Make require use CompUnit interface for importing #694

LLFourn · 2016-01-27T10:23:26Z

Uses the CompUnit object to do the importing of symbols and merging of globalish in &REQUIRE_IMPORT instead of some haphazard CALLER::OUR:: thing. This makes things a bit less brittle and fixes a few bugs.

The following now works:
my $name = "/lib/Module.pm"; require $name <sym>;
require Module <sym>; # where <sym> isn't inside a module/package

bonus: no more "useless use of blah in sink context"
Fixes:

note the first chunk removed in actions is just because it's dead code that can't be reached.

LLFourn · 2016-01-27T10:25:35Z

tests: Raku/roast#100

jonathanstowe · 2016-01-27T11:04:37Z

👍

LLFourn · 2016-01-30T07:19:12Z

src/Perl6/Actions.nqp

@@ -1821,32 +1807,20 @@ class Perl6::Actions is HLL::Actions does STDActions {
                ),
            );
        }
-        $past.push(QAST::Op.new(


Merging of GLOBALish moved to REQUIRE_IMPORT

LLFourn · 2016-01-30T07:33:30Z

I've cleared up this commit a bit and added a bunch of github comments to make things a bit clearer.

salortiz · 2016-02-13T06:21:12Z

👍

lizmat · 2016-02-13T13:28:42Z

Seems this patch causes some breakage for me, e.g.:

$ perl6 t/spec/S01-perl-5-integration/exception_handling.t
1..3
===SORRY!===
ctxlexpad needs an MVMContext

Looks like you planned 3 tests, but ran 0

Complete list:
t/spec/S01-perl-5-integration/return.rakudo.moar (Wstat: 65280 Tests: 0 Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan. You planned 2 tests but ran 0.
t/spec/S01-perl-5-integration/class.rakudo.moar (Wstat: 256 Tests: 1 Failed: 1)
Failed test: 1
Non-zero exit status: 1
Parse errors: Bad plan. You planned 3 tests but ran 1.
t/spec/S01-perl-5-integration/exception_handling.t (Wstat: 65280 Tests: 0 Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan. You planned 3 tests but ran 0.
t/spec/S01-perl-5-integration/import.rakudo.moar (Wstat: 256 Tests: 1 Failed: 1)
Failed test: 1
Non-zero exit status: 1
Parse errors: Bad plan. You planned 3 tests but ran 1.

Is this a known issue with this patch?

LLFourn · 2016-02-13T19:11:39Z

Oh I never tested p5 integration. I'll take a look thanks.
On Sun, 14 Feb 2016 at 12:28 AM, lizmat notifications@github.com wrote:

Seems this patch causes some breakage for me, e.g.:

$ perl6 t/spec/S01-perl-5-integration/exception_handling.t
1..3
===SORRY!===
ctxlexpad needs an MVMContext
Looks like you planned 3 tests, but ran 0

Complete list:
t/spec/S01-perl-5-integration/return.rakudo.moar (Wstat: 65280 Tests: 0
Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan. You planned 2 tests but ran 0.
t/spec/S01-perl-5-integration/class.rakudo.moar (Wstat: 256 Tests: 1
Failed: 1)
Failed test: 1
Non-zero exit status: 1
Parse errors: Bad plan. You planned 3 tests but ran 1.
t/spec/S01-perl-5-integration/exception_handling.t (Wstat: 65280 Tests: 0
Failed: 0)
Non-zero exit status: 255
Parse errors: Bad plan. You planned 3 tests but ran 0.
t/spec/S01-perl-5-integration/import.rakudo.moar (Wstat: 256 Tests: 1
Failed: 1)
Failed test: 1
Non-zero exit status: 1
Parse errors: Bad plan. You planned 3 tests but ran 1.

Is this a known issue with this patch?

—
Reply to this email directly or view it on GitHub
#694 (comment).

LLFourn · 2016-02-13T22:12:10Z

So the problem with this patch is that it uses UNIT:: and there is a bug in rakudo:

BEGIN { UNIT; Nil }

An exception occurred while evaluating a BEGIN
at -e:1
Exception details:
  ===SORRY!=== Error while compiling 
  ctxlexpad needs an MVMContext

And CU::R::Perl5 uses require which is naturally run at BEGIN time. I'll see what I can do.

The following now works: 1. require 'lib/Module.pm' <sym>; 2. require Module <sym>; # where <sym> isn't inside a module Also, no more erroneous "useless use of.." sink context when doing `require ::('Test') <&ok>`.

LLFourn · 2016-02-14T00:51:02Z

changed CALLER::UNIT::GLOBALish to CALLER::LEXICAL::GLOBALish:: to find the globalish to merge into. It seems to work at compile time and runtime (IP5 tests pass now).

Tests for require at compile time added in:
LLFourn/roast@2246a97

LLFourn mentioned this pull request Jan 27, 2016

Test for require file where symbol isn't in module Raku/roast#100

Closed

LLFourn force-pushed the fix-require branch 2 times, most recently from b87cec3 to d924ae2 Compare January 30, 2016 07:15

LLFourn reviewed Jan 30, 2016
View reviewed changes

Make require use CompUnit interface for importing

f565f80

The following now works: 1. require 'lib/Module.pm' <sym>; 2. require Module <sym>; # where <sym> isn't inside a module Also, no more erroneous "useless use of.." sink context when doing `require ::('Test') <&ok>`.

LLFourn force-pushed the fix-require branch from d924ae2 to f565f80 Compare February 14, 2016 00:41

lizmat merged commit f565f80 into rakudo:nom Feb 14, 2016

p6rt mentioned this pull request Jan 5, 2020

Importing sub using require causes Null PMC access Raku/old-issue-tracker#3162

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Make require use CompUnit interface for importing #694

Make require use CompUnit interface for importing #694

LLFourn commented Jan 27, 2016

LLFourn commented Jan 27, 2016

jonathanstowe commented Jan 27, 2016

LLFourn Jan 30, 2016

LLFourn commented Jan 30, 2016

salortiz commented Feb 13, 2016

lizmat commented Feb 13, 2016

LLFourn commented Feb 13, 2016

LLFourn commented Feb 13, 2016

LLFourn commented Feb 14, 2016

Make require use CompUnit interface for importing #694

Make require use CompUnit interface for importing #694

Conversation

LLFourn commented Jan 27, 2016

LLFourn commented Jan 27, 2016

jonathanstowe commented Jan 27, 2016

LLFourn Jan 30, 2016

Choose a reason for hiding this comment

LLFourn commented Jan 30, 2016

salortiz commented Feb 13, 2016

lizmat commented Feb 13, 2016

Looks like you planned 3 tests, but ran 0

LLFourn commented Feb 13, 2016

LLFourn commented Feb 13, 2016

LLFourn commented Feb 14, 2016