TREE_SIDE_EFFECTS vs STATEMENT_LIST vs recompute_side_effects

Richard Kenner kenner@vlsi1.ultra.nyu.edu
Thu Jun 24 18:45:00 GMT 2004


    The deeper issue is in this case the bit propagates up through
    the COND_EXPR, eventually to a COMPONENT_REF.

OK, so you're talking about the case of a non-void-type COND_EXPR, then?

But in that case, the arms will be replaced with variables, not GOTO_EXPRs.
Why are side effects being recomputed then?



More information about the Gcc mailing list