RFR(S)  : 8226910 : make it possible to use jtreg's -match via run-test framework
igor.ignatyev at oracle.com
Sat Jul 6 18:50:16 UTC 2019
> On Jul 6, 2019, at 1:58 AM, David Holmes <david.holmes at oracle.com> wrote:
> Hi Igor,
> On 6/07/2019 1:09 pm, Igor Ignatyev wrote:
>> -- Igor
>>> On Jun 27, 2019, at 3:25 PM, Igor Ignatyev <igor.ignatyev at oracle.com> wrote:
>>>> 25 lines changed: 18 ins; 3 del; 4 mod;
>>> Hi all,
>>> could you please review this small patch which adds JTREG_RUN_PROBLEM_LISTS options to run-test framework? when JTREG_RUN_PROBLEM_LISTS is set to true, jtreg will use problem lists as values of -match: instead of -exclude, which effectively means it will run only problem listed tests.
> + Set to `true` of `false`.
> typo: s/of/or/
fixed .md, regenerated .html.
> Build changes seem okay - I can't attest to the operation of the flag.
here is how I verified that it does that it supposed to:
$ make test "JTREG=OPTIONS=-l;RUN_PROBLEM_LISTS=true" TEST=open/test/hotspot/jtreg/:hotspot_all
lists 53 tests, the same command w/o RUN_PROBLEM_LISTS (or w/ RUN_PROBLEM_LISTS=false) lists 6698 tests.
$ make test "JTREG=OPTIONS=-l;RUN_PROBLEM_LISTS=true;EXTRA_PROBLEM_LISTS=ProblemList-aot.txt
lists 81 tests, the same command w/o RUN_PROBLEM_LISTS lists 6670 tests.
>>> doc/building.html got changed when I ran update-build-docs, I can exclude it from the patch, but it seems it will keep changing every time we run update-build-docs, so I decided to at least bring it up.
> Weird it seems to have removed line-breaks in that paragraph. What platform did you build on?
I built on macos. now when I wrote that, I remember pandoc used to produce different results on macos. so I've rerun it on linux on the source w/o my change, and doc/building.html still got changed in the exact same way.
>>> JBS: https://bugs.openjdk.java.net/browse/JDK-8226910
>>> webrev: http://cr.openjdk.java.net/~iignatyev//8226910/webrev.00/index.html
>>> -- Igor
More information about the build-dev