Reviewers: Kasper Lund, iposva, Message: From the little I now understand of ARM encoding, this optimization was never possible on ARM. The minimal forward branch that you can represent (zero in the instruction field) will be 8 bytes forward, so you must branch over at least one instruction. This means "jump to next" is not possible on ARM, and therefor this optimization was never possible.
I verified that indeed this code was never eliminating any jumps, and that we are generating identical code before and after this change. I am not at all familiar with the ARM code, so Ivan please look closely and make sure I didn't do anything stupid. Description: Remove the ARM jump elimination. This completely removes jump elimination. Please review this at http://codereview.chromium.org/3807 Affected files: M src/assembler-arm.cc M src/assembler-arm.h M src/assembler-ia32.cc M src/flag-definitions.h --~--~---------~--~----~------------~-------~--~----~ v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev -~----------~----~----~----~------~----~------~--~---
