Re: PostgreSQL vs SQL/XML Standards

On 10/25/18 11:08 AM, Alvaro Herrera wrote:

> XQilla seems to depend on Xerces, and seems to have died in 2011.

¿Eh? The latest release, 2.3.4 [1], is dated 2018-07-03.

It looks like the latest development has been happening on the
xquilla_2_3 branch. Sometimes project "activity" statistics rely
exclusively on the "master" branch ("xqilla" branch in this case),
and are deceptive if the project isn't being developed exclusively
by coding on master and backpatching to others.

I've noticed I'm facing the same thing in PL/Java ... plenty of
development lately, but on the REL1_5_STABLE branch. GitHub's
project statistics (and also Open Hub's) are just looking at master
and saying the project's been dead for two years. Now that 1.5.1 is
released, as soon as I get some of REL1_5_STABLE merged /up/ into
master, the statistics will probably magically show it's been alive
all along.

> Zorba appears to have been taken propietary, from the looks of its last
> commits.

It does seem harder to see what's going on there, but the commits
with "copyright changed" as the message turn out to be changing
only the copyright holder in the Apache 2.0 license from
"The FLWOR Foundation" to "zorba.io". But Matthias Brantner
participated with interest in the 2010 thread here where Zorba
was brought up before[2], so he may know something.

> Integrating a C runtime of a Java library sounds nightmarish --
> I wouldn't even think about that.

Or whether or not nightmarish, certainly duplicative of something
we can kinda already do.

In a way, some of the pressure is off, because if you need
a true XMLQUERY or XMLTABLE, you can get them with the Saxon-in-PL/Java
implementation, and right now in any supported PG version. You just
have to spell them funny, doing without the sugary syntax built into
the parser. They're missing some of the automatic casts from
the standard at the moment, which isn't really a loss of function,
as explicit casts can be added to any query needing them ... a temporary
annoyance until the rest of that example's in place.

But a roadmap that could lead to eventual availability of one of the
C/C++ implementations would be nice too.

-Chap

[1]: https://sourceforge.net/projects/xqilla/files/
[2]:
https://www.postgresql.org/message-id/7DDDB18E-041F-4238-B91D-3277EB1CE5BC%4028msec.com