Perl 6 - the future is here, just unevenly distributed

IRC log for #parrotsketch, 2009-02-03

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

All times shown according to UTC.

Time Nick Message
12:03 kid51 joined #parrotsketch
12:04 kid51 * Preposting because I'll be at $job during meeting time, as usual.
12:04 kid51 * Sinus cold kept me inside but made real thinking painful.
12:04 kid51 * So I decided to plow through old RTs instead.  Much easier on the sinuses.
12:05 kid51 * Focused mainly on those filed in last 3 months of RT's heavy use.
12:05 kid51 * Tended to close tickets which merely reported build errors on Linux or Darwin.
12:05 kid51 * Identified some where the discussion appeared left up in the air, requested update from discussants.
12:05 kid51 * Identified others where a quick review from a core committer would probably lead to resolution. (Mainly: allison, particle, coke, infinoid, pmichaud)
12:05 kid51 * One or two that I can work on myself.
12:05 kid51 * So if you were nudged or nagged by my postings this weekend, please take a look.  Aiming to get below 500 RTs open/new/stalled.  (529 as of 0316 UCT today).
12:06 kid51 * Also, we will need to develop a plan to address the dozens of RTs opened by PTC for XXX or TODO items.  Some of the TODO comments may simply have been random musings by Parrot developers in eons past.  Someone with extensive Parrot history could probably evaluate them quickly.
12:06 kid51 * Also, if you're making changes in lib/Parrot/*.pm, please run perl Configure.pl --test before committing to make sure you're not breaking any tests.
12:07 kid51 * Also, we should urge use of branches for changes that have potential either to break build or cause new failures in 'make test' that can't be resolved in less than an hour.
12:07 kid51 EOR
12:25 integral joined #parrotsketch
16:10 davidfetter joined #parrotsketch
16:21 pmichaud joined #parrotsketch
16:33 particle i'll likely miss #parrotsketch due to a schedule conflict.
16:34 particle ~ modified config subsystem to allow the use of ccache on windows, with msvc, icc, bcc, and mingw
16:34 particle .end
16:34 particle please do a review of roadmap tickets in my absense--i'll try to update mine before the meeting, but this may not be possible.
17:18 kj joined #parrotsketch
17:19 kj preposting my non-report, as I won't make it for #ps:
17:19 kj no report.
17:19 kj .eor\
17:19 kj left #parrotsketch
17:19 masak joined #parrotsketch
17:24 rurban joined #parrotsketch
17:56 PacoLinux joined #parrotsketch
17:58 Whiteknight joined #parrotsketch
18:11 Tene joined #parrotsketch
18:19 allison joined #parrotsketch
18:24 masak joined #parrotsketch
18:29 barney joined #parrotsketch
18:30 cotto HI.
18:31 masak tja!
18:31 barney Servus
18:32 rurban hi
18:32 pmichaud hello.
18:33 chromatic joined #parrotsketch
18:35 Whiteknight hello
18:35 chromatic Who else is here?
18:36 pmichaud Present.
18:36 * masak here
18:36 * cotto too
18:36 * rurban also
18:36 allison hi
18:36 chromatic Okay, let's go.
18:36 chromatic allison?
18:36 allison - A successful migration of the subversion repository to svn.parrot.org.
18:37 allison - Took up chromatic's ticket challenge and closed a handful a day, applying patches, fixing bugs, clearing out some old TODOs that don't fit with current architecture.
18:37 allison - Started, completed, and merged in the second string refactor branch, a large-scale function name cleanup.
18:37 allison - A few fixes on the Pod parser, mostly handed it off to kj (with thanks).
18:37 mberends joined #parrotsketch
18:37 allison EOR
18:38 chromatic barney?
18:38 barney Was and will be busy with new freelance project.
18:38 barney [Pipp]
18:38 barney Added 'standard' and 'Reflection' to get_loaded_extensions().
18:38 barney Re-added support for the phc variant, as James Dupont wants to hack on it.
18:38 barney Plan to move to github this weekend
18:38 barney .eor
18:38 chromatic I didn't make much progress.
18:38 chromatic cotto?
