Branch: refs/heads/blead
Home: https://github.com/Perl/perl5
Commit: bea115698e41f6acdaa450b343adbf1082c2e99e
https://github.com/Perl/perl5/commit/bea115698e41f6acdaa450b343adbf1082c2e99e
Author: Richard Leach <richardle...@users.noreply.github.com>
Date: 2023-09-07 (Thu, 07 Sep 2023)
Changed paths:
M pp.c
Log Message:
-----------
S_negate_string - allow for TARG==sv
Adds a case in `S_negate_string` to allow for TARG==sv, which then
means that the `pp_negate` and `pp_i_negate` OPs can support the
TARGMY optimization
Commit: 359c2cd723a18f596af21897fb196d657d81752e
https://github.com/Perl/perl5/commit/359c2cd723a18f596af21897fb196d657d81752e
Author: Richard Leach <richardle...@users.noreply.github.com>
Date: 2023-09-07 (Thu, 07 Sep 2023)
Changed paths:
M lib/B/Op_private.pm
M opcode.h
M pp.c
M regen/op_private
M regen/opcodes
Log Message:
-----------
pp_(i_)negate: add the TARGMY optimization
This mostly means using the existing machinery in regen/, but also not
bothering to replace *PL_stack_sp when (targ == sv), as that would
effectively be a busywork no-op.
Commit: c904517849a782954f8d958163a18f6cc192a485
https://github.com/Perl/perl5/commit/c904517849a782954f8d958163a18f6cc192a485
Author: Richard Leach <richardle...@users.noreply.github.com>
Date: 2023-09-07 (Thu, 07 Sep 2023)
Changed paths:
M pp.c
Log Message:
-----------
pp_(i_)negate: Don't needlessly call S_negate_string
Arguably it will be more common for users to want to negate an IV or NV,
rather than a PV. Prior to this commit, `S_negate_string` was always called
at the start of the OP and, if the target wasn't `SvPOKp`, that function
just immediately returned. This commit moves the `SvPOKp` check into the
`pp_` functions to avoid needless function calls.
Compare: https://github.com/Perl/perl5/compare/5e28db930c33...c904517849a7
