boost::math::quantile() incorrect for poisson_distribution

Hi,

We spotted an apparent bias in the inverse cumulative poisson distribution function,
looking deeper, we found boost is generating different results when compared with scipy and statslib (https://github.com/kthohr/stats)

The value of mean is not important, but for p < 0.5, the result is offset by -1

ie:
```
double mean = 2719.13;
double p = 0.49;
double r_boost = boost::math::quantile(boost::math::poisson_distribution<>(mean), p);
double r_stats = stats::qpois(p,mean);
r_boost --> 2717
r_stats --> 2718
```

In Scipy:
```
import scipy.stats
mean = 2719.13
r = 0.49
print(scipy.stats.poisson.ppf(r, mean))
printout --> 2718.0
```

Code:
```
// Alternative statistics library: statslib
// https://github.com/kthohr/stats
#include
#include
#include

double mean = 2719.13;

void test_range( double a, double b )
{
for (double p = a; p < b+0.0001; p += 0.01)
{
double r_boost = boost::math::quantile(boost::math::poisson_distribution<>(mean), p);
double r_stats = stats::qpois(p,mean);
double diff = r_stats - r_boost;
std::cout << "Mean:" << mean << " p:" << p << " Boost=" << r_boost << " statslib=" << r_stats << " difference=" << diff << (diff != 0 ? " DIFFERENT" : " same") << std::endl;
}
std::cout << std::endl;
}

int main()
{
test_range(0.01, 0.04);
test_range(0.47, 0.52);
test_range(0.96, 0.99);
return 0;
}
```

Output:
```
Mean:2719.13 p:0.01 Boost=2598 statslib=2599 difference=1 DIFFERENT
Mean:2719.13 p:0.02 Boost=2612 statslib=2613 difference=1 DIFFERENT
Mean:2719.13 p:0.03 Boost=2620 statslib=2621 difference=1 DIFFERENT
Mean:2719.13 p:0.04 Boost=2627 statslib=2628 difference=1 DIFFERENT

Mean:2719.13 p:0.47 Boost=2714 statslib=2715 difference=1 DIFFERENT
Mean:2719.13 p:0.48 Boost=2715 statslib=2716 difference=1 DIFFERENT
Mean:2719.13 p:0.49 Boost=2717 statslib=2718 difference=1 DIFFERENT
Mean:2719.13 p:0.5 Boost=2719 statslib=2719 difference=0 same
Mean:2719.13 p:0.51 Boost=2720 statslib=2720 difference=0 same
Mean:2719.13 p:0.52 Boost=2722 statslib=2722 difference=0 same

Mean:2719.13 p:0.96 Boost=2811 statslib=2811 difference=0 same
Mean:2719.13 p:0.97 Boost=2818 statslib=2818 difference=0 same
Mean:2719.13 p:0.98 Boost=2827 statslib=2827 difference=0 same
Mean:2719.13 p:0.99 Boost=2841 statslib=2841 difference=0 same
```