18:38 cotto * making slow but steady progress converting PMCs to use ATTR instead of UnionVal macros
18:38 cotto * found 2 moderately obscure bugs that I initially missed due to poor formatting of test results
18:38 cotto * also found one wierd PCC bug after converting FixedIntegerArray
18:38 cotto * am pausing to improve Parrot's test reporting (or get someone else to do so) to avoid more such bugs
18:38 cotto * also, attempting to fix any bugs I cause
18:38 cotto goto NEXT();
18:39 chromatic japhb?
18:39 Coke joined #parrotsketch
18:39 chromatic Coke?
18:40 Coke hio. I apparently need a calendar reminder for this meeting.
18:40 Coke I have tried to pull out some more deprecated features in the past week. Tcl is currently failing a bunch of tests against parrot trunk and I haven't figured out why (could be my fault.) working on some
18:41 Coke documentation of sorts that will hopefully hit this weekend.
18:41 Coke .
18:41 chromatic masak?
18:41 masak * big refactor of Druid, taking advantage of some of the OO stuff
18:41 masak * found about the usual number of bugs
18:41 masak * ...with the normal distribution of nastiness
18:41 masak * looking forward to someone acting as a counterweight to my bug reporting :)
18:41 masak eor
18:41 chromatic pmichaud
18:42 pmichaud Got distracted with various family things this past week.
18:42 pmichaud Been mainly focusing on adjusting the rakudo repository and re-designing its build sequence.
18:42 pmichaud I expect to have that fairly complete by Thursday.
18:42 pmichaud other than that, just cleaning things up and preparing for Frozen Perl.
18:42 pmichaud EOR
18:42 chromatic rurban?
18:42 rurban - Shook native_pbc: Enabled t/native_pbc/integer.t and number.t tests to get
18:42 rurban this fixed before 1.0.
18:42 rurban Found at least 3 bugs with 64-bit, and fixed one and a half, TT #254.
18:42 rurban This caused some smokes to blow up but now it should be settled.
18:42 rurban I can reproduce all yet detected bugs. I work on solaris-64int on this.
18:43 rurban But I still need a bigendian 64int box for smokes and native pbc's, and
18:43 rurban for the funky floattype 1 (12-byte double BE).
18:43 rurban - Added missing @linkflags@ to the linker, detected on solaris-64int, TT #262
18:43 rurban - Need review for a pbc_disassemble roundtrip patch with new options --bare -b, ...
18:43 rurban (assembable disassemblies) => TT #258
18:43 rurban - Need review for a PackFile API rename patch: Parrot_loadbc => Parrot_pbc_load,
18:43 rurban Parrot_readbc => Parrot_pbc_read plus one new debug argument. => TT #266
18:43 rurban Thanks to particle. I need the 3rd arg for easier packfile debugging.
18:43 rurban - Created parrot MSVC6 project files (dsp/dsw) which might be of use to someone.
18:43 rurban Some files allow /ZI ("Edit and Continue"), some not. #include .str fails
18:43 rurban - I'll produce a single RT #40817-track-generated-files against trunk, again.
18:43 rurban EOR
18:43 chromatic Tene?
18:44 chromatic Whiteknight?
18:44 Whiteknight * Doing some misc GC-related cleanups
18:44 Whiteknight * Going through a lot of old tickes, performing triage and getting things resolved
18:44 Whiteknight * Misc twiddling on the various books.
18:44 Whiteknight * Sent out a proposal to the list yesterday about changing the way "self" is used. Would appreciate feedback
18:44 Whiteknight * Other miscellaneous digging into various issues but nothing to show for it all.
18:45 Whiteknight EOR
18:45 chromatic Did I miss anyone?
18:45 chromatic Question time.
18:45 chromatic Who has a question?
18:46 rurban installable langages, how to proceed?
18:46 chromatic What's the status of the PDD there?
18:46 rurban There's none
18:47 chromatic Languages without C-based dynpmcs or dynops are just Parrot libraries.
18:47 chromatic Those should be easy to install.
18:47 Coke there's also the question of how to /build/ them.
18:47 rurban there's the draft pdd30, where to install hem? that's not decided
18:48 rurban the build question is 2nd (CPAN or a Parrot install.pbc)
18:49 pmichaud in 2008-09 it was suggested that we have  parrot/runtime/languages/
18:49 rurban /usr/lib/parrot/runtime/languages ?
