apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@lyra.org>
Subject Re: documented 1.0 showstoppers
Date Fri, 04 Jun 2004 19:52:09 GMT
On Fri, Jun 04, 2004 at 12:43:22PM -0700, Stas Bekman wrote:
>...
> You have a problem here, 1.0.1 will have to be backwards compatible with 
> 1.0.0. So if now you have API problems which went so far unnoticed (or 
> ignored) you can't commit to them in 1.0.0.

You're assuming that a problem in the API exists. But you also talk about
"coverage" which is generally about the implementation.

Implementation bugs are easy to fix in a 1.0.1. API changes would require
a 1.1 release. Which we can also do IF we find a problem.

> Sure, here is a perfect solution for that problem. Introduce a new doxygen 
> marker. And start marking those solid, used and tested APIs as ready for 
> 1.0. Release 1.0  which will have both kinds. The ones that are untested 
> should be marked as 'subject to change', and then there is no requirement 
> to have them tested.

That seems a bit silly. What makes you believe that the library is so
buggy that we have to mark areas as "steer clear". People have been using
this thing for years.

As far as I'm concerned, there appear to be a couple dodgy places, but
that doesn't mean that we have to go and revamp the entire docset with new
markers.

> >If you want to increase coverage testing, then go ahead. By all means. But
> >I'm with David: what exists on June 16 is going to be 1.0.0. If you don't
> >like what that will be, then apply your efforts to fix it.
> 
> My efforts to fix it go totally ignored so far.

What are you talking about? You're a committer on APR. If you want to
increase the coverage of the test suite, then start writing tests and
committing them. I don't see what is stopping you.

Cheers,
-g

-- 
Greg Stein, http://www.lyra.org/

Mime
View raw message