BUG #19450: Where is checksum_block.inc.c after master install?

On Sun, Apr 5, 2026 at 8:40 PM PG Bug reporting form
<noreply@postgresql.org> wrote:

The following bug has been logged on the website:

Bug reference: 19450
Logged by: RekGRpth
Email address: rekgrpth@gmail.com
PostgreSQL version: 18.3
Operating system: docker alpine
Description:

There is no checksum_block.inc.c after master install

I'll look into that. What broke, so I can reproduce?

BTW, the bug reporting form is meant for released versions, and this
does not affect 18.3.

--
John Naylor
Amazon Web Services

John Naylor <johncnaylorls@gmail.com> writes:

There is no checksum_block.inc.c after master install

I'll look into that. What broke, so I can reproduce?

src/include/Makefile knows what it's supposed to install out
of that subtree, and it thinks storage/*.h is sufficient.
I didn't check to see if the meson system has the same oversight.

One could argue that the real bug is having put a .c file into
the include/ tree in the first place. Why was it done like that?
Couldn't it be a .h file?

regards, tom lane

On Mon, Apr 6, 2026 at 9:50 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:

src/include/Makefile knows what it's supposed to install out
of that subtree, and it thinks storage/*.h is sufficient.
I didn't check to see if the meson system has the same oversight.

One could argue that the real bug is having put a .c file into
the include/ tree in the first place. Why was it done like that?
Couldn't it be a .h file?

That was the way it was first coded. I thought of this way to avoid
adding an exception to headerscheck. I can reverse that decision
easily, but I may not get to it today.

--
John Naylor
Amazon Web Services

John Naylor <johncnaylorls@gmail.com> writes:

On Mon, Apr 6, 2026 at 9:50 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:

One could argue that the real bug is having put a .c file into
the include/ tree in the first place. Why was it done like that?
Couldn't it be a .h file?

That was the way it was first coded. I thought of this way to avoid
adding an exception to headerscheck. I can reverse that decision
easily, but I may not get to it today.

Ah, the good ol' law of conservation of cruft. But on the whole
I think naming it .h not .c is less crufty. Agreed that there's
no great urgency about changing it.

regards, tom lane