Perl 6 - the future is here, just unevenly distributed

IRC log for #inline, 2014-11-25

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

All times shown according to UTC.

Time Nick Message
00:54 mohawk sample: http://www.cpantesters.org/cpan/report/51dba6c0-7326-11e4-9ce3-9a752656b80b
00:56 mohawk ingy, ^^
02:40 ingy hi mohawk
02:41 ingy I finally have a little time for open source today :)
02:41 ingy and in general :[
02:41 ingy davido___: ping ping ping
02:44 mohawk ingy, more fail: http://www.cpantesters.org/cpan/report/244e9c68-740b-11e4-8f8b-0cb810862755
02:53 ingy I wonder how these could possibly fail
02:59 ingy davido___: I'd like to give the blib stuff a go
02:59 ingy I'll work on it locally until you show up :)
03:01 mohawk ingy, i'd like to change the test condition to be something only you see else skip_all, please tell me what that would be
03:03 ingy mohawk: i wouldn't mind that if it failed for me, but it doesn't so I want to figure out what's going on
03:03 ingy otherwise we're just ignoring it
03:03 ingy I assume it passes for you?
03:03 mohawk it does
03:04 mohawk problem is it fails for x others
03:04 mohawk and it's kind of unreliable
03:04 ingy have you tried on a clean perl?
03:04 ingy with the minimal deps installed
03:05 mohawk no
03:05 ingy that's what I'd do next
03:05 mohawk and i don't know what you mean by that
03:05 ingy I can do it or you can
03:05 mohawk probably ideal if you do
03:05 ingy ok
03:05 mohawk i'm fighting an EUMM that's working even though it shouldn't
03:05 * mohawk weeps
03:06 ingy I'll see if I can in a bit
03:06 mohawk ok
03:06 mohawk i want you to in the meantime think of a condition that will be true only for you
03:06 mohawk so we can eliminate false runs
03:07 ingy I don't want to shut up the failure though
03:08 * mohawk sighs in a theatrical way
03:08 mohawk they're not informative right now
03:08 mohawk you wrote the code, and you're not seeing the fails
03:08 ingy nobody's dying
03:09 mohawk so you're saying you don't mind me switching them off?
03:09 mohawk great
03:09 ingy you are saying that you don't care if they fail?
03:09 ingy great
03:10 ingy just work on EUMM
03:10 ingy I'll get around to figuring it out
03:12 mohawk i'm saying i don't like getting mailed with fails that i didn't make
03:12 mohawk so i'm going to switch that shiz off
03:13 mohawk so unless the fail is arch-dependent, tell me
03:13 mohawk else i'll make it ingy-specific
03:13 mohawk if you don't tell me how to do that, it'll be "if 0"
03:13 ingy I want to figure out why the failures happen
03:14 ingy if I can't you can do that
03:14 mohawk thanks for that
03:14 mohawk i don't know enough about pegex to do so
03:14 mohawk if you can't tell from the test results, improve the tests
03:14 ingy I doubt it's pegex
03:14 ingy it's probably a missing dep
03:14 ingy but not sure
03:14 mohawk ok
03:15 mohawk i'll put out a fix in the meantime
03:15 ingy ?
03:16 ingy please don't release anything related to that
03:16 ingy ie don't turn off the failing test
03:17 mohawk got a better suggestion?
03:18 ingy just wait
03:18 mohawk for what, more spam based on the test that's failing?
03:18 mohawk no thanks
03:18 ingy ok then fix it in your email
03:20 ingy I said up front that I don't want to ignore this until I've had a chance to see if it is something simple.
03:21 mohawk ok
03:22 mohawk so what you're saying is you don't care if i get spammed by fail messages from your test that fails
03:22 mohawk remember, you can look up test results from latest I::C on cpantesters.org
03:22 mohawk that's public
03:23 ingy I have tests failing for years
03:23 mohawk you're my hero
03:24 ingy I have to run. ttul
03:24 mohawk i'm not prepared to accept being spammed with failing tests for years
03:24 mohawk i'm going to release a version with skip_all
03:24 ingy don't
03:25 mohawk got a better suggestion that fixes my email problem?
03:45 davido___ mohawk : https://prefs.cpantesters.org/cgi-bin/pages.cgi?act=author-distros  (must log in first)
03:47 davido___ And https://prefs.cpantesters.org/cgi-bin/pages.cgi?act=author-default but that's probably too blunt of a blade.
04:36 davido___ ingy hi.
04:36 ingy hi
04:38 davido___ it's not up yet?
04:38 ingy no
04:40 davido___ sec
04:42 davido___ k
04:45 willthechill joined #inline
04:45 d^i joined #inline
04:46 davido___ in
05:08 davido___ sec... checking something.
05:21 davido__ joined #inline
09:04 sivoais_ joined #inline
16:39 ingy hi davido__
16:40 davido__ hi
