[tor-bugs] #19079 [Core Tor/Tor]: clang -m32 -ftrapv seems buggy with 64-bit signed integer multiply

[Tor Bug Tracker & Wiki]

#19079: clang -m32 -ftrapv seems buggy with 64-bit signed integer multiply
-------------------------------+------------------------------------
 Reporter:  nickm              |          Owner:  nickm
     Type:  defect             |         Status:  needs_revision
 Priority:  High               |      Milestone:  Tor: 0.2.9.x-final
Component:  Core Tor/Tor       |        Version:
 Severity:  Normal             |     Resolution:
 Keywords:  TorCoreTeam201605  |  Actual Points:  .1
Parent ID:                     |         Points:  .1
 Reviewer:                     |        Sponsor:
-------------------------------+------------------------------------
Changes (by teor):

 * status:  needs_review => needs_revision


Comment:

 code review:

 6d6c828 Include __mulodi4 in libor_ctime when it fixes clang -m32 -ftrapv

 You forgot an atoi on argv[3], and you said on IRC it won't matter.

 3303460 Add __mulodi4 source to src/ext

 Do we need unit tests for this?
 Or should we just assume the llvm source is ok?
 (I wonder about bugs like http://kqueue.org/blog/2013/09/17/cltq/ , the
 specific instance isn't an issue here because the types are all the same,
 but in general, these sort of subtle compiler bugs concern me.)

 Other than that, looks good to merge.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/19079#comment:4>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online