choroba has asked for the wisdom of the Perl Monks concerning the following question:

Dear Monks,

as some of you have noticed from my grumbling in the ChatterBox, I'm getting weird CPAN Testers' reports. With Tux's help, I was able to identify the trigger: the problem happens in threaded 5.10.0 only.

The minimal example is here:

eval ' "x" =~ ?x? '

(in fact, eval '??' is even shorter, but probably harder to decipher, and doesn't show so clearly that it has to match to trigger the exception).

In threaded 5.10.0 (compiled with a few bugs in make test even with the latest Devel::PatchPerl applied), it throws

Modification of a read-only value attempted at ...

Interestingly, the exception goes away if I add reset after the match:

eval ' "x" =~ ?x?; reset '

It's clearly a Perl bug, and it's been fixed, as later versions don't show the same behaviour. Nevertheless, I'd be interested in any explanation why "match-only-once" in stringy eval caused this particular error (normal match or the same code outside of eval don't fail).

($q=q:Sq=~/;[c](.)(.)/;chr(-||-|5+lengthSq)`"S|oS2"`map{chr |+ord }map{substrSq`S_+|`|}3E|-|`7**2-3:)=~y+S|`+$1,++print+eval$q,q,a,

Replies are listed 'Best First'.
Re: Threaded 5.10.0 + ?PATTERN? = Modification of a read-only value attempted
by haukex (Archbishop) on Jun 26, 2018 at 22:05 UTC

    Apparently, the patch you need to apply to fix this is 9cddf794fc5 (included in 5.10.1). (Update: Not really an answer to your "why" question, though, sorry.)

    $ git checkout -f v5.10.0 $ git clean -dxf $ patchperl $ git show 9cddf794fc5 | git apply $ sh Configure -des -Dusethreads $ make # etc.