Обсуждение: compiler warnings in copy.c

My compiler is unhappy with the latest changes to copy.c:

gcc -Wall -Wmissing-prototypes -Wpointer-arith
-Wdeclaration-after-statement -Wendif-labels
-Wmissing-format-attribute -Wformat-security -fno-strict-aliasing
-fwrapv -fexcess-precision=standard -g -O2 -Wall -Werror
-I../../../src/include -D_GNU_SOURCE -I/usr/include/libxml2   -c -o
copy.o copy.c -MMD -MP -MF .deps/copy.Po
copy.c: In function ‘DoCopy’:
copy.c:924:30: error: ‘rte’ may be used uninitialized in this function
[-Werror=uninitialized]
copy.c:793:17: note: ‘rte’ was declared here
cc1: all warnings being treated as errors

From what I can see, this is a pretty legitimate complaint.  If
stmt->relation == NULL, then rte never gets initialized, but we still
do cstate->range_table = list_make1(rte).  That can't be good.

Also, you seem to have pushed these commits with a date more than two
weeks in the past.  Please don't do that!

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

* Robert Haas (robertmhaas@gmail.com) wrote:
> My compiler is unhappy with the latest changes to copy.c:
>
> gcc -Wall -Wmissing-prototypes -Wpointer-arith
> -Wdeclaration-after-statement -Wendif-labels
> -Wmissing-format-attribute -Wformat-security -fno-strict-aliasing
> -fwrapv -fexcess-precision=standard -g -O2 -Wall -Werror
> -I../../../src/include -D_GNU_SOURCE -I/usr/include/libxml2   -c -o
> copy.o copy.c -MMD -MP -MF .deps/copy.Po
> copy.c: In function ‘DoCopy’:
> copy.c:924:30: error: ‘rte’ may be used uninitialized in this function
> [-Werror=uninitialized]
> copy.c:793:17: note: ‘rte’ was declared here
> cc1: all warnings being treated as errors

Huh, interesting that mine didn't.

> >From what I can see, this is a pretty legitimate complaint.  If
> stmt->relation == NULL, then rte never gets initialized, but we still
> do cstate->range_table = list_make1(rte).  That can't be good.

Yeah, I'll fix that.

> Also, you seem to have pushed these commits with a date more than two
> weeks in the past.  Please don't do that!

Oh, wow, sorry about that.  I had expected a rebase to update the date.
Thanks!
    Stephen

On 2015-01-28 15:05:11 -0500, Stephen Frost wrote:
> > Also, you seem to have pushed these commits with a date more than two
> > weeks in the past.  Please don't do that!
> 
> Oh, wow, sorry about that.  I had expected a rebase to update the date.

It updates the committer, but not the author date. Use --pretty=fuller
to see all the details. You can pass rebase --ignore-date to also reset
the author date. Or commit --amend --reset

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services

* Andres Freund (andres@2ndquadrant.com) wrote:
> On 2015-01-28 15:05:11 -0500, Stephen Frost wrote:
> > > Also, you seem to have pushed these commits with a date more than two
> > > weeks in the past.  Please don't do that!
> >
> > Oh, wow, sorry about that.  I had expected a rebase to update the date.
>
> It updates the committer, but not the author date. Use --pretty=fuller
> to see all the details. You can pass rebase --ignore-date to also reset
> the author date. Or commit --amend --reset

Thanks for the clarification as to what was happening.  I've modified my
aliases to use 'git am --ignore-date' which appears to have fixed this.
Thanks!
    Stephen