From @avar

This causes t/operators/numify.t in the spec tests to fail. The only
grammar rule that parses numers with e is for decimals​:
 
  token dec_number {
  \d+[_\d+]* '.' \d+[_\d+]* [ <[Ee]> <[+\-]>? \d+ ]?
  {*}
  }

From zev-perl6@strangersgate.com

The fix is straightforward, but this change should also be made in
STD.pm. This fixes RT #​49910.

Zev

From zev-perl6@strangersgate.com

Index: languages/perl6/src/parser/grammar.pg
===================================================================
--- languages/perl6/src/parser/grammar.pg	(revision 25537)
+++ languages/perl6/src/parser/grammar.pg	(working copy)
@@ -532,7 +532,7 @@
         | \d+[_\d+]*
             {{ say "Leading 0 does not indicate octal in Perl 6" }}  # FIXME
         ]
-    | \d+[_\d+]*
+    | \d+[_\d+]* [ <[Ee]> <[+\-]>? \d+ ]?
     ]
     {*}
 }

From zev-perl6@strangersgate.com

There is a patch in RT #​50550.

Zev

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

From @jquelin

On 08/02/05 15​:21 -0800, Zev Benjamin via RT wrote​:

There is a patch in RT #​50550.

why on earth did i get a copy of your answer?
seems that it forged a from​:
  From perl6-compiler-return-1739-jquelin=gmail.com@​perl.org

something's fishy in there...

jérôme
--
jquelin@​gmail.com

From @coke

On Feb 5, 2008 4​:41 PM, via RT Zev Benjamin
<parrotbug-followup@​parrotcode.org> wrote​:

# New Ticket Created by Zev Benjamin
# Please include the string​: [perl #​50550]
# in the subject line of all future correspondence about this issue.
# <URL​: http://rt.perl.org/rt3/Ticket/Display.html?id=50550 >

The fix is straightforward, but this change should also be made in
STD.pm. This fixes RT #​49910.

Zev

FYI, I've merged the two tickets.

--
Will "Coke" Coleda

From anjohnson@iee.org

On Tue Feb 05 13​:41​:02 2008, zev-perl6@​strangersgate.com wrote​:

The fix is straightforward, but this change should also be made in
STD.pm. This fixes RT #​49910.

This patch would make 3e-4 a valid integer literal, even though it's not
an integer. Now 300e-2 *is* a whole number, but I'm not sure it should
be accepted as an integer literal.

This is part of the question of what distinguishes an integer from a
floating point number; it can't just be the presence of a decimal point
as the above examples show, but S02 isn't specific. The current version
of dec_number in grammer.pg is too simplistic, it doesn't allow
exponents without both a decimal and at least one fractional digit.

Here are a few examples that need answers from the language lawyers​:
  .1 Legal, or is a leading 0 required?
  1e1 Integer or Float?
  1e-1 Integer or Float?
  10e-1 Integer or Float?
  .1e1 If legal, Integer or Float?
  1.e1 Legal?

- Andrew

From @pmichaud

On Fri Mar 07 23​:08​:14 2008, songmaster wrote​:

On Tue Feb 05 13​:41​:02 2008, zev-perl6@​strangersgate.com wrote​:

The fix is straightforward, but this change should also be made in
STD.pm. This fixes RT #​49910.

This patch would make 3e-4 a valid integer literal, even though it's not
an integer. Now 300e-2 *is* a whole number, but I'm not sure it should
be accepted as an integer literal.

STD.pm has since been updated to have the correct parsing -- the answer
is that an integer literal never has a dot or 'e'.

In r26812 I've just updated rakudo to match STD.pm, so this ticket can
be closed.

This is part of the question of what distinguishes an integer from a
floating point number; it can't just be the presence of a decimal point
as the above examples show, but S02 isn't specific. [...]
Here are a few examples that need answers from the language lawyers​:
.1 Legal, or is a leading 0 required?
1e1 Integer or Float?
1e-1 Integer or Float?
10e-1 Integer or Float?
.1e1 If legal, Integer or Float?
1.e1 Legal?

Just for completeness​:

  .1 Not a legal number (leading digit required)
  1e1 Float
  1e-1 Float
  10e-1 Float
  .1e1 Not a legal number (leading digit required)
  1.e1 Not a legal number (digits required after dot)

Closing ticket, thanks!

Pm

@pmichaud - Status changed from 'open' to 'resolved'