From allen@grumman.com

  $ perl -we '$a = 0e0'
  Bareword found where operator expected at -e line 1, near "0e0"
  (Missing operator before e0?)
  syntax error at -e line 1, next token ???
  Execution of -e aborted due to compilation errors.

Both with perl5.004_04 and perl5.005_60. Perhaps it is trying to parse it
as an octal number, but the error message does not seem to indicate that.

John.

From [Unknown Contact. See original ticket]

This patch makes 0e0 scan as 'decimal' (thus as 0.0e0 would)
rather than as octal/hex/binary. Your line numbers WILL vary, as
this comes from an already-patched toke.c, but it will apply
properly despite the warnings from patch.

/spider

--- toke.c.prev	Thu Aug 19 16:22:48 1999
+++ toke.c	Fri Aug 20 11:45:14 1999
@@ -6398,7 +6401,7 @@ Perl_scan_num(pTHX_ char *start)
 		s += 2;
 	    }
 	    /* check for a decimal in disguise */
-	    else if (s[1] == '.')
+	    else if (strchr(".Ee", s[1]))
 		goto decimal;
 	    /* so it must be octal */
 	    else

From [Unknown Contact. See original ticket]

On Fri, 20 Aug 1999, Spider Boardman wrote​:

This patch makes 0e0 scan as 'decimal' (thus as 0.0e0 would)
rather than as octal/hex/binary. Your line numbers WILL vary, as
this comes from an already-patched toke.c, but it will apply
properly despite the warnings from patch.

[patch deleted]

Wow, that was fast. Talk about service! Thanks.

John.

From [Unknown Contact. See original ticket]

On Fri, 20 Aug 1999, Spider Boardman wrote​:

This patch makes 0e0 scan as 'decimal' (thus as 0.0e0 would)
rather than as octal/hex/binary. Your line numbers WILL vary, as
this comes from an already-patched toke.c, but it will apply
properly despite the warnings from patch.

/spider

--- toke.c.prev Thu Aug 19 16​:22​:48 1999
+++ toke.c Fri Aug 20 11​:45​:14 1999
@​@​ -6398,7 +6401,7 @​@​ Perl_scan_num(pTHX_ char *start)
s += 2;
}
/* check for a decimal in disguise */
- else if (s[1] == '.')
+ else if (strchr(".Ee", s[1]))
goto decimal;
/* so it must be octal */
else

Thanks.

(Though perhaps we should use (s[1] == '.' || s[1] == 'e' || s[1] == 'E')
for speed, instead of calling strchr?)

John.

From [Unknown Contact. See original ticket]

On Fri, 20 Aug 1999 12​:06​:46 -0400 (EDT), "John L. Allen" wrote (in part)​:

John> (Though perhaps we should use (s[1] == '.' || s[1] == 'e'
John> || s[1] == 'E') for speed, instead of calling strchr?)

With the compiler I'm using, it's not measurable (strchr is
inlined). You have raised a valid point, though. I'm inclined
to leave it up to the pumpkings for the various release streams
to decide. ;=}

/spider