On Tue, Jun 19, 2012 at 4:31 AM, Peter Eisentraut <pete...@gmx.net> wrote:
> On ons, 2012-01-18 at 21:21 +0200, Peter Eisentraut wrote:
>> On lör, 2012-01-07 at 16:41 -0500, Tom Lane wrote:
>> > Peter Eisentraut <pete...@gmx.net> writes:
>> > > I suggest that we change PostgresMain(), PostmasterMain(), BackendRun(),
>> > > WalSenderMain(), and WalSndLoop() to return void as well.
>> >
>> > I agree this code is not very consistent or useful, but one question:
>> > what should the callers do if one of these functions *does* return?
>>
>> I was thinking of a two-pronged approach:  First, add
>> __attribute__((noreturn)) to the functions.  This will cause a suitable
>> compiler to verify on a source-code level that nothing actually returns
>> from the function.  And second, at the call site, put an abort();  /*
>> not reached */.  Together, this will make the code cleaner and more
>> consistent, and will also help the compiler out a bit about the control
>> flow.
>
> Patch for 9.3 attached.

+1. Should this call around line 4114 of postmaster.c not bother with
proc_exit() anymore:

        /* And run the backend */
        proc_exit(BackendRun(&port));

I was hoping that some of the clang static analyzer complaints would
go away with these changes, though it looks like only one[1] did. I
would be interested to see the similar elog/ereport patch you
mentioned previously, perhaps that will eliminate a bunch of warnings.

Josh

[1] this one goes away with the patch:
http://kupershmidt.org/pg/scan-build-2012-06-19-master/report-E2cUqJ.html#EndPath

-- 
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers

Reply via email to