18:49 pmichaud sorry
18:49 pmichaud runtime/parrot/languages
18:50 pmichaud i.e., as a sibling to runtime/parrot/library
18:50 pmichaud (looking up the reference now)
18:50 rurban /usr/lib/runtime/parrot/languages ?
18:50 allison I was just thinking about this this morning, specifically for building packages
18:50 allison /usr/lib/parrot/languages
18:51 rurban We have to think of an absolute path which is acceptable to packagers, not as now, where we still create /usr/runtime ansd such mess
18:51 pmichaud previous discussion was at http://irclog.perlgeek.de/par​rotsketch/2008-09-09#i_559127
18:51 allison then /usr/lib/parrot/runtime is our install of the directory runtime/parrot
18:51 diakopter joined #parrotsketch
18:51 rurban sorry, there's no /usr/lib/parrot/runtime
18:52 allison rurban: not currently, no, but moving to that instead of /usr/runtime
18:52 rurban there's just /usr/lib/parrot/library dynext and include
18:53 rurban In my latest package I've toyed with /usr/lib/parrot/languages
18:53 allison library is a reasonable substitute for runtime
18:53 rurban /usr/lib/parrot/languages/abc/abc.pbc
18:53 allison though, perhaps confusing to have lib/.../library
18:54 pmichaud I think 'library' corresponds to runtime/parrot/library of the build tree
18:54 rurban sure
18:54 allison pmichaud: yes, but what makes sense in the repository might not make sense in the install
18:54 chromatic 'library' is fine as a path component to me; we should not assume that it'll always be under /usr/lib -- it could be under /opt or $HOME or wherever.
18:54 allison chromatic: yes, that's determined by the prefix
18:55 allison chromatic: configurable at in config process
18:55 rurban /usr/lib is the default $libdir    It's just that some languages also have their runtime subpath and that doesn't fit
18:55 chromatic Then who cares if lib/.../library ?
18:55 allison chromatic: it's just about sane defaults
18:55 rurban We just have to decide where to put the languages and how
18:56 allison (because most packaging will use the defaults)
18:56 pmichaud I propose  /usr/lib/parrot/languages/abc.pbc
18:56 TimToady don't forget to think about versions and authorities
18:56 Coke pmichaud: /usr/lib/parrot/languages/abc/abc.pbc, at least, so we can install multiple pbcs for a single lang.
18:56 allison pmichaud: which is also what I propose, and what rurban proposed after me, so, consider it done
18:56 rurban I proposed /usr/lib/parrot/library/abc.pbc
18:57 rurban in my draft
18:57 allison oh, sorry, I missed the extra language directory
18:57 pmichaud Coke:  no, I'd like additional pbcs to go into .../parrot/languages/abc/
18:57 allison what I wrote this morning was /usr/lib/parrot/languages/pynie/pynie.pbc
18:57 pmichaud but the base level one should be    .../parrot/languages/abc.pbc
18:57 Coke pmichaud: I disagree, but don't care enough to write any code. =-)
18:58 pmichaud I'm basing this on the notion that 'abc' would be linked to 'parrot' and automatically load abc.pbc
18:58 pmichaud (as discussed last fall)
18:58 Coke pmichaud: adding another dir there doesn't really impact that.
18:58 barney but there usually it the exe  /usr/lib/parrot/languages/abc
18:58 TimToady don't forget to think about versions and authorities :)
18:58 allison pmichaud: it does make library loading simpler if there's only one directory to look in
18:58 pmichaud TimToady: I haven't forgotten.
18:58 pmichaud TimToady: we'll deal with that in Rakudo, yes.
18:58 allison TimToady: noted, and will consider
18:59 Coke Also, need a place for oplibs, dynpmcs...
18:59 rurban ok, sounds good. then the call would be load_bytecode 'languages/pynie.pbc'
18:59 Coke (versions, authorities) I can't see that going in for 1.0
18:59 TimToady just pointing out that if you do the symlink trick to parrot, it might have to find version numbers in the path, not just the final component
18:59 allison Coke: yes, I'm thinking it goes languages/foo//oplibs languages/foo/dynpmcs and languages/foo/library
18:59 pmichaud I'm okay with /usr/lib/parrot/languages/abc/abc.pbc then.  I'm more interested in seeing the "load correct language based on argv[0]" working.
