RFR: 8365053: Refresh hotspot precompiled.hpp with headers based on current frequency [v18]
Magnus Ihse Bursie
ihse at openjdk.org
Fri Aug 29 11:07:59 UTC 2025
On Wed, 27 Aug 2025 16:45:55 GMT, Francesco Andreuzzi <duke at openjdk.org> wrote:
>> Ran the script locally, and noticed that on my Mac this was the difference:
>>
>>
>> -#include "oops/oopHandle.inline.hpp"
>> #include "oops/oop.inline.hpp"
>> +#include "oops/oopHandle.inline.hpp"
>>
>>
>> Something is a bit off with sorting, I think.
>>
>> Indeed, this fails:
>>
>>
>> $ diff --git a/test/hotspot/jtreg/sources/TestIncludesAreSorted.java b/test/hotspot/jtreg/sources/TestIncludesAreSorted.java
>> index 1a304a44944..287aad01b78 100644
>> --- a/test/hotspot/jtreg/sources/TestIncludesAreSorted.java
>> +++ b/test/hotspot/jtreg/sources/TestIncludesAreSorted.java
>> @@ -58,6 +58,7 @@ public class TestIncludesAreSorted {
>> "share/metaprogramming",
>> "share/oops",
>> "share/opto",
>> + "share/precompiled",
>> "share/services",
>> "share/utilities"
>> };
>>
>> $ CONF=macosx-aarch64-server-fastdebug make images test TEST=sources/
>> STDERR:
>> java.lang.RuntimeException: 1 files with unsorted headers found:
>> /Users/shipilev/Work/shipilev-jdk/src/hotspot/share/precompiled/precompiled.hpp
>>
>>
>> This is a minor thing, and we can technically deal with it once we sort the includes in `precompiled.hpp` in the constellation of issues that are dedicated to include sorting. But it would also be great to do this right from the get-go.
>>
>> See if there is anything easy missing in the script?
>
>> See if there is anything easy missing in the script?
>
> Thanks @shipilev, I addressed this in 6693cefe01e03cd05a11a93cf04d0d6af78f42bb. I use `SortIncludes.java` instead of `sort`, and I also added `precompiled` to `TestIncludesAreSorted.java`.
Once again thank you @fandreuz for the effort you put into this!
-------------
PR Comment: https://git.openjdk.org/jdk/pull/26681#issuecomment-3236658521
More information about the hotspot-dev
mailing list