[Webkit-unassigned] [Bug 199759] Bytecode cache should use FileSystem

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=199759

--- Comment #13 from Christopher Reid <chris.reid at sony.com> ---
(In reply to Yusuke Suzuki from comment #12)
> Comment on attachment 374164 [details]
> Updated Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=374164&action=review
> 
> r=me with nits.
> 
> > Source/JavaScriptCore/API/JSScript.mm:164
> > +    FileSystem::MappedFileData mappedFile(fd, FileSystem::MappedFileMode::Private, success);
> 
> We should change this interface to something like this instead of passing
> boolean `success`.
> 
> Optional<MappedFileData> mappedFile = FileSystem::MappedFileData::open(...);
> 
> But this is not directly related to this patch. So I'm OK with this. Just
> commented for further refactoring.
> 

That's a good improvement. I'll look at doing that in a followup patch.

> > Source/JavaScriptCore/runtime/CachePayload.cpp:82
> >  #if !OS(WINDOWS)
> >          munmap(m_data, m_size);
> >  #else
> > -        RELEASE_ASSERT_NOT_REACHED();
> > +        UnmapViewOfFile(m_data);
> >  #endif
> >      } else
> 
> Can we extract this part in WTF::FileSystem as
> `FileSystem::unmapViewOfFile(buffer, size)` (and using it in
> MappedFileData::~MappedFileData)?

Sounds good I'll update that before landing.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190717/18ee8539/attachment.html>