18:59 Coke we don't have a way to say "I want version 3.7 of the tcl runtime."
19:00 rurban dynpmc's should better be put globally
19:00 Coke rurban: why?
19:00 rurban how to call it?
19:00 allison We could work that in as languages/tcl37/... but at that point I think we should really have a language config file in /etc
19:00 Coke rurban: same way we do now, by using .loadlib?
19:01 rurban .loadlib languages/foo/oplib
19:01 pmichaud does .loadlib look in subdirs?
19:01 allison rurban: no need for dynpmcs to be global, and it would be troublesome for filenaming
19:01 allison pmichaud: no, but you can add to the search path
19:01 pmichaud allison: how does one do that?
19:02 allison pmichaud: so, that would be part of installing a language
19:02 pmichaud I thought .loadlib got compiled into the .pbc
19:02 allison someone was just working on that code a couple of weeks ago... have to look
19:03 allison or not (too lengthy for the moment)
19:03 allison but, I'll look at it today
19:03 rurban I hope it will work without polluting LD_RUN_PATH / PATH / DYNLIB_PATH
19:03 allison I'm going to build a pynie package
19:03 pmichaud punie, please?
19:03 pmichaud pynie's broken at the moment, I think.
19:04 allison pmichaud: okay, or abc, doesn't really matter which language
19:04 pmichaud right, abc is another good choice.
19:04 pmichaud downside of those languages is that they don't have dynops or dynlibs
19:04 rurban WMLScript is the most troublesome
19:04 allison yeah, that's why I wanted a meatier one, but not quite as complex as perl6
19:04 allison maybe lua
19:04 pmichaud pynie doesn't have dynops or dynlibs either, I don't think.
19:05 allison ah, okay, haven't looked yet
19:05 pmichaud a simple (and very useful) one to get working would be nqp.
19:05 rurban forth is also problematic because it has multiple unmergable pbc's
19:05 pmichaud because it's needed for compiling all of the others.
19:05 pmichaud Perl6Grammar is another one that we need for building other languages.
19:05 allison pmichaud: nqp may also be more complex than the average language
19:06 rurban nqp and Perl6Grammar are libs not languages
19:06 pmichaud nqp is incredibly simple.
19:06 pmichaud it's just the .pbc file
19:06 pmichaud the whole point of nqp is that it doesn't have a runtime :-)
19:06 allison I'd kind of expect that we'd install nqp by default, but it still could go into /usr/lib/languages/nqp
19:06 pmichaud rurban:  nqp is a compiler.
19:06 rurban I know :)
19:06 pmichaud rurban: Perl6Grammar is a compiler.
19:07 pmichaud Thus I doubt they're 'libs'
19:07 allison maybe we should have /usr/lib/parrot/compilers too?
19:07 allison though, the distinction between languages and compilers is fairly thin
19:07 rurban but it should go into /usr/lib/parrot/library/nqp.pbc shouldn't it?
19:07 pmichaud is there a substantial distinction between 'language' and 'compiler'?
19:07 pmichaud rurban: I would expect to be able to run programs written in NQP with just a 'nqp' command.
19:07 allison at the notional level yes, but at the install level?
19:08 pmichaud in that sense, it acts like a language.
19:08 particle i thought the difference was something like: compilers are shipped with parrot, languages aren't
19:08 pmichaud particle: that's been the traditional difference for the naming of the directories, yes.
19:08 pmichaud particle: I don't think that was intended to indicate how they'd be used.
19:08 pmichaud or invoked.
19:08 allison pmichaud: yes, that's the other thing I'm working on today, the busybox aliasing so you can run any language as 'languagename' instead of 'parrot languagename.pbc'
19:08 pmichaud anyway, I think that 'nqp' should be invokable like any other language.
19:09 allison (which requires a standard location for the languagename.pbc file, which is how I got started on this)
19:09 rurban we have now  /usr/lib/parrot/library/P6object.pbc  /usr/lib/parrot/library/PGE/Perl6Grammer.pir
19:09 pmichaud allison:  yes.  I'm just saying nqp.pbc belongs there also.
