Re: Proposal for Signal Detection Refactoring - Mailing list pgsql-hackers

On 01/10/2018 14:00, Chris Travers wrote:
>
>
> On Wed, Sep 26, 2018 at 9:54 AM Chris Travers <chris.travers@adjust.com
> <mailto:chris.travers@adjust.com>> wrote:
>
>
>
>     On Tue, Sep 25, 2018 at 3:23 PM Tom Lane <tgl@sss.pgh.pa.us
>     <mailto:tgl@sss.pgh.pa.us>> wrote:
>
>         Chris Travers <chris.travers@adjust.com
>         <mailto:chris.travers@adjust.com>> writes:
>         > However,  what I think one could do is use a struct of volatile
>         > sig_atomic_t members and macros for checking/setting.  Simply
>         writing a
>         > value is safe in C89 and higher.
>
>         Yeah, we could group those flags in a struct, but what's the point? 

>
>
>     This was one of two things I noticed in my previous patch on
>     interrupts and loops where I wasn't sure what the best practice in
>     our code is.
>
>     If we don't want to make this change, then would there be any
>     objection to me writing up a README describing the flags, and best
>     practices in terms of checking them in our code based on the current
>     places we use them?  If the current approach will be unlikely to
>     change in the future, then at least we can document that the way I
>     went about things is consistent with current best practices so next
>     time someone doesn't really wonder.
>
>
> Attaching a first draft of a readme.  Feedback welcome.  I noticed
> further that we used to document signals and what they did with carious
> processes but that this was removed after 7.0, probably due to
> complexity reasons.

diff --git a/src/backend/utils/init/README b/src/backend/utils/init/README
new file mode 100644
index 0000000000..0472687f53
--- /dev/null
+++ b/src/backend/utils/init/README
@@ -0,0 +1,96 @@
+src/backend/utils/misc/README

Not only are the directory names mismatched here, but I wonder whether
this file would be long in either of these directories.

+Implementation Notes on Globals and Signal/Event Handling
+=========================================================
+
+The approch to signal handling in PostgreSQL is designed to strictly
conform
+with the C89 standard and designed to run with as few platform
assumptions as
+possible.

We use C99 now, so why would be design this to be conforming to C89?

More generally, I'd like this material to be code comments.  It's the
kind of stuff that gets outdated before long if it's kept separate.

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