| grizzlysmit | hi guys I have the following problem on Ubuntu 26.04 $ fez upload --file=archive/2026-08-04T09:47:15.576225+10:00-GUI::Editors-0.1.17.tar.gz =<< FATAL: Something went wrong: Dist exists. Do you need to run 'fez login' again? logging it in again achieves nothing even though the login succeeds any ideas | 00:02 | |
| Geth | rakudo: ugexe++ created pull request #6512: Make .split and .trans honor capture markers |
00:13 | |
| rakudo: ugexe++ created pull request #6513: Make MAIN recognize enum values that are not in scope |
01:15 | ||
| rakudo: ugexe++ created pull request #6514: RakuAST: Make native scalars usable in BEGIN-time code |
02:28 | ||
| rakudo: ugexe++ created pull request #6515: Do not apply unspace to whitespace inside regexes |
02:57 | ||
| rakudo: ugexe++ created pull request #6516: Defer array parameterization for generic types |
03:13 | ||
|
06:47
sp1983 joined
07:04
sp1983 left
08:11
sp1983 joined
08:14
sp1983 left
08:50
coleman left,
coleman joined
|
|||
| lizmat | grizzlysmit perhaps it's a shell issue, try quoteing the --file argument | 08:57 | |
| Geth | rakudo/main: e76d78c19b | (Nick Logan)++ (committed using GitHub Web editor) | src/core.c/CompUnit/Repository/Installation.rakumod Remove bin wrapper scripts on uninstall Previously uninstalling a distribution with a bin/ script would leave the generated wrapper scripts behind. generate-bin-wrapper names the wrappers after the script with its .raku extension stripped, but uninstall appended the wrapper extensions to the unstripped name, so for bin/synopsis.raku it tried to delete names like synopsis.raku-m ... (14 more lines) |
09:43 | |
| rakudo/main: 82d85db0ae | (Nick Logan)++ (committed using GitHub Web editor) | 2 files Defer array parameterization for generic types Previously parameterizing array checked the primspec of its type argument immediately, so a still-generic type parameter such as the T of `has array[T] $.a` inside a parametric role died at role compile time with "Can only parameterize array with a native type, not T". Blob[T] in the same position worked, because currying a parametric role ... (9 more lines) |
09:54 | ||
| rakudo/main: e4f01cf0ac | (Nick Logan)++ (committed using GitHub Web editor) | 3 files Make explicitly-manage return the managed string Previously explicitly-manage returned the internal CStr box produced by its final assignment. Passing that return value directly to a native sub skipped the ExplicitlyManagedString handling in the dispatcher, so the argument was marshalled as a plain Str and MoarVM freed the temporary buffer after the call. A C function that takes ownership of the string ... (7 more lines) |
09:58 | ||
|
09:58
sp1983 joined
|
|||
| rakudo/main: 78e581c4ed | (Nick Logan)++ (committed using GitHub Web editor) | 6 files Make .split and .trans honor capture markers Previously .split and .trans read the match extent directly from the $!from and $!pos attributes of each cursor. Capture markers are compiled as named captures called $!from and $!to, and their positions only get transferred out of the capture stack when a full Match object is reified with .MATCH. Since the split iterators never reify, the ... (25 more lines) |
10:01 | ||
| rakudo/main: 4055764419 | (Nick Logan)++ (committed using GitHub Web editor) | 5 files RakuAST: Make native scalars usable in BEGIN-time code Previously any BEGIN-time use of a native-typed scalar died while compiling the BEGIN with "No lexical found with name", e.g. BEGIN my int $foo; ... (27 more lines) |
10:05 | ||
| rakudo/main: 761b728c7f | (Nick Logan)++ (committed using GitHub Web editor) | 2 files Make coercions check subset refinements on every value The raku-coercion dispatcher chose between identity, a constraint mismatch error, and coercer selection while recording a dispatch program whose guards only cover the type and concreteness of the value. A subset target or constraint checks the value itself, so whichever outcome the recorded value produced was replayed for every later value ... (15 more lines) |
10:07 | ||
|
10:23
sp1983 left
|
|||
| Geth | rakudo/main: ee30050cae | (Nick Logan)++ | src/Raku/ast/var-lowering.rakumod RakuAST: lower anonymous variable declarations Previously the lex2local analysis skipped anonymous declarations: they returned early at registration, and the alphabetic name rule would have declined their empty desigilname anyway. An anonymous variable has no lookup path at all, which makes it the safest possible candidate, and the legacy optimizer already lowers them. ... (11 more lines) |
10:26 | |
| rakudo/main: da56524a2b | (Nick Logan)++ | 3 files RakuAST: run the lowering analysis for BEGIN-compiled routines Previously a routine invoked at BEGIN time was compiled by the code stub's compiler thunk during parsing, before the unit's optimize phase, so the lex2local analysis never saw it: the thunk emits and caches the QAST block, and the unit's own emission reuses that cache. Any such routine kept every container declaration as a clone-vivified lexical, ... (13 more lines) |
|||
| rakudo/main: 58c9e91fe5 | (Elizabeth Mattijsen)++ (committed using GitHub Web editor) | 3 files Merge pull request #6506 from ugexe/ugexe/rakuast-lex2local-anon-begin RakuAST: extend lexical lowering to anonymous variables and BEGIN-compiled routines |
|||
| rakudo/main: af1017f2d2 | (Nick Logan)++ (committed using GitHub Web editor) | 2 files RakuAST: stop thunking the right side of test assignments Previously $x //= rhs (and the ||= &&= or= and= forms) compiled to a METAOP_TEST_ASSIGN call with the right operand wrapped in a thunk. The thunk added a frame the source has no block for. A caller-walking lookup in the operand then started one frame too close, so in $cwd //= CALLERS::('$_') the lookup found the routine's own implicit ... (10 more lines) |
10:28 | ||
| rakudo/main: 177d09d2ca | (Nick Logan)++ (committed using GitHub Web editor) | 2 files RakuAST: skip the runtime type check on the implicit method invocant Previously a method composed from a role compiled a runtime type check of its invocant against the $?CLASS typevar, which each concretization resolves to its own class. A private method call is resolved in the caller's package, so calling one on another object doing the same role handed the caller's concretization an invocant of a different class and ... (11 more lines) |
10:29 | ||
|
10:47
sp1983 joined
|
|||