Fixed bug on overflown number literals in scanner #569
Conversation
dutor
requested review from
monadbobo,
ayyt,
CPWstatic,
boshengchen,
laura-ding and
whitewum
dutor
changed the title
| } | ||
| if (yyleng - i > 16) { | ||
| yyterminate(); | ||
| } |
There was a problem hiding this comment.
The maximum value of int64 is 9,223,372,036,854,775,807,
Hexadecimal is 0X7FFF FFFF FFFF FFFF
when value is in [ 0X8000 0000 0000 0000, 0XFFFF FFFF FFFF FFFF] ， the value is still overflow.
So, here add esle statement
There was a problem hiding this comment.
Hex is a way to specify integers as raw bytes, there is no such thing overflow in terms of literals.
There was a problem hiding this comment.
if （yyleng - i == 16 && yytext[i] > 7 ), Does the hexadecimal value not overflow?
the value in decimal is great than 9,223,372,036,854,775,807.
There was a problem hiding this comment.
Yeah， you are right, [-9,223,372,036,854,775,808 9,223,372,036,854,775,807] is [0, 0XFFFF FFFF FFFF FFFF]
👍
There was a problem hiding this comment.
you can replace magic number "18" with std::numeric_limits<int64_t>::digits10 .
There was a problem hiding this comment.
@monadbobo But here we are dealing with hexdecimal, instead of base 10.
| if (yyleng - i > 16) { | ||
| yyterminate(); | ||
| } | ||
| } | ||
| int64_t val = 0; | ||
| sscanf(yytext, "%lx", &val); |
There was a problem hiding this comment.
Should we judge the return value of sscanf?
There was a problem hiding this comment.
Yes, we should have done that, even though the regex and checking above already guarantee the correctness.
|
Unit testing passed. |
|
Unit testing passed. |
* parse date * enhance datetime * add test * fix review * rename Co-authored-by: Sophie <84560950+Sophie-Xie@users.noreply.github.com> Co-authored-by: jakevin <30525741+jackwener@users.noreply.github.com> Co-authored-by: Sophie <84560950+Sophie-Xie@users.noreply.github.com>
As title.
This PR will close #534