RFR: 8351345: [IR Framework] Improve reported disabled IR verification messages [v2]
Hamlin Li
mli at openjdk.org
Fri Mar 7 10:53:46 UTC 2025
On Thu, 6 Mar 2025 21:45:44 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:
>> Hamlin Li has updated the pull request incrementally with one additional commit since the last revision:
>>
>> refine
>
> test/hotspot/jtreg/compiler/lib/ir_framework/TestFramework.java line 752:
>
>> 750: if (!flag.startsWith("-D") && !flag.startsWith("-e") && JTREG_WHITELIST_FLAGS.stream().noneMatch(flag::contains)) {
>> 751: // Found VM flag that is not whitelisted
>> 752: System.out.println("Non-whitelisted JTreg VM or Javaoptions flag: " + flag);
>
> That's a good idea! Is the intention to just report the first non-whitelisted flag found or all of them? I guess just one of them is fine to indicate the reason for not performing IR matching (could be verbose to report all non-whitelisted flags otherwise).
>
> I would merge this message with the already existing one here and remove that one in favor of the new one:
> https://github.com/openjdk/jdk/blob/a23fb0af65f491ef655ba114fcc8032a09a55213/test/hotspot/jtreg/compiler/lib/ir_framework/TestFramework.java#L612-L615
>
> Another thought while cleaning this up: We could also improve this message
> https://github.com/openjdk/jdk/blob/a23fb0af65f491ef655ba114fcc8032a09a55213/test/hotspot/jtreg/compiler/lib/ir_framework/TestFramework.java#L597-L602
> and split it into three separate bailouts + messages. Could be part of this RFE (you could then set a new title for the issue to something like `[IR Framework] Improve reported disabled IR verification messages`).
Thanks for the suggestion! It make sense to me. Could you have another look?
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/23931#discussion_r1984854330
More information about the hotspot-compiler-dev
mailing list