<<   August 7, 2001   >>

* Short after YAPC::Europe, some foreign guests (o.a. Leon)

* JP is late from the restaurant, 7 man were waiting

* Abigail worked out Schwerns conference talk for total private
object data by using source filters (Sir, yes sir!!)

* Johan welcomes our geusts

* Some YAPC::Europe impressions

- extremely impressed by how all went, almost professional
- short team report of the problems encountered before and
during the conference
- what talks were fun

* Agenda.

- proposals?
- Ann: pure social

* Abigail: Randall pushes Ingy to do Inline RecDescent

* CPANTS / NAPC / CPAN

- Automated testing
- Smoke tests (spare CPU cycles)
- Module scoring, scaling, rating
- NAPC is making binary
- Quality service
- Liz: CPANTS@home (sandbox machines)
- Security problems might hinder people to participate
- chroot might help, tested modules are not installed
- if you are *that* scared, don't even download /anything/
- Someone in Germany already did this (Johan will look him up)
- How well does the tests cover the module functionality. More
problematic with XS and TIE stuff
- make it *reaaallllly* easy
- user replies: usefulness (see scaling/scoring/rating)
- Code review
- Docs before and/or during coding
- recommendations

- CPAN needs a serious rewrite
- depenancies: advice and automation

- general discussion about what info in a CPANTS database
is useful for whom.

* other useless bits

- intercal and 'Inline intercal'
- 'Inline forth' should be 'forth Inline'
- Filter::Simple, (Filter::Util::Col was bad)
- recommendations in module Foo::Bar for Foo::Bar::Simple
should apply moreoften (like LWP => LWP::Simple)
- Named arguments vs. speed
- Internal routines vs. API
- can one write a valid perl program with only uppercase letters?
JP gives it a try:

$!=16; # presumed error message = "Invalid File"
($A,$B)=$!~=/^..(...)......(.)/; # val e
`$^X $B$A`;

Now do it *without* backticks. The problem is that we cannot
call 'eval' (all lowercase) and we cannot create coderefs to
fill things like %SIG, die handlers, or overloading. Stings are
only evaluated at compile time, so building "@{[...]}" won't
be parsed runtime.

Several people now started hacking this and eventually turned
up with code that met the original request.

BEGIN{$^H=2097152};$_="\050?\173\160\162\151\156\164\042".
"\112\165\163\164 \141\156\157\164\150\145\162 \160\145".
"\162\154 \150\141\143\153\145\162\012\042\175\051";/$_/;

Now when testing to enter this code in an SMS message (a part
of the original problem), it turned up that it also did not
support the brace (or square brackets for the matter).
Tri-graphs were mentioned, but '??<' and the like are legal
perl, so it is unlikely that this would work.

Ohh, well, we'll just wait for Abigail's brain to finish on
this and we'll see this arise as a JAPH on comp.lang.perl.misc
very soon.

(Would people do this at Visual Basic user meetings?)