more double negation fixes

[tbg]

fixes, e.g., parsing of 1e-/test/-5.

very much related to #1810.

[tamird]

LGTM
On Jul 25, 2015 11:20 AM, "Tobias Schottdorf" [email protected]
wrote:

fixes, e.g., parsing of 1e-/test/-5.

very much related to #1810

#1810.

You can view, comment on, or merge this pull request online at:

#1812
Commit Summary

• insert spaces between binary operators as well
• slightly enhance TestParse2

File Changes

• M sql/parser/expr.go
  https://github.com/cockroachdb/cockroach/pull/1812/files#diff-0 (12)
• M sql/parser/parse_test.go
  https://github.com/cockroachdb/cockroach/pull/1812/files#diff-1 (56)
• M sql/parser/walk_test.go
  https://github.com/cockroachdb/cockroach/pull/1812/files#diff-2 (22)

Patch Links:

• https://github.com/cockroachdb/cockroach/pull/1812.patch
• https://github.com/cockroachdb/cockroach/pull/1812.diff

—
Reply to this email directly or view it on GitHub
#1812.

[petermattis]

LGTM

[bdarnell]

1e- doesn't appear to be commonly accepted; both postgres and python require an exponent value (1e-0).

[tbg]

This only tests the parser, if you tried to evaluate this it (hopefully) wouldn't work. Discarding this in the parser would require a lookahead. Since we've copied that from Postgres, better not to touch it.

[petermattis]

I think this is a bug in scanner.scanNumber(). The postgres scanner has the following rules for scanning floating point numbers:

real            ({integer}|{decimal})[Ee][-+]?{digit}+
realfail1       ({integer}|{decimal})[Ee]
realfail2       ({integer}|{decimal})[Ee][-+]

If the scanner sees - or + it should require the following character to be a digit.

[tbg]

ok. Tracked anywhere? If not, I can go ahead and file it.

[petermattis]

Not tracked yet as far as I know. File away.