Re: taking stdbool.h into use

On 8/16/17 02:32, Thomas Munro wrote:
>> 0001-Fix-bool-int-type-confusion.patch
>
> Looks good.

Committed that one.

>> 0002-Change-TRUE-FALSE-to-true-false.patch
>
> Looks good. What about these?

OK, did more digging. Windows, ICU, and Perl define their own
TRUE/FALSE, so in those uses we don't have to make any changes. ECPG
also has its own definitions. In ECPG I also found some obsolete uses
for bools that have been fixed in the equivalent backend code a long
time ago, so I added patches to fix that. And another use in
src/backend/port/dynloader/darwin.c was probably wrong to begin with as
well.

I also went through all the comments to make equivalent changes. I kept
that in a separate patch for easier viewing, but it should probably be
committed together.

Using a case-insensitive diff mechanism one can verify that no logic was
changed here.

>> 0003-Remove-TRUE-and-FALSE.patch
>
> Looks good. What about these?
>
> src/interfaces/ecpg/include/ecpglib.h:#define FALSE 0
> src/interfaces/ecpg/pgtypeslib/extern.h:#define FALSE 0

Not sure about that. These are available for use by ecpg applications,
and it's perhaps not worth disrupting that.

>> 0004-Remove-BoolPtr-type.patch
>
> Looks good.

Committed that one.

>> 0007-Avoid-use-of-bool-in-thread_test.c.patch
>
> Looks good.

and that one

>> 0005-Make-casting-between-bool-and-GinTernaryValue-more-r.patch
>> 0008-Use-stdbool.h-if-available.patch

These need some more work based on Tom's feedback.

Attached is a new patch set. Based on the discussion so far, 0001
through 0007 might be ready; the other two need some more work.

--
Peter Eisentraut http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services