Last friday I pushed
https://github.com/Perl/perl5/pull/21107
which contains a number of updates to Porting/deparse-skips.txt to
reflect the current reality, and two fixes to Deparse.pm:
The first fix fixes a regression from 5.36, where the new OP_PADSV_STORE
op (which combines a pad op and a scalar assign op) wasn't deparsing
correctly in list context: for example,
$ ./perl -Ilib -MO=Deparse -e'my @a = (my $s = 1)'
my(@a) = local($s = 1);
The second is that although signatures are part of the 5.36 bundle,
Deparse will recognise, and correctly deparse, a signatured sub in the
scope of 'use feature "signatures", but not in the scope of 'use v5.36'.
I think this PR should be merged in time for 5.38.0.
--
In my day, we used to edit the inodes by hand. With magnets.
https://github.com/Perl/perl5/pull/21107
which contains a number of updates to Porting/deparse-skips.txt to
reflect the current reality, and two fixes to Deparse.pm:
The first fix fixes a regression from 5.36, where the new OP_PADSV_STORE
op (which combines a pad op and a scalar assign op) wasn't deparsing
correctly in list context: for example,
$ ./perl -Ilib -MO=Deparse -e'my @a = (my $s = 1)'
my(@a) = local($s = 1);
The second is that although signatures are part of the 5.36 bundle,
Deparse will recognise, and correctly deparse, a signatured sub in the
scope of 'use feature "signatures", but not in the scope of 'use v5.36'.
I think this PR should be merged in time for 5.38.0.
--
In my day, we used to edit the inodes by hand. With magnets.