Re: pg_upgrade check fails on Solaris 10

Hello, hackers!

We got an error for pg_upgrade check on the branch REL_11_STABLE (commit
40ad4202513c72f5c1beeb03e26dfbc8890770c0) on Solaris 10 because IIUC the
argument to the sed command is not enclosed in quotation marks (see
[1]):

$ gmake -C src/bin/pg_upgrade/ check
<...>
MAKE=gmake
bindir="/home/buildfarm/mpolyakova/postgrespro_REL_11_STABLE/inst/bin"
libdir="/home/buildfarm/mpolyakova/postgrespro_REL_11_STABLE/inst/lib"
EXTRA_REGRESS_OPTS="" /bin/sh test.sh --install
test.sh: MSYS/MINGW/: not found
gmake: *** [check] Error 1
gmake: Leaving directory
`/home/buildfarm/mpolyakova/postgrespro_REL_11_STABLE/src/bin/pg_upgrade'
$ sed: command garbled: s/

Attached diff.patch fixes the problem.

About the system: SunOS, Release 5.10, KernelID Generic_141444-09.
About the used shell: according to the manual, it comes from the package
SUNWcsu.

Thanks to Victor Wagner for his help to investigate this issue.

[1] $ man sh
<...>
Quoting
The following characters have a special meaning to the shell and cause
termination of a word unless quoted:
; & ( ) | ^ < > newline space tab

--
Marina Polyakova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

On 2019-Sep-17, Marina Polyakova wrote:

> Hello, hackers!
>
> We got an error for pg_upgrade check on the branch REL_11_STABLE (commit
> 40ad4202513c72f5c1beeb03e26dfbc8890770c0) on Solaris 10 because IIUC the
> argument to the sed command is not enclosed in quotation marks (see [1]):

Hmm, I'm surprised it has taken this long to detect the problem.

> Attached diff.patch fixes the problem.

I have pushed it to all branches that have src/bin/pg_upgrade (namely,
9.5 onwards), thanks. I hope this won't make the msys/mingw machines
angry ;-)

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On 2019-09-18 17:36, Alvaro Herrera wrote:
> On 2019-Sep-17, Marina Polyakova wrote:
>
>> Hello, hackers!
>>
>> We got an error for pg_upgrade check on the branch REL_11_STABLE
>> (commit
>> 40ad4202513c72f5c1beeb03e26dfbc8890770c0) on Solaris 10 because IIUC
>> the
>> argument to the sed command is not enclosed in quotation marks (see
>> [1]):
>
> Hmm, I'm surprised it has taken this long to detect the problem.

Looking at the members of buildfarm [1] castoroides and protosciurus -
IIUC they do not check pg_upgrade. And I was that lucky one who have run
the branch with the latest commits at our buildfarm...

>> Attached diff.patch fixes the problem.
>
> I have pushed it to all branches that have src/bin/pg_upgrade (namely,
> 9.5 onwards), thanks. I hope this won't make the msys/mingw machines
> angry ;-)

Thank you! I ran pg_upgrade tests for MSYS, everything is fine.

The branch REL9_4_STABLE (commit
8a17afe84be6fefe76d0d2f4d26c5ee075e64487) has the same issue - according
to the release table [2] it is still supported, isn't it?...

[1] https://buildfarm.postgresql.org/cgi-bin/show_members.pl
[2] https://www.postgresql.org/support/versioning/

--
Marina Polyakova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

On 2019-Sep-23, Marina Polyakova wrote:

> On 2019-09-18 17:36, Alvaro Herrera wrote:
> > On 2019-Sep-17, Marina Polyakova wrote:
> >

> > > We got an error for pg_upgrade check on the branch REL_11_STABLE
> > > (commit
> > > 40ad4202513c72f5c1beeb03e26dfbc8890770c0) on Solaris 10 because IIUC
> > > the
> > > argument to the sed command is not enclosed in quotation marks (see
> > > [1]):
> >
> > Hmm, I'm surprised it has taken this long to detect the problem.
>
> Looking at the members of buildfarm [1] castoroides and protosciurus - IIUC
> they do not check pg_upgrade. And I was that lucky one who have run the
> branch with the latest commits at our buildfarm...

Argh.

But I meant "how come nobody runs pg_upgrade tests on old Solaris?"

> > I have pushed it to all branches that have src/bin/pg_upgrade (namely,
> > 9.5 onwards), thanks. I hope this won't make the msys/mingw machines
> > angry ;-)
>
> Thank you! I ran pg_upgrade tests for MSYS, everything is fine.
>
> The branch REL9_4_STABLE (commit 8a17afe84be6fefe76d0d2f4d26c5ee075e64487)
> has the same issue - according to the release table [2] it is still
> supported, isn't it?...

Yeah, but pg_upgrade is in contrib/ in 9.4, so nowhere as good as from
9.5 onwards; and it's going to die in a couple of months anyway, so I'm
not thrilled about fixing this there.

If you *need* to have this fixed in 9.4, we can do that, but do you?

--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On 2019-09-23 19:41, Alvaro Herrera wrote:
> On 2019-Sep-23, Marina Polyakova wrote:
>> The branch REL9_4_STABLE (commit
>> 8a17afe84be6fefe76d0d2f4d26c5ee075e64487)
>> has the same issue - according to the release table [2] it is still
>> supported, isn't it?...
>
> Yeah, but pg_upgrade is in contrib/ in 9.4, so nowhere as good as from
> 9.5 onwards; and it's going to die in a couple of months anyway, so I'm
> not thrilled about fixing this there.
>
> If you *need* to have this fixed in 9.4, we can do that, but do you?

No, we don't. I just noticed :-)

--
Marina Polyakova
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company