19:09 allison pmichaud: in that case, I would put nqp in /usr/lib/parrot/languages/nqp (to keep it standard)
19:09 pmichaud allison: and that nqp is a very simple test case, and an immediately useful one because rakudo/abc/pynie/punie/pipp/etc. all need to run nqp in order to build.
19:10 rurban Files included in the =parrot-perl6= package:
19:10 rurban /usr/bin/perl6.exe
19:10 rurban /usr/lib/parrot/dynext/perl6_group.dll
19:10 rurban /usr/lib/parrot/languages/perl6/perl6.pbc
19:10 allison pmichaud: true, okay I'll dig into it and pick a language to "debianize" as a test case
19:10 rurban /usr/lib/parrot/library/P6object.pbc
19:10 rurban /usr/lib/parrot/library/P6object.pir
19:10 rurban /usr/share/doc/parrot-0.9.0​/languages/perl6/MAINTAINER
19:10 rurban /usr/share/doc/parrot-0.9.0/languages/perl6/README
19:10 rurban /usr/share/doc/parrot-0.9.0/languages/perl6/STATUS
19:10 rurban /usr/share/doc/parrot-0.9.0/langu​ages/perl6/compiler_overview.pod
19:10 rurban /usr/share/doc/parrot-0.9.0/​languages/perl6/glossary.pod
19:10 rurban /usr/share/doc/parrot-0.9.0/langu​ages/perl6/spectest-progress.csv
19:10 rurban /usr/share/man/man1/perl6.1.gz
19:10 pmichaud rurban: all of those are going away.
19:10 pmichaud (well, not P6object)
19:11 rurban /usr/lib/parrot/library/PGE/Perl6Grammer.pir is in parrot
19:11 allison rurban: that one stays
19:11 pmichaud but should it remain  in library/
19:11 pmichaud ?
19:11 pmichaud it's really a compiler.
19:11 allison yes, it's a core parrot library
19:11 pmichaud it's not a library, at least not in the sense of "load this library so one can do xyz"
19:11 allison it's also not a language
19:12 pmichaud yes it is.
19:12 pmichaud It's the compiler for Perl 6 grammars.
19:12 barney let's call it a 'tool'
19:12 pmichaud I'm happy to call it a tool.
19:12 allison it's either going in /usr/lib/parrot/library or /usr/lib/parrot/languages/pge, I prefer the former
19:12 pmichaud but Perl6Grammar is the thing that takes a grammar (such as the pod grammar) and converts it to executable PIR.
19:13 particle it's a compiler/language, not a library like Data/Escape
19:13 particle languages/Perl6Grammar/ seems best to me
19:13 allison yes, the line between tools and languages gets thinner when we're writing DSLs as part of the tools
19:13 * chromatic bonks you all on the head
19:13 pmichaud I'd be very surprised for anyone to do   load_bytecode 'Perl6Grammar.pbc'
19:13 allison but, in general, I want the core tools to be installed as core libraries
19:14 pmichaud the normal invocation for Perl6Grammar is from the command line.
19:14 allison particle: no, it's not a separate language
19:14 rurban It's load_bytecode 'PGE/Perl6Grammar.pbc'
19:14 rurban No load_bytecode 'PGE/Perl6Grammar.pir'
19:14 allison okay, we're bikeshedding now
19:14 pmichaud rurban: but nothing actually does that.
19:14 allison this part isn't critical
19:14 pmichaud allison: it is to rakudo.
19:14 allison (have to figure out a strategy for languages, but the rest can wait)
19:14 rurban ok,
19:15 pmichaud I'll rephrase
19:15 allison rakudo needs a way to load PGE/Perl6Grammar.pbc, it already has that as long as it's installed in a standard library location
19:15 pmichaud allison: wrong.
19:15 pmichaud rakudo need a way to *run* PGE/Perl6Grammar.pbc
19:15 pmichaud Rakudo does not load PGE/Perl6Grammar.pbc .
19:15 allison ?
19:15 pmichaud that has been my point.
19:15 pmichaud PGE/Perl6Grammar.pbc is never loaded as a library.  It is always run as a command.
