From @zoffixznet

IRC​: https://irclog.perlgeek.de/perl6-dev/2017-10-14#i_15302654

1) Spotted that Str.trans grabs caller's $/ with $/ := nqp​::getlexcaller('$/')
  The same is not done for Cool.trans candidate, so presumably that makes whatever
  feature needs caller's $/ broken for Cool, while working for Str. However...

2) Why does Str.trans grabs caller's $/ in the first place? The commit that added
  it seems to claim to fix something that isn't meant to work as tr/// replaces per
  char and doesn't use special regex syntax (basically, author of the commit seems
  to be confusing it with s///).

  So I tried removing all of the $/ business from Str.trans, but that caused 2 tests
  to fail, one of which does something like this​:

  <Zoffix__> m​: say '"foo & bar"'.trans(/(' '+)/ => {' ' ~ (' ' x ($0.chars - 1))}, /(\W)/ => {"&#{ord($0)};"}); say $/
  <camelia> rakudo-moar de2b9ff72​: OUTPUT​: «&#​34;foo  &#​38;   bar&#​34;␤Nil␤»

  As can be seen from the eval above, while the $/ is used internally for trans stuff, the trans's caller's $/ is not
  set to anything. So unless I'm misunderstanding something, at least the part in Str.trans's proto that grabs caller's $/
  and possibly some other occurrences inside Str.trance need to be removed as they're useless wasted cycles.

  However, if there *is* some reason for why Str.trans proto (and routine) grab caller's $/, the same thing needs to be added
  to Cool.trans, to make it work right too.

[1] rakudo/rakudo@ec79843
[2] https://gist.github.com/zoffixznet/56ee3544201dbd18a419514ed70a782e

From @geekosaur

On Sat, Oct 14, 2017 at 1​:15 PM, Zoffix Znet <perl6-bugs-followup@​perl.org>
wrote​:

2) Why does Str.trans grabs caller's $/ in the first place? The commit
that added
it seems to claim to fix something that isn't meant to work as tr///
replaces per
char and doesn't use special regex syntax (basically, author of the
commit seems
to be confusing it with s///).

Are you sure this isn't just a copypasto?

--
brandon s allbery kf8nh sine nomine associates
allbery.b@​gmail.com ballbery@​sinenomine.net
unix, openafs, kerberos, infrastructure, xmonad http://sinenomine.net

The RT System itself - Status changed from 'new' to 'open'