[jdk17u-dev] RFR: 8311583: tableswitch broken by JDK-8310577

[Martin Balao]

Hi,

The 17u backport of 8311583 does not apply cleanly because of the absence of 8310577. In particular, the data type used in `Bytecodes::special_length_at` is `jlong`/`jint` instead of `int`. In my view, 8310577 should not be considered a dependency because: 1) the data type will be replaced by `int64_t` anyways and we can cast to `int` where appropriate (in the same line), 2) 8310577 is an enhancement, 3) touches multiple unrelated files (far beyond the scope of 8311583) and 4) has broken things —and has follow up fixes that would be needed, propagating risks—. In fact, 8311583 is a follow up fix of 8310577. I want 8311583 in 17u as a dependency of another fix.

No regressions observed in the hotspot/jtreg/runtime/verifier tests category.

Thanks,
Martin.-

-------------

Commit messages:
 - Backport 135f64e603266e0bb8852edf77d21a9cd79388ce

Changes: https://git.openjdk.org/jdk17u-dev/pull/1988/files
 Webrev: https://webrevs.openjdk.org/?repo=jdk17u-dev&pr=1988&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8311583
  Stats: 153 lines in 4 files changed: 146 ins; 0 del; 7 mod
  Patch: https://git.openjdk.org/jdk17u-dev/pull/1988.diff
  Fetch: git fetch https://git.openjdk.org/jdk17u-dev.git pull/1988/head:pull/1988

PR: https://git.openjdk.org/jdk17u-dev/pull/1988