Fix division of numerator 0

[chfast]

In case the division numerator is 0 its computed number of words m is also 0. Then avoid accessing un[m-1].

This is not a critical bug because in the worst case we increased the m by 1 but results were computed correctly.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.94%. Comparing base (150f2de) to head (a54970b).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #310   +/-   ##
=======================================
  Coverage   99.94%   99.94%           
=======================================
  Files          10       10           
  Lines        1890     1890           
=======================================
  Hits         1889     1889           
  Misses          1        1           

Flag	Coverage Δ
32bit	99.94% <100.00%> (ø)
gcc	99.41% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
include/intx/intx.hpp	99.86% <100.00%> (ø)

[gumb0]

This is not a critical bug because in the worst case we increased the m by 1 but results were computed correctly.

It could segfault theoretically on out of buffer access?

[chfast]

This is not a critical bug because in the worst case we increased the m by 1 but results were computed correctly.

It could segfault theoretically on out of buffer access?

No. The un is followed by vn in the normalized_div_args struct. So we were accessing the lowest word of the vn. So the comparison was garbage but not invalid memory access.

[chfast]

In other words, executing tests in constexpr works as pretty good undefined behavior sanitizer :)

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud