At 09:47 AM 4/16/2001 -0700, Peter Scott wrote:
>At 12:11 PM 4/16/01 -0400, Dan Sugalski wrote:
>>There are a number of reasons to *not* claim to parse perl 5 code.
>>
>>*) We won't load any perl 5 XS code
>>*) We won't be getting the corner cases, and perl5 has a *lot*.
>>*) It complicates the interpreter if we need to add code to support
>>things that perl 6 doesn't do. (*ocugh*typegobs*cough*)
>>*) It makes the parser that much more work to write. Even if we switch
>>parser code entirely when going from perl 5 to perl 6 and back, someone
>>still has to write the code in the first place.
>>*) It's even more complexity, which is just that much more stuff we can
>>potentially break
>>*) It sets up the expectation (rightly so, IMNSHO) in the user community
>>that we eat and process all perl 5 code correctly, and we just aren't
>>going to. Despite our best efforts we're not going to.
>
>As a very low-tech solution, why not bundle perl 5 *with* perl 6 so that
>once perl 6 detects that it's been fed perl 5 code, it can send it to the
>perl 5 compiler/interpreter.
Besides the size and clunkiness issues, there's another problem. String
evals in a perl 5 section of code will expect to be parsed as perl 5, which
kinda precludes the whole "compile perl 5 to bytecode and pass it through
the p526 converter" solution. Makes mixing and matching perl5 and perl 6
code rather more difficult. :(
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
