|
🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm Set by lizmat on 8 June 2022. |
|||
|
00:21
finanalyst left
|
|||
| [Coke] | lizmat: desired URL to replace the issue in github.com/Raku/raku.org/issues/229 ? | 00:58 | |
| github.com/Raku/doc/pull/4513 - JJ says Test is Rakudo, not Raku, but that's not right, is it? | 02:05 | ||
|
05:24
guifa left
|
|||
| ab5tract | It's certainly a fuzzy distinction. I don't see how one could implement a new Raku compiler/interpreter/whatever-er that doesn't include a Test that conforms to the existing API | 08:41 | |
| that said, according to a strict viewpoint, Test is indeed Rakudo and not Raku | 08:44 | ||
| lizmat | I'd say that Test is Raku as it is used in roast, and that specifies Raku ? | 09:25 | |
|
10:11
sena_kun joined
|
|||
| Geth | roast: bd38ef0bfd | (Elizabeth Mattijsen)++ | S03-metaops/reduce.t Fix tests wrt to [>>+<<] This has actually little todo with #1705, but needs a fix in the handling of the HYPER sub in rakudo. Until then, change the test to a state where they can be TODOed |
10:42 | |
| roast: 7a0b23bea1 | (Elizabeth Mattijsen)++ | S32-list/reduce.t Adapt test to no-automatic handling 1 arg case Since the magic of handling the 1-arg case for reduction with left associativity will be removed as it is *too* magic (and the zero argument case was **not** magically handled). |
10:47 | ||
| rakudo/main: 9bcb01e966 | (Elizabeth Mattijsen)++ | src/core.c/metaops.rakumod Remove the magic handling of left one-arg reduction As this is considered to be the conclusion of #1705: - there is magic handling for 1-arg with left-associativity - there is NO magic handling for 0-arg with left-associativity - there is NO magic handling for 1 and 0-arg with any other associativity This broke a few spectests, which have been fixed with - github.com/Raku/roast/commit/bd38ef0bfd - github.com/Raku/roast/commit/7a0b23bea1 |
10:55 | ||
|
11:10
finanalyst joined
11:12
sena_kun left
11:34
finanalyst left
|
|||
| Geth | rakudo/main: b2198d9e18 | (Elizabeth Mattijsen)++ | src/core.c/metaops.rakumod Make reduction of hyperop work on 0 / 1 arguments This was first noted in github.com/Raku/old-issue-tracker/issues/3475 in 2014. Its partial fix in the end caused inconsistency in handling ... (7 more lines) |
11:59 | |
| roast: d34202519b | (Elizabeth Mattijsen)++ | S03-metaops/reduce.t unTODO now passing tests Becaise of github.com/rakudo/rakudo/commit/b2198d9e18 And make them a little bit more specific. |
12:11 | ||
| roast: 75eb5f103b | (Elizabeth Mattijsen)++ | S03-metaops/reduce.t Replace TODOed test (from 2008) with a test that tests for the currently expected semantics of [,], namely as a producer of a List |
12:19 | ||
| lizmat | one fewer test file that needs fudging :-) | ||
|
12:26
ilogger2 left,
ilogger2_ joined
13:07
finanalyst joined
|
|||
| lizmat | m: dd [and] | 13:11 | |
| camelia | Bool::True | ||
| lizmat | m: dd [!and] | ||
| camelia | Bool::True | ||
| lizmat | that feels wrong | ||
| ab5tract | lizmat: R#5690 | 13:36 | |
| linkable6 | R#5690 [open]: github.com/rakudo/rakudo/pull/5690 Address long-standing discrepancy for != and METAOP_NEGATE | ||
| ab5tract | I think it's the same underlying issue? | ||
| lizmat | I was about to commit a fix to METAOP_NEGATE | 13:37 | |
| ab5tract | cool :) | 13:38 | |
| lizmat | but reading that now, feels not good | 13:40 | |
| sub METAOP_NEGATE(\op) is implementation-detail { | |||
| - -> |c { c.elems > 1 ?? !op.(|c) !! True } | |||
| + -> |c { !op.(|c) } | |||
| or at least incomplete | 13:41 | ||
| m: say &infix:<==>() | |||
| camelia | True | ||
| lizmat | m: say &infix:<!=>() | ||
| camelia | True | ||
| lizmat | m: say &infix:<!==>() | ||
| camelia | True | ||
| lizmat | feels the latter two should both be False ? | ||
| or am I missing something | |||
| it should then also affect: | 13:42 | ||
| multi sub infix:<!=>( --> True) { } | |||
| multi sub infix:<!=>($ --> True) { } | |||
| afk for a few hours& | 13:55 | ||
|
14:03
finanalyst left
|
|||
| ab5tract | lizmat: Ah, yes that PR is not polished yet. I've marked it now as a draft | 16:51 | |
| lizmat | after cycling, decided not to commit that fix to METAOP_NEGATE | 16:52 | |
| Geth | rakudo/main: 4f8cfb1268 | (Elizabeth Mattijsen)++ | src/core.c/Hash.rakumod Give Hash.clone the same semantics as Array.clone As in: re-containerize any containerized values when cloning, but don't go any deeper than the values in the instance itself. In respnse to #1997 |
18:48 | |
|
19:24
jdv left
|
|||
| Geth | rakudo/main: 72e2dcd13d | (Elizabeth Mattijsen)++ | src/core.c/Hash/Object.rakumod Give Hash::Object.clone the same semantics as Array.clone github.com/rakudo/rakudo/commit/4f8cfb1268 gave Hash.clone the same semantics as Array.clone. But that broke .clone on object hashes. This commit fixes that. In response to #1997 |
19:30 | |
|
19:36
jdv joined
|
|||
| Geth | roast: 6ea6294fa4 | (Elizabeth Mattijsen)++ | 2 files Add tests for #1997 As fixed by: - github.com/rakudo/rakudo/commit/4f8cfb1268 - github.com/rakudo/rakudo/commit/72e2dcd13d |
19:41 | |
|
19:41
sena_kun joined
22:29
guifa joined
22:42
sena_kun left
23:12
sena_kun joined
23:15
sena_kun left
|
|||