This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Loop invariants


I've been working all day today to try to fix the ICE in the loop code
when compiling i-cobol.adb, but am having trouble understanding what
is meant by an "invariant" in the SSA context.

It looks like nearly all of the function is inside a loop.  "j" is
the array index and we have

	T.502_17 = (<unnamed type>)j.500_60;

as a statement within the loop. 

Also within the loop is a statement

	(*T.517_68)[T.502_17]{lb:1 sz:1} = 48;

It's trying to see where it can move the latter statement and running into
max_loop of the first being zero, which causes the SIGSEGV.  I'm not at all
sure what's *supposed* to be happening here and need some help.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]