Leopold Toetsch wrote:
Nick Glencross wrote:
Leopold Toetsch wrote:
Fixed, thanks for testing.
Hmmm... Not seeing any improvement at my end ...
I don't see any indication that it is still broken. I ran "make test" on
two different boxes w/o failures (the 2 cases are
imcc/t/imcpasm/opt1_71.pir and _72.pir).
What is still failing? What says gdb/$DEBUGGER? Which paltform/OS?
At the moment I'm on Windows/cygwin, and don't have access to Linux.
$ svn update
At revision 8234.
make clean & rebuild
$ perl -Ilib imcc/t/imcpasm/opt1.t
1..72
ok 1 - branch opt if
ok 2 - branch opt gt
...
ok 69 - constant add big nums
ok 70 # skip constant concat N/Y
not ok 71 - segv - last ins changed
# Failed test (imcc/t/imcpasm/opt1.t at line 1038)
# '# IMCC does produce b0rken PASM files
# # see http://[EMAIL PROTECTED]/rt3/Ticket/Display.html?id=32392
# main:
# @pcc_sub_call_0:
# set I0, 1
# set I1, 0
# set I2, 0
# set I3, 0
# set I4, 0
# set S1, ""
# set_p_pc P0, func
# invokecc
# end
# func:
# inc I30
# '
# doesn't match '/func:
# inc I\d+
# null I0
# null I3
# returncc
# /
# '
ok 72 - segv - last ins deleted
# Looks like you failed 1 test of 72.
It's doing the same from the command-line. Was the fix for opt1_71.pir
in r8227?
Cheers,
Nick
