Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-03-14

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

All times shown according to UTC.

Time Nick Message
00:13 idiosyncrat joined #marpa
00:13 idiosyncrat choroba generously did a last test on the new Marpa::R2 release.
00:14 idiosyncrat And I have just uploaded Marpa-R2-4.000000
00:14 idiosyncrat It's the first full, indexed release in almost 2 years and ...
00:15 idiosyncrat since Marpa::R2 is stable (serious bug fix only and serious bugs in R2 are rarely found) ...
00:15 idiosyncrat this might even be the last R2 release.
00:19 idiosyncrat choroba: Thanks!
00:26 idiosyncrat All tests PASS on CPANtesters.
00:27 idiosyncrat This release is simply the last release with the release # changed and all was OK with it, so all *should* be OK with this one.
00:34 idiosyncrat Re the issue with R3 and Andreas's testbeds, here is one of his regression matrix suites: http://analysis.cpantesters.org/solved?distv=Marpa-R3-4.001_040
00:38 idiosyncrat I think the ones green boxes with high numbers indicate the most significance, and these point to Andreas "smoke testing" something, but I'm not 100% sure what.
00:38 idiosyncrat s/one green boxes/ones with green boxes/
00:44 idiosyncrat Back to the R2 release, 98 tests including linux & freebsd, all of which are OK.
00:44 idiosyncrat So looks like a successful launch.
01:06 idiosyncrat Oops! Found a problem in R2.  The two tests I added to the test suite require perl 5.10.1  -- Marpa::R2 had only required perl 5.10.0
01:07 idiosyncrat I'm not sure what to do about this.
02:49 ilbot3 joined #marpa
02:49 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
03:12 ronsavage joined #marpa
03:12 ronsavage JK: Re Marpa-R2-4.000000. All tests, yours and mine, pass for me.
05:40 ronsavage joined #marpa
05:41 ronsavage JK: Re http://irclog.perlgeek.de/marpa/2017-03-14#i_14258564. 5.10.0 was so buggy I'd adandon it myself (as I think I've said before), and make 5.10.1 the pre-req.
06:42 hobbs 5.6.0, 5.8.0, 5.10.0, none of them are, in retrospect, worth bothering with even if you support the rest of that series
07:08 sirdancealot joined #marpa
07:28 sirdancealot joined #marpa
08:22 idiosyncrat joined #marpa
08:22 idiosyncrat Good night!
09:51 sirdancealot joined #marpa
10:32 sirdancealot joined #marpa
11:52 sirdancealot joined #marpa
13:30 sirdancealot joined #marpa
14:25 sirdancealot joined #marpa
15:12 idiosyncrat joined #marpa
15:14 idiosyncrat hobbs, ronsavage: Thanks for the feedback.
15:15 idiosyncrat I've promised stability with R2, so I'm inclined to keep 5.10.0 support.
15:16 idiosyncrat I think what I'll do is a new developer's release cycle, and a new indexed release.  The fix is (should be) trivial.
15:35 idiosyncrat Allison Randal has tweeted a few Parrot lessons.
15:35 idiosyncrat https://twitter.com/allisonrandal/status/841668083039260673
15:37 idiosyncrat There have never been 5 Perl who could develop all of Marpa.  Sometimes the solution space is complicated and your choice is deal with it, or walk away from a great technology.
15:37 idiosyncrat https://twitter.com/allisonrandal/status/841668378616033280
15:39 idiosyncrat Marpa was considered dead on arrival for the first years of its life.  If what you do is genunely innovative, then you will be misunderstood, almost by definition.
15:40 idiosyncrat If I were to try to pinpoint a reason for the Parrot failure, it was failure to develop a constituency.
15:41 idiosyncrat "Build it and they will come" only happens in the movies.  You have to have a "they".
15:42 idiosyncrat I can sympathize with the Parrot quest, because Marpa also had a real, unique, technical vision -- it was not developed from the sales team's PowerPoint slides. :-)
15:43 idiosyncrat But at some point you have to turn away from the Holy Grail and serve a real constituency.
15:43 idiosyncrat Connecting this back to the previous discussion --
15:44 idiosyncrat Maintaining Marpa::R2 is much less fun and exciting than a lot of what I have planned for R3.
15:45 idiosyncrat And Marpa, in a sense, is all about the search for new stuff, which is happening in R3, and which is not happening in R2.
15:46 idiosyncrat But R2 has a real, live constituency and, while I hope some courageous souls are currently using R3, R3 is not rooted in a constituency.
15:48 idiosyncrat In early days I considered using Parrot in the way I'm using Lua now, but when I asked about structures, I learned that something so trivial and low-level was not gonna happen in
15:48 idiosyncrat Parrot.
15:49 idiosyncrat Which, to be sure, is fine, because at that point I didn't bring any constituency with me, and Allison could reasonable wonder if I ever would.
15:49 idiosyncrat s/reasonable/reasonably/
15:49 idiosyncrat But Parrot didn't find any other constituency either.
16:01 IRCFrEAK joined #marpa
16:08 sirdancealot joined #marpa
20:04 sirdancealot joined #marpa
21:08 sirdancealot joined #marpa
21:52 ronsavage joined #marpa
22:00 ronsavage JK: All good points, but seriously, how many of your R2 constituency are still using 5.10.0? I'd say 0 because if they are prepared to adopt Marpa they are certainly prepared to adopt a more recent Perl, right?
23:12 idiosyncrat joined #marpa
23:12 idiosyncrat ronsavage: my musings were more prompted by Allison's tweets than the R2 5.10.0 matter but, yes, the are relevant.
23:13 idiosyncrat I'm going to announce R2 4.000000 on the Google list, and at the same time, I'll bring up the 5.10.0 matter.
23:13 idiosyncrat It'll be interesting if anyone reading that list cares about 5.10.0.
23:14 idiosyncrat But the Parrot matter is relevant -- as a constituency, I was (in a very polite, friendly and even helpful way let me add) told I was not a Parrot priority.
23:19 idiosyncrat_ joined #marpa
23:20 idiosyncrat_ But in the end Parrot never really (IIRC) made any actual constituency a priority.
23:20 idiosyncrat_ So, should I kick 5.10.0 (and its users if there are any) to the curb?
23:20 idiosyncrat_ Maybe.
23:20 idiosyncrat_ But I'm in no hurry to do so.
23:22 idiosyncrat_ Btw, I do note that for years it's been hard to make a new installation of 5.10.0 -- so if anybody is using 5.10.0, I expect they are are a legacy system, one packaged during the short window after 5.10.0 and 5.10.1.
23:24 idiosyncrat_ But I've worked in shops where you can only use the vendor-shipped Perl, and where OS's are not updated often.  For example, I've had root access to systems running payroll for a Fortune 500 company.
23:25 idiosyncrat_ On those boxes, our instructions were to upgrade nothing without an onerous series of approvals which, by intent, made all but the most urgent changes impossible.
23:27 idiosyncrat_ I was required, not just to obey this restriction, but to see that others did, something I was very happy to do -- I liked getting my paycheck on time as much as anybody. :-)
23:28 idiosyncrat_ Anyway, it's quite possible there are some R2 users out there forced to run 5.10.0, and for whom the latest release's regression is a serious annoyance.

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