We currently get this: TEST-UNEXPECTED-FAIL | lirasm | lirasm --execute ../nanojit-src/lirasm/tests/hardfloat/ui2d.in expected output Output is: 4.29497e+09 actual output Output is: 4.29497e+009 That's really annoying. We currently use |cout << d| to print the double result. Anyone know if printf("%f") or printf("%g") gives more consistent results across all platforms?
Looks like we won't be able to eliminate compiler/lib differences, but I wonder if post-processing is a feasible option.
Created attachment 487469 [details] [diff] [review] sed to the rescue Not pretty but should do the job to make windows happy.
Attachment #487469 - Flags: review?(nnethercote)
Comment on attachment 487469 [details] [diff] [review] sed to the rescue r=me if you add a brief comment explaining why the sed command is there. Thanks!
Attachment #487469 - Flags: review?(nnethercote) → review+
(In reply to comment #4) > rreitmai http://hg.mozilla.org/projects/nanojit-central/rev/fa319402f3bb Gak, change applied to input file, not output... fix coming.
Ignore comment 5, the problem was cygwin sed's treatment of escape. fix in http://hg.mozilla.org/projects/nanojit-central/rev/cf0963c16a8c
http://hg.mozilla.org/tracemonkey/rev/104e20381ef1 http://hg.mozilla.org/tracemonkey/rev/2b77342e96dd http://hg.mozilla.org/tracemonkey/rev/a23a09d2a9b6
Whiteboard: fixed-in-nanojit → fixed-in-nanojit, fixed-in-tracemonkey
Status: ASSIGNED → RESOLVED
Last Resolved: 7 years ago
Resolution: --- → FIXED
Component: Nanojit → Nanojit
Product: Core → Core Graveyard
You need to log in before you can comment on or make changes to this bug.