Re: ANNOUNCE: smokers@perl.org Discussion of perl's daily build and smoke test
Jarkko Hietaniemi <[EMAIL PROTECTED]> wrote: > On Mon, Feb 19, 2001 at 04:01:25PM +0100, H.Merijn Brand wrote: > > On Mon, 19 Feb 2001 08:49:04 -0600, Jarkko Hietaniemi <[EMAIL PROTECTED]> wrote: > > > On Mon, Feb 19, 2001 at 03:47:12PM +0100, Johan Vromans wrote: > > > > As an active non-smoker, I'd appreciate a different name. > > > > > > Likewise. What's wrong with builders? > > > > Same here. Testers? > > perl-builders? Not funny. p
Re: ANNOUNCE: smokers@perl.org Discussion of perl's daily build and smoke test
"H.Merijn Brand" <[EMAIL PROTECTED]> wrote: > On Mon, 19 Feb 2001 08:49:04 -0600, Jarkko Hietaniemi <[EMAIL PROTECTED]> wrote: > > On Mon, Feb 19, 2001 at 03:47:12PM +0100, Johan Vromans wrote: > > > As an active non-smoker, I'd appreciate a different name. > > > > Likewise. What's wrong with builders? Because "PerlBuilder" is a commercial product. It's a poor quality competitor that I wouldn't recommend to anybody. If you put the words together like this, you'll create an accidental association. p
Re: Proposed basic criteria for accepting new core modules
Michael G Schwern <[EMAIL PROTECTED]> wrote:
> I'd like to propose two simple critereon for all future module
> additions to the core:
>
> 1) It has a reasonable amount of POD documentation. At minimum it
> must have something, even if its just a NAME, SYNOPSIS and
> DESCRIPTION.
This is arbitrary. I don't disagree that it's a must, but it might be
better to qualify this out a bit to see how this might work. I do,
however, think that it would be a good absolute to have a requirement of
standard internal documentation, not just for the core, but to CPAN as
well. Currently, a lot of authors are recklessly putting in documentation
in HTML files, readme's, etc. which makes building a PITA since special
steps have to be taken to get this help to any end user of a compiled
collection (and if you compile them yourself, you need to be aware of
every module author's individual idiosyncracies). I've been begging Dave
Roth for four years now to actually put documentation for ODBC.pm into
ODBC.pm (among several others). "Reasonable amount", however, is
arbitrary. Please qualify it.
> 2) It has a reasonable amount of tests. Minimum must be more than
> simply a compile test.
Although I fail to see any problem here, I know one's going to pop up. It
sounds like an absolute where one isn't needed. For example, for a
previous employer, I spent two months writing a Y2K perl testing suite to
sell to high-priced companies to prove that their Perl was Y2K safe. Of
course, the boss and I both knew that we were wanting to charge a lot of
money for basically saying if(0){you have a y2k problem in the Perl
core)... I'm sure there will be items where testing would be superfluous
or misleading.
> What is "reasonable" in each case is ultimately left up to the
Pumpking.
Always qualify your suggestions in group posts, please. The P5P may be
used to working as an oligarchical monarchy, but the P6 groups have a
happy democracy with only an occasional trump from above. Having abstained
from the P5P for so long, it's been a long time since I've seen a
suggestion with no content.
p
Re: Proposed basic criteria for accepting new core modules
Michael G Schwern <[EMAIL PROTECTED]> wrote: > Sorry, it wasn't clear to the perl-qa folks that I'm talking about > Perl5, not Perl6. > > No nonononono. Please don't drag CPAN into this particular > conversation. I'm just trying to nail down p5p on this simple issue. > > Sorry for not clarifying, this is purely a Perl5 thing. Ah, totally different story. For what it's worth, good luck. I remember this issue popping up in the P5P a couple of times before, and Murray Nesbitt and I have both addressed individual authors about it. However, I'll reassert, to us it's as important that authors write documentation in an orderly manner as it is that they write it at all, otherwise it's hard to find later and can get overlooked (and may as well not exist). I have seen this improve over the past couple of years, but there are still a couple of stubborn ones out there. Aldo and Dave Roth are mentionable. (I do that on purpose in an ongoing effort to try to hint them [strongly] into standards compliance, not to say that either aren't decent programmers.) ;-) p
Re: Anyone seriously using 5.004?
I do know of a couple who are. They mentioned it on perl-5-meta. On Wednesday 22 August 2001 21:11, Michael G Schwern wrote: > I've got 5.004_04, 5.004_05, 5.005_03, 5.6.1 and bleadperl installed > to test against. Should I bother with 5.004? Is anyone *seriously* > using it still? > > > PS When I say 5.004, I mean the original 5.004 release.
