Fix bug in edge case in populating PointerMaps
authordanno@chromium.org <danno@chromium.org@ce2b1a6d-e550-0410-aec6-3dcde31c8c00>
Thu, 24 Jan 2013 15:05:38 +0000 (15:05 +0000)
committerdanno@chromium.org <danno@chromium.org@ce2b1a6d-e550-0410-aec6-3dcde31c8c00>
Thu, 24 Jan 2013 15:05:38 +0000 (15:05 +0000)
commit3c523437389b8c956edf7fb4745901640b437f52
treefb371885df370e9745eb6ed13a4bb64eddd456dc
parentdd92c41aa1361615c24458335137be40b153fc1c
Fix bug in edge case in populating PointerMaps

If a live range was split immediately before an instruction that required a
safe-point pointer map, it was possible for the safe-point to incorrectly use the
register/stack slot information of the preceeding live range rather than the one
that actually covers the safe-point.

R=mstarzinger@chromium.org,vegorov@chromium.org

Review URL: https://chromiumcodereview.appspot.com/12040052

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13493 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
src/lithium-allocator.cc