https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118013
David Malcolm <dmalcolm at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2024-12-13
Summary|bogus "infinite loop" |bogus "infinite loop"
|warning |warning due to -fanalyzer
| |not restricting range of
| |iterator variable
Ever confirmed|0 |1
--- Comment #1 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Thanks for filing this bug report.
Confirmed; reproducer on GCC trunk with Compiler Explorer:
https://godbolt.org/z/TGh8sbjYc
which shows the control flow in a slightly different way (hopefully more
readable).
It looks like the analyzer is considering the (impossible) path in which we
exit the inner loop due to
(j >= nsubst)
but also fail to execute the increment of "i" due to:
j != nsubst
i.e. that:
j > nsubst
which of course isn't possible, since j is being incremented up from 0 until it
reaches nsubst, and thus can only be in the range [0, nsubst].
Looks like the analyzer needs to be smarter about iteration variables.
