Fix hull_test isMeshConvex() vertices check

[cjmayo]

Don't know why this isn't failing on GitHub but for me Hull.FailingTest1 and Hull.FailingTest2 cause coredumps:

[ RUN      ] Hull.FailingTest1
manifold_test: /usr/include/glm/detail/type_vec3.inl:188: constexpr const T& glm::vec<3, T, Q>::operator[](length_type) const [with T = int; glm::qualifier Q = glm::packed_highp; length_type = int]: Assertion `(i) >= 0 && (i) < (this->length())' failed.

gdb:

#4  0x00007f3c1f970f26 in __assert_fail () from /usr/lib64/libc.so.6
#5  0x000056477e252666 in glm::vec<3, int, (glm::qualifier)0>::operator[] (this=<optimized out>, i=3)
    at /usr/include/glm/detail/type_vec3.inl:186
#6  glm::vec<3, int, (glm::qualifier)0>::operator[] (this=<optimized out>, i=3) at /usr/include/glm/detail/type_vec3.inl:186
#7  isMeshConvex (hullManifold=..., epsilon=epsilon@entry=8.9962799999999992e-06)
    at manifold-e26e01a2310db36c23240fb7ad2e9692f6411a92/test/hull_test.cpp:43
#8  0x000056477e252ba5 in Hull_FailingTest1_Test::TestBody (this=<optimized out>)
    at manifold-e26e01a2310db36c23240fb7ad2e9692f6411a92/test/hull_test.cpp:165

[elalish]

Good catch, thanks! FYI @Kushal-Shah-03

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 87.26%. Comparing base (d437097) to head (2d866b8).
Report is 75 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #898      +/-   ##
==========================================
- Coverage   91.84%   87.26%   -4.58%     
==========================================
  Files          37       66      +29     
  Lines        4976     9442    +4466     
  Branches        0     1033    +1033     
==========================================
+ Hits         4570     8240    +3670     
- Misses        406     1202     +796     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Kushal-Shah-03]

Thanks for the catch, that might explain why some of the tests had a larger epsilon than expected, I had just assumed it was due to the conversion error. I will run my tests again.