16:41 ingy mohawk: I could not replicate the pegex failure on a clean perl, so go ahead and disable them unless PERL_INLINE_DEVELOPER_TEST is set
16:42 davido__ What's wrong with setting "TODO" on them? We would still *see* the reports, but wouldn't get emailed failures.
16:42 ingy what happens when a TODO passes?
16:42 ingy I thought that was a real error
16:42 davido__ Passing todo doesn't cause a fail, just a message, I think.... let me verify.
16:43 ingy but if not, then yes, that's better
16:43 davido__ It'll take me a couple minutes....
16:43 ingy np
16:43 ingy i'll set up the pairup
16:44 davido__ because it's not a test::More decision, it's the TAP harness I think.
16:44 ingy the tests are all passing on travis too
16:45 ingy I need to make the fails noisier somehow
16:46 davido__ I verified.  TODO passing is ok, you just get a message in prove -v mode.
16:46 davido__ but still counts as all tests successful.
16:47 davido__ failing tests have 'false' return values, so you can detect that within the test script and conditionally spew 'diag' or 'note' information.
16:48 davido__ oh, hold on....... thought of something....
16:51 davido__ All of the failures come from 'kent' and 'njh'....
16:52 ingy ?
16:52 davido__ testers
16:52 davido__ kent is running blead, so we can discount that one.
16:52 davido__ njh is running everything from 5.8.8 up to 5.18.3....
16:52 ingy what about andk
16:53 davido__ http://analysis.cpantesters.org/solved?distv=Inline-C-0.66
16:54 davido__ look at the "Regression#2" box.  A .942 relevancy.  Shows the mode of failure.
16:54 davido__ 99.72% of the time, that test ends with "eq_1" instead of "const"
16:55 ingy huh
16:55 ingy so you think njh has a bad setup?
16:56 davido__ It doesn't look good.  does he have any passes?
16:56 davido__ he has one pass.
16:56 davido__ but a totally different environment.
16:57 davido__ See Regression#3.  item 9 is totally different from 10-20.
16:57 davido__ what we probably should be trying to figure out is how, besides PERLBREW_PATH, he differs from everyone else.
16:58 ingy I guess
16:58 davido__ but I think Regression#2 is interesting.
16:58 davido__ has a much higher statistical relevance.
16:58 ingy I'd like to work on the grant now. but maybe you and mohawk can discuss this later?
16:58 davido__ sure.
16:58 ingy :)
16:59 ingy cpantesters++ # pretty cool stuff
17:00 davido__ is the pairup open?
17:00 davido__ oh, in now.  it rejected me a moment ago.
17:00 davido__ ...and... out again.
17:01 ingy it's messed up
17:01 ingy hmm
17:01 ingy what happened in there?
17:02 davido__ I ctrl-a d last night, so it wasn't me.
17:03 ingy fixed it. at least for me
17:03 GitHub188 [inline-c-pm] mohawk2 pushed 2 new commits to master: http://git.io/h4VchA
17:03 GitHub188 inline-c-pm/master 9bd6b3e Ed J: Skip test/pegex-parser.t unless env PERL_INLINE_DEVELOPER_TEST set
17:03 GitHub188 inline-c-pm/master b227148 Ed J: CPAN Release 0.67...
17:04 GitHub172 [inline-c-pm] mohawk2 tagged 0.67 at master: http://git.io/120aWg
17:04 mohawk ingy, by your command
17:04 d^i joined #inline
17:04 davido__ 0.67 is out.
17:04 ingy mohawk: ?
17:05 mohawk original battlestar galactica reference
17:05 mohawk i did the change
17:05 ingy mohawk: fine.
17:06 davido__ mohawk: Did you know you can turn off FAIL reports emailed to you? (You probably do know, but if not, thought I'd mention)
17:06 mohawk i didn't know
17:07 davido__ i'll show you the link... just a moment.
17:07 ingy also what abut the todo discussion of the last 10 mins?
17:07 mohawk but unless it's far more fine-grained than i think, it's not really a solution to the problem just addressed ;-)
17:07 davido__ it's by distribution.
17:07 mohawk davido__, ok
17:07 mohawk ingy, let me scrollback - what's the question?
17:07 davido__ https://prefs.cpantesters.org/
17:08 davido__ and after you log in with pause credentials: https://prefs.cpantesters.org/cgi-bin/pages.cgi?act=author-distros
17:08 ingy everything we just discussed after I said you could release it
17:08 mohawk seen
17:08 davido__ I checked last night, and it has to be the person who made the release to set it.
17:09 mohawk yes, TODO passing just reports under -v as davido__ said
17:09 davido__ yes, but a fail can be used to trigger an unless( ok 0 ) { diag }
17:10 mohawk true
17:10 davido__ Oh, the final thought mohawk is looking at http://analysis.cpantesters.org/solved?distv=Inline-C-0.66
17:10 mohawk the way i see it, pegex is pure perl without system interactions, so running it on all the systems doesn't seem to add much
17:11 davido__ that makes sense, except for the person for whom it inexplicably fails to parse correctly.
17:11 davido__ in this case, njh
17:12 mohawk davido__, i don't think you can say it failed to parse correctly
17:12 mohawk what you can say here is the $::data var was unset
17:13 davido__ It looks (Regression#2) like it failed to get "const"
17:13 davido__ oh, no, i'm misreading that.
17:13 davido__ apologies. :)
17:14 davido__ anyway, i've got to focus or I'll get myself into a time crunch....
17:14 davido__ but at some point i'm happy to work on the tests more.
17:14 mohawk great
17:15 mohawk let's talk when you're past this crunch
17:15 davido__ Yeah. driving to San Francisco tonight for a few days.
17:15 davido__ but bringing laptop++ :)
17:16 mohawk (david's laptop)++ ;-)
17:19 davido__ do you need me to go larger ingy?
17:19 ingy no
17:21 davido__ will we keep the perl-inline-module generate command available?
17:22 ingy not sure
17:22 ingy it will change a bit if we do
17:22 ingy brb
17:28 ingy back
17:28 davido__ k
17:36 davido__ wow, bird just flew hard into my window.
17:37 mohawk ha ha
17:37 mohawk "thap"
17:45 ingy davido__: so 3 incantations that will auto build stubs if they don't exist
17:45 davido__ yes.
17:45 ingy by putting a code ref in @INC
17:45 davido__ that's a cool trick (code ref in @inc)
17:45 ingy we'll put at the end
17:46 ingy so if something goes out of date, we detect that separately
17:48 ingy davido__: let's use IO::All for I/O in Inline::Module
17:48 davido__ :)
17:48 ingy since it is author side only
17:49 davido__ Yeah, that's fine.  We don't care too much about author deps.
18:05 davido__ brb
18:06 ingy davido__: http://www.perlmonks.org/?node_id=357925
18:06 ingy it's you!
18:06 ingy 10 years ago
18:06 davido__ yeah, that was in my early perl days.
18:07 davido__ well, early revisiting perl.  I became interested in it in the late 90's, but when I joined PerlMonks in late 2003, that's when I got into it more.
18:08 davido__ ok, 5 mins.
18:10 ingy there's a problem here
18:11 ingy we need to figure out when the thing we are making is really an ::Inline or just some other missinbg thing
18:16 davido__ back
18:21 ingy brb
18:22 davido__ k
18:22 ingy getting closre
18:30 ingy back
18:30 davido__ k.  i'm running short now, but will be back after 12:40.
18:30 ingy ok
18:30 davido__ i have a few more mins....
18:33 davido__ ok.... i'll be back in just over an hour.
18:34 ingy ok
19:43 davido__ back
19:53 ingy the inline/stub part of postamble may go away
19:54 davido__ so simplifying the postamble section too.
19:54 ingy yeah
19:54 davido__ even better.
19:54 ingy I think the whole thing might be getting simpler
19:54 davido__ sometimes that's how it is.
19:54 ingy but I not 100% yet
19:57 davido__ that's good.
19:58 davido__ can that be -MInline::Module=lib also then?
19:59 ingy yes
19:59 davido__ good
19:59 davido__ export?
19:59 davido__ (in the 2nd example
20:01 davido__ is a space necessary between e and 1?
20:01 ingy no
20:01 davido__ :)
20:02 davido__ # or 'lib'
20:02 ingy good enough for now
20:02 ingy will likely change
20:03 davido__ cool though.
20:03 ingy removes a lot of code
20:03 ingy and bin/ goes away
20:04 davido__ yeah.  maintain functionality, remove code, big win.
20:04 ingy I think the bin/ would just be confusing
20:04 ingy no real win
20:04 davido__ I think actually this way the author should have a better feel for what is happening.
20:05 davido__ i mean it's not all *that* different, but it just feels more familiar than running some bin/
20:07 GitHub1 [inline-module-pm] ingydotnet deleted postamble at ec556f0: https://github.com/ingydotnet/inline-module-pm/commit/ec556f0
20:09 GitHub110 [inline-module-pm] ingydotnet created autostub (+1 new commit): http://git.io/GQ0_iA
20:09 GitHub110 inline-module-pm/autostub 373084b Ingy döt Net: Half done work on autostub branch
20:09 davido__ good.
20:09 davido__ see you tomorrow.
20:09 GitHub144 [acme-math-xs-pm] ingydotnet pushed 1 new commit to eumm: http://git.io/3f3AcQ
20:09 GitHub144 acme-math-xs-pm/eumm 7d8f80b Ingy döt Net: Deleting the generated code....
20:10 ingy ok
20:15 travis-ci ingydotnet/inline-module-pm#21 (autostub - 373084b : Ingy döt Net): The build passed.
20:15 travis-ci Change view : https://github.com/ingydotnet/inline-module-pm/commit/373084ba64db
20:15 travis-ci Build details : http://travis-ci.org/ingydotnet/inline-module-pm/builds/42121356
23:09 davido__ joined #inline

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