[foreign-jextract] RFR: 8263018: Improve API for lifecycle of native resources [v2]

[Maurizio Cimadamore]

On Wed, 24 Mar 2021 12:32:03 GMT, Jorn Vernee <jvernee at openjdk.org> wrote:

>> Maurizio Cimadamore has updated the pull request with a new target base due to a merge or a rebase. The pull request now contains 98 commits:
>> 
>>  - Merge branch 'foreign-jextract' into jextract+resourceScope
>>  - Merge branch 'foreign-memaccess+abi' into jextract+resourceScope
>>  - Merge branch 'foreign-jextract' into jextract+resourceScope
>>  - Add overloads accepting ResourceScope parameter
>>  - Merge branch 'resourceScope' into jextract+resourceScope
>>  - Add ResourceScope overloads in CLinker
>>    Rewrite StdLib test not to use NativeScope
>>    Tweak names of allocator factories
>>    Fix javadoc
>>  - All tests pass
>>    * Fix upcall overload taking NativeScope
>>    * Generate NativeScope in output package
>>    * Add overload for functions returning a struct so than a NativeScope can be passed
>>    * Removed some now unused features from RuntimeHelper (ti make scope non-transferrable)
>>    * Fixup vararg invoker to deal with allocator parameters
>>    * Fixup tests to (a) depend on the generated NativeScope and (b) not to use TWR with segments directly
>>    Notes:
>>    * the internal libclang probably needs to be re-generated
>>  - Merge branch 'resourceScope' into jextract+resourceScope
>>  - Fix javadoc of SegmentAllocator
>>  - Address rest of review comments from Paul
>>  - ... and 88 more: https://git.openjdk.java.net/panama-foreign/compare/4e696684...6316ddca
>
> src/jdk.incubator.jextract/share/classes/jdk/internal/clang/TranslationUnit.java line 69:
> 
>> 67:     public final void save(Path path) throws TranslationUnitSaveException {
>> 68:         MemorySegment pathStr = CLinker.toCString(path.toAbsolutePath().toString());
>> 69:         SaveError res = SaveError.valueOf(Index_h.clang_saveTranslationUnit(tu, pathStr, 0));
> 
> This function takes the address of the pathStr and passes it to a native function, so I don't believe using a GC managed segment here is safe.

true

> src/jdk.incubator.jextract/share/classes/jdk/internal/clang/Type.java line 115:
> 
>> 113:     private long getOffsetOf0(String fieldName) {
>> 114:         MemorySegment cfname = CLinker.toCString(fieldName);
>> 115:         return Index_h.clang_Type_getOffsetOf(type, cfname);
> 
> Same here, this function takes the address of `cfname` so can't use a managed segment

true

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

PR: https://git.openjdk.java.net/panama-foreign/pull/474