Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-05-16

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 ronsavage JK: Yes, good idea about not using LuaJIT; nice in theory but actually problematic.
00:18 ronsavage This article http://perlhacks.com/2016/05/taming-qmail/?utm​_source=feedburner&utm_medium=feed&utm​_campaign=Feed%3A+PerlHacks+%28Perl+Hacks%29 contains an interesting comment about moving from Moose to Moo for performance reasons.
01:48 ilbot3 joined #marpa
01:48 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Pastebin: http://scsys.co.uk:8002/marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today
04:11 idiosyncrat_ Just uploaded Marpa-R3-4.001_007, with Lua.  A few tests have passed already.
04:28 idiosyncrat_ 31 passes, no unexpected failures, which is a bit of a surprise.
04:29 idiosyncrat_ This is my first every release made w/o getting it working on all 3 of my test machines.
04:31 idiosyncrat_ I couldn't get the freestanding Lua to link on my Darwin PPC Tiger, which builds dual i386/PPC archives (the idea was that all Mac executables and libraries would work on all platforms, whether i386 or PPC).
04:32 idiosyncrat_ I'll see if I can get a newer Perl to install on it -- if not, I'll give up on it as a development/testing machine.
04:34 idiosyncrat_ It was nice to test all my bit-twiddling on a number of diverse architectures and my 3 test machines included i386, ARM and PPC.
04:35 idiosyncrat_ Some freebsd tests are coming in, also successes.
04:38 idiosyncrat_ For those curious, here
04:39 idiosyncrat_ 's a discussion of this architecture and these "universal" binaries in the perl docs: http://perldoc.perl.org/perlmaco​sx.html#Universal-Binary-support
05:10 idiosyncrat_ Good night!
05:12 idiosyncrat_ Oh yes
05:12 idiosyncrat_ rns: The new release almost certainly fails on Windows & Cygwin.
05:14 idiosyncrat_ But Roberto has suggestions in his Makefile: http://www.lua.org/manual/5.3/readme.html
05:15 idiosyncrat_ Specifically the Makefile has a recipte for mingw
06:57 ronsavage joined #marpa
08:15 rns joined #marpa
08:16 rns idiosyncrat: re http://irclog.perlgeek.de/m​arpa/2016-05-16#i_12483822 -- cygwin ok, windows failed -- will look into it later, thanks for the link.
08:16 rns left #marpa
08:44 koo7 joined #marpa
08:45 pczarn joined #marpa
11:15 pczarn My engine sorts items in each set with a O(n log n) sort. Am I thinking right that the parser's time complexity will be O(n), O(n**2 log n) or O(n**3 log**2 n)?
11:15 pczarn It's unfortunate, but close enough.
13:17 JPGainsborough joined #marpa
13:35 kaare_ joined #marpa
13:41 pczarn I'd like to fix and update the sequence rewrite document. Can I make it more similar to my implementation?
15:24 idiosyncrat_ joined #marpa
15:25 idiosyncrat_ pczarn: I've forgotten where that doc is.  Could you remind me with a link?
15:26 idiosyncrat_ Re revising, fixing.  It's purpose IIRC was to reflect my plans, thinking, so maybe for your you'd want to start your own doc.
15:26 idiosyncrat_ On the other hand, if my doc is wrong, I certainly want to fix that!
15:27 idiosyncrat_ pczarn: re an extra log n -- here's a secret.
15:28 idiosyncrat_ Most theoreticians, when it comes time to implement, throw in an extra O(log n) and never tell anyone.  In fact, if there is an O(1) alternative to the O(log n), but you don't bother to implement it, that's considered legit.
15:28 idiosyncrat_ In Libmapra I sweated the difference between O(1) and O(log n), so my O(n) is *really* AFAICT O(n).
15:29 idiosyncrat_ Doing that in this sort of practical application is unusual.
15:30 idiosyncrat_ If you were publishing, the referees probably wouldn't care about the difference.
15:32 idiosyncrat_ Richard Lipton, one of my professors, attributes a saying to another of them Stan Eisenstat: "O(n) = O(n log n)"
15:35 idiosyncrat_ https://rjlipton.wordpress.com/2011/01/1​9/we-believe-a-lot-but-can-prove-little/
16:34 pczarn idiosyncrat_: done, https://github.com/jeffreykegler/kollos/pull/60
16:36 pczarn now, I'd like to rename "terminator" separation to "trailing", and do some rearrangement that is harder to sum up.
16:38 pczarn For example, replacing the 6-tuple with a 5-tuple.
16:39 pczarn Or 4-tuple.
16:54 choroba joined #marpa
17:06 idiosyncrat_ pczarn: I'm immersed in other stuff at this point, so I can't make time for the detailed thinking on that stuff.  Maybe you want to fork?
17:06 idiosyncrat_ When I come back to the stuff, roles may reverse, and I'll be studying your document.
19:16 koo7 joined #marpa
19:22 pczarn joined #marpa
23:18 ronsavage joined #marpa

| Channels | #marpa index | Today | | Search | Google Search | Plain-Text | summary