Oleg Bartunov <oleg(at)sai(dot)msu(dot)su> writes:
> On Tue, 9 Oct 2007, Andrew Dunstan wrote:
>> Are we in beta or not? To me, beta means nothing but bug fixes go in,
>> period. No ifs, no buts, no maybes, no exceptions. And that should
>> definitely go for contrib as well. We need a bit of self-discipline around
> the original intention was to have text search companion with code examples
> for dictionaries and parser API and ability to keep them in workable
> condition, instead of having them in SGML documentation.
Yes. I think that we can justify new contrib code here as a documentation
bug fix: right now, the examples in sections 12.9 and 12.10 are wrong
(obsolete), impossible to maintain (which is why they're obsolete;
they failed to track code changes), and not in a format that's directly
useful as a template for new code.
However, the dict_regex code is not ready for a number of reasons,
quite aside from the question of whether we want to add a dependency.
I concur with the idea of dropping it for now.
I'm a bit inclined to drop dict_roman as well, as it seems of
questionable real use, and code-example-wise it adds nothing over
dict_intdict. That would leave us with two dictionary examples and one
parser example, and I think both of the dictionaries look useful enough
to be worth keeping. (In particular I'd not vote for having only
dict_intdict since it does not illustrate how to use a config file
... and we surely don't want to encourage people to not do that
regards, tom lane
In response to
pgsql-patches by date
|Next:||From: Kevin Grittner||Date: 2007-10-10 15:12:31|
|Subject: Re: quote_literal with NULL|
|Previous:||From: Magnus Hagander||Date: 2007-10-10 11:50:30|
|Subject: Re: Additional windows codepages|