19:16 allison pmichaud: you're talking about:
19:16 allison $(PARROT) $(NQP_DIR)/nqp.pbc $(PGE_LIBRARY)/Perl6Grammar.pir $(SOURCES)
19:16 barney Does Rakudo need to run  PGE/Perl6Grammar.pbc after rakudo.pbc has been built?
19:17 pmichaud barney: no.
19:17 pmichaud Perl6Grammar is only used for building languages.
19:17 allison but, that means it needs to be available as part of the build tools
19:17 Whiteknight I thought the intent of Perl6Grammar was to be used in Rakudo
19:17 allison that is, part of the 'parrot-dev' package, or some such
19:17 pmichaud (yes, it's possible that someone _can_ load it as a library later, but that would be done in the same sense that someone could load perl6.pbc as a library to run Perl 6 code)
19:18 allison Whiteknight: no, it's just the standard way of compiling a PGE-based parser
19:18 rurban So it doesn't need to be put in parrot, just parrot-devel
19:18 pmichaud the only reason that Perl6Grammar.pbc lives in library/PGE/Perl6Grammar.pbc now is because there's never been anywhere else good to put it.
19:18 allison rurban: yes, but that means all language packages have a build requirement of the parrot-devel (which seems reasonable)
19:19 pmichaud (perhaps it could've gone into compilers/pge.  But that wasn't obvious at the time it was created.)
19:19 pmichaud what is "parrot-devel"?
19:19 rurban a packlage name
19:19 pmichaud does it get installed?
19:19 rurban we have those names in the MANIFEST
19:19 allison pmichaud: an "invented on the fly" name for a package that contains the build tools
19:19 rurban users can choose
19:19 pmichaud so, not by default?
19:19 rurban basic runtime parrot or parrot-devel for more
19:20 allison pmichaud: yes, not on for a default install of parrot
19:20 Coke I wouldn't include the language development tools by default, no.
19:20 allison pmichaud: like how many C packages don't install the headers by default
19:20 barney pre, parrot runtime engine
19:20 pmichaud fair enough.
19:20 chromatic Is this discussion still useful for #ps, or does it belong on #parrot now?
19:20 rurban over to parrot
19:20 pmichaud I'm fine either way.
19:20 Coke Why don't we table it for a second until we find out if there's anything else.
19:20 chromatic Other questions?  Milestone review?
19:20 Coke then the meeting can close, and this discussion can continue.
19:21 rurban for that's all  ,thanks
19:21 allison chromatic: so far it's been critical to what I'm working on today, but it's wrappable now
19:21 pmichaud I agree with allison.
19:21 chromatic Okay, let's go over the milestones quickly then.
19:21 chromatic PDD 26 AST, pmichaud and Tene -- progress?
19:21 pmichaud not this week, but I'm also not feeling behind.
19:21 pmichaud it may be a good plane-task.
19:22 chromatic PDD 19 PIR, Whiteknight and Tene?
19:22 pmichaud (since I'll be several hours on a plane this weekend)
19:22 * particle steps away for another job-related phone call &
19:22 chromatic User documentation, allison?
19:22 chromatic PDD 23 exceptions docs, Tene?
19:22 Whiteknight eh?
19:23 allison did some documentation cleanup, mostly focused on dev tasks, would like to hand off documentation task
19:23 chromatic https://trac.parrot.org/parrot/wiki/ParrotRoadmap
19:23 chromatic Installable dynops and dynpmc tools, pmichaud, allison, particle?
19:23 allison (or at least be sure I'm not the only one working on it, it's more of a group sprint than a single person task)
19:24 pmichaud installable dynops -- I added rakudo requirements to the ticket, I need to find out more about parrot-devel
19:24 allison pmichaud: what problems did you run into moving Rakudo out of the repository?
19:24 pmichaud allison: currently Rakudo depends on having parrot's reconfigure.pl script... but I can get away from that since parrot-config exists.
19:25 pmichaud I expect to have that resolved today/tomorrow.
19:25 pmichaud (most likely today)
19:25 allison pmichaud: okay, that would be a good dependency to eliminate
19:25 pmichaud actually, we may get to the point where rakudo distribution is just a .pbc :-)
19:25 pmichaud (and some docs)
19:25 allison pmichaud: that works
19:26 pmichaud oh, we can't do that.
