[11u] RFR: 8249821: Separate libharfbuzz from libfontmanager

Lindenmaier, Goetz goetz.lindenmaier at sap.com
Tue Sep 22 09:12:52 UTC 2020


Hi Mario,

Thanks for reviewing. 

You are right, in itself, this change makes no sense for 11.
But it is a prerequisite for "8250894: Provide a configure option 
to build and run against the platform libharfbuzz"
which in my eyes is a change that makes sense for 11.
It will help to adapt the VM to new os releases that come 
with more recent libs.

> so the only question is if we want to keep
> this extra comment at this point or not:
> // Solaris does not link because the implementation in .cc specifies
> num_glyphs as const.

There are several other adptions in the file for solaris, 
that are commented similarly. See line 284
http://cr.openjdk.java.net/~goetz/wr20/8249821-separate_harfbuzz-jdk11/01/src/java.desktop/share/native/libharfbuzz/hb-subset-cff-common.hh.html

I guess these comments help in case  libharfbuzz is
updated to find places where adaptions for solaris are 
needed.

Best regards,
  Goetz.



> -----Original Message-----
> From: Mario Torre <neugens at redhat.com>
> Sent: Tuesday, September 22, 2020 10:19 AM
> To: Lindenmaier, Goetz <goetz.lindenmaier at sap.com>
> Cc: jdk-updates-dev at openjdk.java.net
> Subject: Re: [11u] RFR: 8249821: Separate libharfbuzz from libfontmanager
> 
> Hi Goetz,
> 
> Looks good to me.
> 
> I'm a bit on the fence because I think this kind of changes shouldn't
> really be backported, but the fonts subsystems is often target for
> security patches and backports so I guess not having this will
> certainly makes future work a lot more complicated than necessary that
> this is a justified change, so the only question is if we want to keep
> this extra comment at this point or not:
> 
> // Solaris does not link because the implementation in .cc specifies
> num_glyphs as const.
> 
> Other than that, from my side is OK, you can consider it reviewed.
> 
> Cheers,
> Mario
> 
> On Tue, Sep 22, 2020 at 9:35 AM Lindenmaier, Goetz
> <goetz.lindenmaier at sap.com> wrote:
> >
> > Hi,
> >
> > I downport this for parity with 11.0.10-oracle.
> >
> > I had to fix a row of things in the makefile.
> >  - some utility functions are missing in the 11 build
> >  - I implemented the solaris build. I checked Solaris sparc works.
> >
> > The rest is just a plain file move, except for adding a const
> > in hb_subset_cff_common.hh  hb_serialize_cff_fdselect().
> > The implementation in .cc declares the argument const, and
> > solaris fails to link without seeing the const in the header.
> >
> > http://cr.openjdk.java.net/~goetz/wr20/8249821-separate_harfbuzz-
> jdk11/01/
> >
> > Please review.
> >
> > https://bugs.openjdk.java.net/browse/JDK-8249821
> > https://hg.openjdk.java.net/jdk/client/rev/cda47533b1ee
> >
> > Best regards,
> >   Goetz.
> >
> 
> 
> --
> Mario Torre
> Associate Manager, Software Engineering
> Red Hat GmbH <https://www.redhat.com>
> 9704 A60C B4BE A8B8 0F30  9205 5D7E 4952 3F65 7898



More information about the jdk-updates-dev mailing list