From @drforr

Inserting a newline before '=>' seems to break context handling. The
sample code uses 0x20 space, not \t, but the problem appears to be with
\h in the parser, not the tabbing.

--cut here--
{
complementary
  => 1 # Newline and tab here
  ?? True
  !! False
}
--cut here--

--cut here--
===SORRY!=== Error while compiling /home/jgoff/bug.t
Preceding context expects a term, but found infix => instead
at /home/jgoff/bug.t​:3
------> =>⏏ 1
--cut here--

From @labster

On Mon May 04 13​:24​:19 2015, drforr@​pobox.com wrote​:

Inserting a newline before '=>' seems to break context handling. The
sample code uses 0x20 space, not \t, but the problem appears to be with
\h in the parser, not the tabbing.

--cut here--
{
[4​:15pm] TimToady​: the \h* is intentional for LTM purposes, and it can't be liberalized to \s* without potentially breaking heredocs
[4​:16pm] TimToady​: it also reduces the LTM lookahead greatly on identifiers without a subsequent =>
[4​:17pm] TimToady​: if you let it go past \n then it probably has to scan a whole bunch of spaces on the next line on the offchance someone wanted to put a newline before the =>
[4​:17pm] TimToady​: and you'll note that this is the first time in a decade that someone has actually attempted it :)
[4​:18pm] TimToady​: so it's gonna stay \h*

It sounds like this is intentional behavior.

complementary
=> 1 # Newline and tab here
?? True
!! False
}
--cut here--

--cut here--
===SORRY!=== Error while compiling /home/jgoff/bug.t
Preceding context expects a term, but found infix => instead
at /home/jgoff/bug.t​:3
------> =>⏏ 1
--cut here--

From @masak

labster (>)​:

It sounds like this is intentional behavior.

Indeed. Rejecting ticket.

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

@masak - Status changed from 'open' to 'rejected'