19:26 pmichaud darn.
19:26 pmichaud (dynops and dynlibs)
19:26 chromatic Support policy: I'll check it in somewhere to the repository.  It just needs a name.
19:26 Coke chromatic: perhaps a wiki page would be better?
19:26 allison docs/project/support_policy.pod
19:27 Coke we still have missed tasks from the 0.8.2 release, btw.
19:27 chromatic I'm sure we can link it off of our website.
19:27 Coke (according to the roadmap)
19:27 chromatic Windows porting, allison, particle?
19:27 pmichaud Coke: weren't the missed tasks moved forward?
19:28 allison Coke: we're not updating them in the earlier milestones once they've been moved to a later milestone
19:28 Coke allison: that's incredibly confusing.
19:28 pmichaud Coke: hopefully we're moving away from the roadmap wiki page anyway :-)
19:28 Coke I suggest if we're going to keep that page we at least change the status to 'rescheduled'
19:28 pmichaud +1
19:28 allison Coke: yes, good idea
19:29 Coke I'll do that right now.
19:30 Coke only one left that wasn't brought forward: nan/inf, pdd14-numbers
19:31 allison Coke: IIRC, that was landed. chromatic?
19:31 chromatic particle did the nan/inf, but I haven't made BigInt work with the IBM standard decimal library yet.
19:31 Coke also: garbage collectable contexts
19:32 particle coke: that was moved to long term roadmap iirc
19:32 allison Coke: those were moved out to 2.0 or so
19:32 particle gc contexts
19:32 particle i need to document nan/inf in the pdd
19:32 chromatic Windows porting, allison, particle?
19:32 particle haven't attacked it this past week
19:32 allison Coke: so, move the pdd14-numbers item to 0.9.1
19:32 Coke allison: ah. that means they're on a different page. blah.
19:33 allison chromatic: also haven't been running compiles on Windows this week
19:33 allison Coke: just say "rescheduled"
19:33 Whiteknight particle: what needs to be done on Windows?
19:34 Coke (blah. a better use of my time here is to kill these pages entirely)
19:34 pmichaud (Coke +1)
19:34 barney (+ Coke 1)
19:34 particle Whiteknight: there are many skip/todo tests on windows, some of which (like threads) are unimplemented or partially implemented features
19:34 allison Coke: I created https://trac.parrot.org/parrot/report/14 so all you have to do is tag a ticket as 'roadmap' and we have our roadmap page
19:35 chromatic PDD 20 lexicals, pmichaud?  Plane time?
19:35 pmichaud chromatic: yes.
19:35 Coke allison: there's already a report page for that.
19:35 * particle wishes for named report links.
19:35 chromatic MMD and tickets resolution, allison?
19:35 * Tene finally here, was stuck in traffic for an hour.
19:35 allison Coke: yes, that one (I created it a while ago)
19:35 Coke no, https://trac.parrot.org/parrot/roadmap
19:36 allison Coke: that's different, that's all tickets that are tagged for a milestone, no mater what kind of ticket they are (bug, patch, todo)
19:36 allison Coke: which means the actual roadmap tasks get lost
19:37 Coke I don't think we need yet another ticket status, but if you think it helps, ok.
19:37 allison Coke: the new report is just roadmap tasks, so it exactly corresponds to the old Roadmap wiki page
19:37 Coke (we already have critical, blocker...)
19:37 Tene Queue a couple question slots for me.
19:37 allison Coke: aye, but none of those say whether it's a task on our project plan for a particular release
19:38 allison Coke: I want to eliminate the Roadmap wiki page entirely
19:38 Coke that's why we have /milestones/
19:38 pmichaud milestones are our roadmap.
19:38 allison Coke: nope, that also includes stuff that isn't part of the project plan
19:38 pmichaud tasks not associated with a milestone aren't on the roadmap.
19:38 allison but many tasks associated with a milestone aren't on the roadmap
19:38 pmichaud right.  I think they should not be associated with a milestone then.
19:39 pmichaud i.e., the milestones should be our roadmap.  Things not on our roadmap should not be associated with a milestone on that roadmap.
