summaryrefslogtreecommitdiff
path: root/test/scaladoc
diff options
context:
space:
mode:
authorGrzegorz Kossakowski <grzegorz.kossakowski@gmail.com>2012-11-09 22:58:47 -0800
committerPaul Phillips <paulp@improving.org>2012-12-19 23:18:26 -0800
commitb6dd2d2e14aa88ea42622ac6a6092bba8ddb300a (patch)
treeab9e82096e325b8847066f319afd03b88640ea08 /test/scaladoc
parentba6a3d6f87c0007196a4a82aaef58a290f87b864 (diff)
downloadscala-b6dd2d2e14aa88ea42622ac6a6092bba8ddb300a.tar.gz
scala-b6dd2d2e14aa88ea42622ac6a6092bba8ddb300a.tar.bz2
scala-b6dd2d2e14aa88ea42622ac6a6092bba8ddb300a.zip
Do not recompute stack frames when instrumenting bytecode.
It turns out that we do not need to do that. See comment in `ProfilerVisitor.java`. Also, since recomputing stack frame map was the only reason we needed to implement `getCommonSuperClass` we can now remove its implementation that was causing problems on Java 7 due to a cyclic dependency involving class loader because we would try to load a class we are currently transforming and transformer is triggered just before classloading. //cc @namin who worked on this code with me.
Diffstat (limited to 'test/scaladoc')
0 files changed, 0 insertions, 0 deletions