About xmllint checking for the validity of postgres.xml in 9.5

On Wed, Dec 3, 2014 at 12:09 PM, Michael Paquier
<michael.paquier@gmail.com> wrote:

Since commit 5d93ce2d, the output of xmllint is checked by passing
--valid to it. Isn't that a regression with what we were doing for
pre-9.4 versions? For example, with 9.4 and older versions it is
possible to compile man pages even if the xml spec is not entirely
valid when using docbook 4.2.

Another thing coming to my mind is why don't we simply have a variable
to pass flags to xmllint similarly to xsltproc? Packagers would be
then free to pass the arguments they want. (Note that in some of the
environments where I build the docs postgres.xml is found as invalid,
making build fail for master only, not for older branches).

In any case, attached is a patch showing the idea, bringing more
flexibility in the build, default value being "--valid --noout" if the
flag is not passed by the caller.
Regards,
--
Michael

Michael Paquier wrote:

Hi all,

Since commit 5d93ce2d, the output of xmllint is checked by passing
--valid to it. Isn't that a regression with what we were doing for
pre-9.4 versions? For example, with 9.4 and older versions it is
possible to compile man pages even if the xml spec is not entirely
valid when using docbook 4.2.

I don't think this is a regression. It just means we're stricter than
before. Is there a reason behind this tinkering?

--
�lvaro Herrera http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

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

On Wed, Dec 3, 2014 at 9:43 PM, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

Michael Paquier wrote:

Hi all,

Since commit 5d93ce2d, the output of xmllint is checked by passing
--valid to it. Isn't that a regression with what we were doing for
pre-9.4 versions? For example, with 9.4 and older versions it is
possible to compile man pages even if the xml spec is not entirely
valid when using docbook 4.2.

I don't think this is a regression. It just means we're stricter than
before. Is there a reason behind this tinkering?

Just got surprised by how we got strict on master when doing a build
of the docs using docbook 4.2 and some old versions of docbook-dsssl
and docbook-xsl: man pages can still compile even if the spec is not
exactly correct sometimes.
--
Michael

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