19:39 Coke Nevermind. I'll just leave it the way it is.
19:39 Coke (I don't agree with the plan, and I have other ways to spend these tuits.)
19:40 allison I don't much care as long as there's a way to record the tasks that are specifically for the roadmap, and keep them from getting buried under a pile of other tickets
19:40 allison if milestone tag policing is the way to go, go for it
19:41 pmichaud allison: here's a question
19:41 pmichaud (related)
19:41 pmichaud if I come up with a bug that's not specifically on the roadmap, but has to be fixed in order for us to consider the associated item to be "done", how do I mark that?
19:41 pmichaud for example, the fact that Perl6Grammar.pbc isn't being included as part of 'make install'
19:42 pmichaud 'make install' is on the roadmap
19:42 pmichaud as it existed this morning, it didn't install Perl6Grammar.pbc
19:42 allison you can link to it from the ticket that depends on it
19:42 pmichaud yes, but shouldn't it also be a dependency on the 'roadmap'?
19:42 pmichaud i.e., shouldn't the roadmap indicate that that task has to be completed as part of the release?
19:43 pmichaud because without it, we haven't really gotten 'make install' to work.
19:43 allison The one ticket is a dependency for the roadmap
19:43 pmichaud oh -- you're saying modify the roadmap ticket to link to the dependency.
19:43 pmichaud okay.
19:44 allison the other ticket is a dependency for the other ticket
19:44 allison (interrupted by phone)
19:44 pmichaud to me it's just easier to say "this has to be done as part of the release also".
19:44 pmichaud but yes, I can try it the way you describe.
19:45 allison the other possibility is to tag it for the milestone
19:45 allison and, I actually prefer that one
19:45 pmichaud right, that's what I'm thinking
19:45 pmichaud I'd like the milestone to list all of the things that need to get done
19:45 allison so, it's not a "roadmap" task directly, but it is part of our "todo"s for the milestone
19:45 allison yes, I agree with that
19:45 pmichaud okay, I agree with you.
19:45 pmichaud so, 'roadmap' is just to identify the tickets that we specifically named at pds
19:46 pmichaud it's a subset of things that need to be done for the milestone
19:46 allison yes
19:46 pmichaud the milestone lists all of the things we need for the release.
19:46 rurban Perl6Grammar.pbc is just fixed with r36333, RT #40817 should be a dep
19:46 allison aye, he roadmap is a particularly important subset, from the overall plan
19:46 pmichaud okay, that makes sense to me.
19:46 rurban a proper MANIFEST.generated not to miss a built pbc
19:48 pmichaud I have to depart... see you all on #parrot or next week here
19:48 Tene Bye.
19:49 chromatic Tene, would you like to report now?
19:49 Tene Sure.
19:49 Tene I tried to work on making EHs not catch exceptions thrown from themselves, but I couldn't find any way to identify that case uniquely.
19:50 Tene What format do I need to use for PDD docs?  Does that just mean adding to docs/pdds/* ?
19:50 chromatic Yes.
19:50 Tene Okay.
19:50 Tene I'm done, then.
19:51 allison Tene: there isn't any way for an EH to identify an exception that was thrown within itself (that's not part of the spec)
19:51 allison Tene: and yes, new PDDs go in docs/pdds/draft
19:53 chromatic Are there other questions?
19:54 allison I'm closing the 0.9.0 milestone in Trac
19:54 allison do we want to move the remaining active tickets to 0.9.1?
19:54 chromatic Yes.
19:56 allison done
19:57 chromatic Anything else?
19:58 chromatic Okay, let's wrap up this week.  Close more tickets (especially mine).
19:58 Coke left #parrotsketch
19:58 allison Thanks, c!
19:58 chromatic left #parrotsketch
20:03 PacoLinux left #parrotsketch
20:06 masak left #parrotsketch
20:20 mberends left #parrotsketch
20:31 Whiteknight left #parrotsketch
20:40 diakopter left #parrotsketch
20:54 pmichaud left #parrotsketch
21:26 rurban left #parrotsketch
22:00 Whiteknight joined #parrotsketch
22:59 Whiteknight joined #parrotsketch

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