[MLton] svn/HEAD has regressions

Wesley W. Terpstra wesley at terpstra.ca
Thu Apr 29 11:18:23 PDT 2010


I was preparing to make another debian and mingw package when I found:

testing FuhMishra
./bin/regression: line 28:  9598 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing fft
./bin/regression: line 28: 11830 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing flat-array.3
1c1
< 2 101
---
> 2 100
testing kitqsort
./bin/regression: line 28: 13546 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing kitsimple
./bin/regression: line 28: 13651 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing math
./bin/regression: line 28: 14271 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing real-int
./bin/regression: line 28: 15625 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing real
./bin/regression: line 28: 15691 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing testdyn1
./bin/regression: line 28: 17235 Segmentation fault      "$mlton"
"${flags[@]}" "${extraFlags[@]}" -output "$f" "$mlb"
testing time
9c9
< test3a        OK
---
> test3a        WRONG
12c12
< test4a        OK
---
> test4a        WRONG
testing time3
2,12c2,12
< 123.457
< 123
< 123.5
< 123.46
< 123.457
< 123.4568
< 123.45679
< 123.456789
< 123.4567890
< 123.45ve a
< 123.000
---
> 100.000
> 100
> 100.0
> 100.00
> 100.000
> 100.0000
> 100.00000
> 100.000000
> 100.0000000
> 100.000
> 100.000
testing barnes-hut
./bin/regression: line 265: 19522 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing matrix-multiply
./bin/regression: line 265: 19819 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing ray
./bin/regression: line 265: 20092 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing raytrace
./bin/regression: line 265: 20108 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing simple
./bin/regression: line 265: 20123 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing tsp
./bin/regression: line 265: 20233 Segmentation fault      $mlton -output
"$tmpf" "${flags[@]}" -default-ann 'nonexhaustiveMatch ignore' -default-ann
'redundantMatch ignore' "$tmpf.sml"
testing mlprof
./bin/regression: line 285: 20660 Segmentation fault      "$mlton"
"${flags[@]}" -output "$tmpf" "$f.mlb"

These all happen on debian/sid updated to April 29, 2010.

The crashes happen during the compile using MLton (r7452). The bootstrapping
compiler (r7430) did not exhibit the crashes. All the problems seem related
to the handling of floating point numbers.

I don't have time to debug this right now, but will in the next few days.
However, I'm betting someone will have a "Doh!" moment and be able to fix
this immediately. ;)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mlton.org/pipermail/mlton/attachments/20100429/f4366c3e/attachment.html


More information about the MLton mailing list