mirror of https://github.com/CGAL/cgal
Fix a bug when LEDA is used instead of GMP
That is a really trivial bug fix, for a rather nasty bug! I spend several hours on it! Currently the `leda::rational::operator>>` does not clear the `failbit` of the stream, when it reads a number without denominator, and the EOF directly after. The flag `eofbit` and `failbit` are set while trying to read the '/', but in that case the `failbit` should be cleared from the stream status. Trivial bug fix for master.
This commit is contained in:
parent
d3d758421d
commit
baffa5a7f8
|
|
@ -34,4 +34,4 @@
|
|||
3
|
||||
7 4
|
||||
7 5
|
||||
8 4
|
||||
8 4
|
||||
|
|
|
|||
|
|
@ -17,4 +17,4 @@
|
|||
3
|
||||
4 0
|
||||
4 2
|
||||
6 3
|
||||
6 3
|
||||
|
|
|
|||
Loading…
Reference in New Issue