpdnAVzSLm5cxNidVvvLZWwQ@mail.gmail.com
Whole thread Raw
In response to Re: Lexer issues  (Julien Rouhaud <rjuju123@gmail.com>)
Responses Re: Lexer issues  (Julien Rouhaud <rjuju123@gmail.com>)
List pgsql-hackers
Hi Julien,

Sorry for the late reply. I was able to solve the issue. It had to do with the extra syntax I had introduced in gram.y. However, since you mentioned the utility command, can you elaborate a bit more on that?

Thanks,
Patrick

On Tue, Apr 14, 2020 at 5:00 AM Julien Rouhaud <rjuju123@gmail.com> wrote:
Hello,

On Mon, Apr 13, 2020 at 4:04 PM Patrick REED <patrickreed352@gmail.com> wrote:
>
> I am experimenting with postgres and am wondering if there is any tutorial on how to properly add a new command to postgres.
>
> I want to add a new constraint on "CREATE ROLE" that requires an integer, it has an identifier that is not a known (reserved or unreserved keyword) in postgres, say we call it TestPatrick. In other words, I want to do this "CREATE ROLE X TestPatrick=10". I am having an issue with having postgres recognize my new syntax.
>
> I have seen this video: https://www.youtube.com/watch?v=uSEXTcEiXGQ and was able to add have my postgres compile with my added word (modified gram.y, kwlist.h, gram.cpp etc based on the video). However, when I use my syntax on a client session, it still doesn't recognize my syntax... Are there any specific lexer changes I need to make? I followed the example of CONNECTION LIMIT and tried to mimic it for Create ROLE.

I'd think that if you can get a successful compilation with a modified
gram.y (and any kwlist change needed) the new syntax should be
accepted (at least up to the parser, whether the utility command is
properly handled is another thing), since there's a single version of
the CreateRoleStmt.  Is there any chance that you're somehow
connecting to something else than the freshly make-install-ed binary,
or that the error is coming from later stage than parsing?

pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: [PATHC] Fix minor memory leak in pg_basebackup
Next
From: Michael Paquier
Date:
Subject: Re: Making openssl_tls_init_hook OpenSSL specific