Re: getting rid of maintainer-check

The maintainer-check target never really caught on, I think. Most
people don't run it, and that in turn annoys those who do. Also, it
doesn't provide much functionality.

I propose that we get rid of it and roll the functionality into the
regular build.

Specifically:

- Running duplicate_oids during the regular build was already discussed
elsewhere recently. There are some details to be resolved there, but
it's doable.

- Checking for tabs in SGML files can be run during the regular
documentation build without problems.

- The NLS checks can also be run during the regular NLS-enabled build.

That's it. Any concerns?

On Tue, Sep 3, 2013 at 10:41 PM, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
> The maintainer-check target never really caught on, I think. Most
> people don't run it, and that in turn annoys those who do. Also, it
> doesn't provide much functionality.
>
> I propose that we get rid of it and roll the functionality into the
> regular build.
>
> Specifically:
>
> - Running duplicate_oids during the regular build was already discussed
> elsewhere recently. There are some details to be resolved there, but
> it's doable.
>
> - Checking for tabs in SGML files can be run during the regular
> documentation build without problems.
>
> - The NLS checks can also be run during the regular NLS-enabled build.
>
> That's it. Any concerns?

I can't speak for anyone else, but personally I think that sounds like
a significant improvement.

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

On 2013-09-03 22:41:17 -0400, Peter Eisentraut wrote:
> The maintainer-check target never really caught on, I think. Most
> people don't run it, and that in turn annoys those who do. Also, it
> doesn't provide much functionality.
>
> I propose that we get rid of it and roll the functionality into the
> regular build.
>
> Specifically:
>
> - Running duplicate_oids during the regular build was already discussed
> elsewhere recently. There are some details to be resolved there, but
> it's doable.

Maybe we should also badger cpluspluscheck into a state where it can be
run as part of a normal build if a c++ compiler was detected?

I think it misses vpath support and it might be dependant on some
bashims.

Greetings,

Andres Freund

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

On 9/4/13 11:12 AM, Andres Freund wrote:
> Maybe we should also badger cpluspluscheck into a state where it can be
> run as part of a normal build if a c++ compiler was detected?
>
> I think it misses vpath support and it might be dependant on some
> bashims.

That might also be doable. If we could at the same time stick a usable
C++ compiler configuration into PGXS, that would also help the growing
number of extensions that need that and are currently using variously
bad workarounds.

On 03-09-2013 23:41, Peter Eisentraut wrote:
> The maintainer-check target never really caught on, I think. Most
> people don't run it, and that in turn annoys those who do. Also, it
> doesn't provide much functionality.
>
It has its use (before each release) but I agree that it isn't used
during minor version updates (because you need to update only one or two
po files).

> I propose that we get rid of it and roll the functionality into the
> regular build.
>
By 'regular build' you mean --enable-nls? If so, +1.

> - Running duplicate_oids during the regular build was already discussed
> elsewhere recently. There are some details to be resolved there, but
> it's doable.
>
This has been bashing sufficient developers along the years. +1.

> - Checking for tabs in SGML files can be run during the regular
> documentation build without problems.
>
This one too. +1.

--
Euler Taveira Timbira - http://www.timbira.com.br/
PostgreSQL: Consultoria, Desenvolvimento, Suporte 24x7 e Treinamento

On Tue, 2013-09-03 at 22:41 -0400, Peter Eisentraut wrote:
> The maintainer-check target never really caught on, I think. Most
> people don't run it, and that in turn annoys those who do. Also, it
> doesn't provide much functionality.
>
> I propose that we get rid of it and roll the functionality into the
> regular build.

Here is a patch for that. I also integrated Andrew's Perl version of
duplicate_oids.

The MSVC build needs to be updated separately, if they want to have that
same functionality.

On Thu, Sep 12, 2013 at 11:21 PM, Peter Eisentraut <peter_e(at)gmx(dot)net> wrote:
> On Tue, 2013-09-03 at 22:41 -0400, Peter Eisentraut wrote:
>> The maintainer-check target never really caught on, I think. Most
>> people don't run it, and that in turn annoys those who do. Also, it
>> doesn't provide much functionality.
>>
>> I propose that we get rid of it and roll the functionality into the
>> regular build.
>
> Here is a patch for that. I also integrated Andrew's Perl version of
> duplicate_oids.
>
> The MSVC build needs to be updated separately, if they want to have that
> same functionality.

These patches look OK to me, and everyone who has commented has been
in favor of this proposal. I'll mark this Ready for Committer.

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