Segfault toward end of hotspot-comp build (original) (raw)

John Rose john.r.rose at oracle.com
Tue Sep 6 16:29:45 PDT 2011

On Sep 6, 2011, at 10:16 AM, Charles Oliver Nutter wrote:

Ok, looks like it was a dirty build. It worked ok after a clean.

Yup. The names in globals.hpp are accessed in part via a BHE (big hairy enum). Changing the list in globals.hpp (or systemDictionary.hpp or vmSymbols.hpp) can cause *.o files to contain stale enum values.

We used to have automatically-generated *.o -> *hpp dependencies in our makefiles which would cause big painful rebuilds if you updated a key header file.

Now, since we went to a cpp-based solution for portability, I don't know how (or if) that stuff is tracked. It probably depends on how clever your toolchain is.

-- John -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mail.openjdk.java.net/pipermail/hotspot-compiler-dev/attachments/20110906/93462c90/attachment.html

More information about the hotspot-compiler-dev mailing list