Horton

Thread

Commits

  1. Silence uninitialized variable warning with some compiler versions

  2. pg_regress: Include diffs in TAP output

  1. Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    m.litsarev@postgrespro.ru — 2026-06-11T15:08:13Z

    Hi, hackers!

I found warning while compiling the code (introduced in 
5720ae01436480c1dcdd34ece0c3d83b00844bca)
with the folloing message

postgres/build/../src/test/regress/pg_regress.c: In function 
‘results_differ’:
postgres/build/../src/test/regress/pg_regress.c:1577:17: warning: 
‘startpos’ may be used uninitialized in this function 
[-Wmaybe-uninitialized]
  1577 |                 fseek(difffile, startpos, SEEK_SET);

gcc (Ubuntu 11.5.0-10ubuntu1~22~ppa1) 11.5.0
ldd (Ubuntu GLIBC 2.35-0ubuntu3.13) 2.35

./configure --enable-debug --enable-cassert --enable-depend 
--enable-tap-tests --enable-nls --with-perl --with-icu --with-libxml 
--with-libxslt --with-gssapi --with-openssl --with-zstd --with-lz4 
--with-ldap --with-python --prefix=$PGINSTDIR --with-pgport=$PGPORT

CFLAGS="-ggdb -Og -g3 -fno-omit-frame-pointer -fvisibility=default 
-Wno-clobbered"


It definitely seems false-positive from semantical point of view, but to 
keep compiler silent I attached a small patch which fixes it.

Respectfully,
Mikhail Litsarev, PostgresPro

  2. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    m.litsarev@postgrespro.ru — 2026-06-11T15:09:56Z

    The patch attached

  3. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    Ewan Young <kdbase.hack@gmail.com> — 2026-06-12T01:53:43Z

    Hi Mikhail,

On Thu, Jun 11, 2026 at 11:10 PM <m.litsarev@postgrespro.ru> wrote:
>
> The patch attached

Thanks for the report and patch.

I looked at the data flow and agree this is a false positive. startpos is
assigned at the ftell() call inside the first if (difffile) block, and the
only way to reach the use at fseek() with difffile != NULL is for that same
block to have executed (since difffile is reassigned by the fopen(..., "r")
at the end of it).

I should note I couldn't reproduce the warning here: with gcc 12.2 it stays
silent at every optimization level, including -Og.

The fix looks correct to me. The raw diff looks large, but it's almost entirely
re-indentation — git diff -w shows the actual change is just moving the
braces to nest the read block inside the first one.

For the record, the minimal alternative would be long startpos = 0; at the
declaration, which is a one-line diff. I'd lean toward your version, but
I'll leave the final call to a committer.

Regards,
Ewan

  4. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    Heikki Linnakangas <hlinnaka@iki.fi> — 2026-06-15T10:28:43Z

    On 12/06/2026 04:53, Ewan Young wrote:
> On Thu, Jun 11, 2026 at 11:10 PM <m.litsarev@postgrespro.ru> wrote:
>> The patch attached
> 
> Thanks for the report and patch.
> 
> I looked at the data flow and agree this is a false positive. startpos is
> assigned at the ftell() call inside the first if (difffile) block, and the
> only way to reach the use at fseek() with difffile != NULL is for that same
> block to have executed (since difffile is reassigned by the fopen(..., "r")
> at the end of it).
> 
> I should note I couldn't reproduce the warning here: with gcc 12.2 it stays
> silent at every optimization level, including -Og.
> 
> The fix looks correct to me. The raw diff looks large, but it's almost entirely
> re-indentation — git diff -w shows the actual change is just moving the
> braces to nest the read block inside the first one.
> 
> For the record, the minimal alternative would be long startpos = 0; at the
> declaration, which is a one-line diff. I'd lean toward your version, but
> I'll leave the final call to a committer.

LGTM as far as the patch goes.

The function feels too nonchalant about errors while opening the file. 
The diff output file was created earlier in open_result_files() already, 
so it really should exist. I propose the attached, to bail out if the 
fopen() fails. (As the original patch, this looks a lot smaller with 
"diff -w")

- Heikki

  5. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    m.litsarev@postgrespro.ru — 2026-06-16T15:22:44Z

    > ...I propose the attached, to bail out if the fopen() fails. (As the 
> original patch, this looks a lot smaller with "diff -w")
I agree the v2 patch is even better.

Mikhail

  6. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    Heikki Linnakangas <hlinnaka@iki.fi> — 2026-06-17T06:03:28Z

    On 16/06/2026 18:22, m.litsarev@postgrespro.ru wrote:
>> ...I propose the attached, to bail out if the fopen() fails. (As the 
>> original patch, this looks a lot smaller with "diff -w")
> I agree the v2 patch is even better.

Committed, thanks!

- Heikki

  7. Re: Fix warning: ‘startpos’ may be used uninitialized in function ‘results_differ’

    Imran Zaheer <imran.zhir@gmail.com> — 2026-06-18T06:26:56Z

    Hi

I also reported [1] some similar warnings emitted from the old gcc
some time ago. Do you think this should be worth getting fixed?


Thanks
Imran Zaheer


[1]: https://www.postgresql.org/message-id/flat/CA+UBfa=H9RQdZ8Tof+SF7gMC-FTrsJ_w=Vnxs4OiBAaB_go+tQ@mail.gmail.com<span
class="sewitjdq8nk9iqz"></span>


On Wed, Jun 17, 2026 at 11:03 AM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
>
> On 16/06/2026 18:22, m.litsarev@postgrespro.ru wrote:
> >> ...I propose the attached, to bail out if the fopen() fails. (As the
> >> original patch, this looks a lot smaller with "diff -w")
> > I agree the v2 patch is even better.
>
> Committed, thanks!
>
> - Heikki
>
>
>