Re: build remaining Flex files standalone - Mailing list pgsql-hackers

From John Naylor
Subject Re: build remaining Flex files standalone
Date
Msg-id CAFBsxsGq=jLMofR3BN1S1xJcaJsYcQnkWFRtRDObYVtDST0+Vg@mail.gmail.com
Whole thread Raw
In response to build remaining Flex files standalone  (John Naylor <john.naylor@enterprisedb.com>)
Responses Re: build remaining Flex files standalone
List pgsql-hackers
Here are the rest. Most of it was pretty straightforward, with the
main exception of jsonpath_scan.c, which is not quite finished. That
one passes tests but still has one compiler warning. I'm unsure how
much of what is there already is really necessary or was cargo-culted
from elsewhere without explanation. For starters, I'm not sure why the
grammar has a forward declaration of "union YYSTYPE". It's noteworthy
that it used to compile standalone, but with a bit more stuff, and
that was reverted in 550b9d26f80fa30. I can hack on it some more later
but I ran out of steam today.

Other questions thus far:

- "BISONFLAGS += -d" is now in every make file with a .y file -- can
we just force that everywhere?

- Include order seems to matter for the grammar's .h file. I didn't
test if that was the case every time, and after a few miscompiles just
always made it the last inclusion, but I'm wondering if we should keep
those inclusions outside %top{} and put it at the start of the next
%{} ?

- contrib/cubeparse.y now has a global variable -- not terrific, but I
wanted to get something working first.

- I'm actually okay with guc-file.c now, but I'll still welcome
comments on that.

-- 
John Naylor
EDB: http://www.enterprisedb.com

Attachment

pgsql-hackers by date:

Previous
From: Amit Kapila
Date:
Subject: Re: [PATCH] Use indexes on the subscriber when REPLICA IDENTITY is full on the publisher
Next
From: Bharath Rupireddy
Date:
Subject: Re: Patch proposal: New hooks in the connection path