Friday June 22, 2007
Perl 6 Design Meeting Minutes for 20 June 2007
The Perl 6 design team met on 20 June 2007. Larry, Damian, Patrick, Nicholas, and chromatic attended. These are the minutes.
- mostly in talk preparation mode
- in the phase of building up sufficient panic to write something down
- Damian, did you put together anything on "What's new in Perl 6?"
- I will by next week
- what day do you need it by?
- Larry's presentation's on Tuesday
- I have to put one together for YAPC::EU
- I can probably pretty much wing it
- anything you have would be nice
- mostly in the throes of upgrading to FC 7
- I tried an upgrade first
- now it's a fresh install
- then you spend days just running to stay in place
- seems to have settled down finally
- hacking more on the spec and the grammar
- nothing major there
- looking at some of the tests to see what kind of state they're in
- trying to get them in a portable state
- by and large, that'll take an army of ants rather than just one person
- did find and add more missing things from the standard grammar
- it's certainly stabilizing
- I've basically been destroying all goodwill on p6l
- discussing aspects of the new POD syntax
- basically trying to have a rational argument with people coming at it from opposing positions
- surprisingly difficult
- things seem to have calmed down after Larry ruled on the issue
- reworking the Perl 6 POD parser module to accept Larry's ruling, as best it can
- working on an extension to Perl 6 POD which allows you to extract bits of code and put into documentation
- otherwise preparing for my extensive September trip
- busy week here
- we have the 01-sanity tests running under Perl 6 for Parrot
- we're looking at the Test.pm file from Pugs
- it requires a lot of things not in the sanity tests
- we've today put together a throwaway Test.pm so that we can run several of the other tests without having to make all of the real Test.pm work
- when we get more features, we'll merge the two back into a single unit
- I can also now parse the action stubs in the standard grammar
- we don't need a pre-processing step now
- it may be a throwaway thing, but I can play with it and do experiments with it
- I'll check it in later today; I held off until last night's Parrot release
- we've been doing a lot of cleanup of the Perl 6 language directory today
- trying to make it easier for contributors to that part
- also wrote a message to the lists about the plans for further development
- any specific plans for the hackathon?
- plan to use the presentations to build on what people know and show what we need to do
- then I'll incorporate plenty of Pugs tests so we'll have plenty of visible things to do
- worked this weekend on nqp, Not Quite Perl 6
- little language to work on the compilation phase
- uses the same kind of grammar constructs that are in the standard grammar
- Patrick, everything's really cool!
- Larry mentioned an army of ants and things that need to be done
- is there a way to hook those up to the microgrants?
- I can come up with a list of those
- I should do some myself
- that's an excellent idea
- Larry, you ruled that the way POD works in Perl 6 is that the POD parser is free to treat everything that looks like POD as POD
- the Perl 6 parser is free to treat everything that looks like a string, even if it contains POD, as a string
- is that the only case in where the Perl parser prefers Perl to POD?
- what if an assignment operator occurs in column one?
- do something more like Perl 5 does?
- only recognize POD where the parser expects a statement?
- I don't think we have to go that far
- if code gets misparsed as a chunk of POD, it will likely produce some sort of message indicating a misparse
- I want to advance that position myself
- POD 6 has a careful design
- intended to interleave POD6 and Perl; easier than it was, as it doesn't require blank lines
- people can document their code at a finer grain
- one line of code, one line of POD, one line of code, one line of POD
- could document the creation of a list of something
- that won't work if the parser treats things that look like assignments as assignments
- in Perl, you only get an
= in the first column if you write in a very bad style
- I'm not inclined to cater to that
- that makes my life a little bit easier
- especially for retrofitting Perl 6 POD back onto Perl 5
- I can identify Perl 5 strings and be careful with those
- I'm happy with that
- I'll update S26 and S02 or S03 with the interaction between Perl and POD
- it's the "Doctor, it hurts if I do this!" isn't it?