[8u] RFR: 8255003: Build failures on Solaris (was Re: Build failures on illumos (Solaris))

[Severin Gehwolf]

On Wed, 2020-10-21 at 19:02 +0300, Andrew Brygin wrote:
> Hello Severin,
> 
> On 21/10/2020 10:17, Severin Gehwolf wrote:
> > Hi,
> > 
> > On Tue, 2020-10-20 at 21:14 +0300, Andrew Brygin wrote:
> > > I have updated the fix according to following comment:
> > > 
> > > https://bugs.openjdk.java.net/browse/JDK-8255003?focusedCommentId=14375484&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14375484
> > > 
> > > The removal of Java_sun_awt_X11_XToolkit_getDefaultScreenData will be
> > > tracked as JDK-8251840.
> > 
> > Thank you!
> > 
> > > Please take a look to updated webrev:
> > > 
> > > http://cr.openjdk.java.net/~bae/8u/8255003/webrev.01/
> > 
> > Does this actually build on Windows? I'm thinking not, since it's
> > missing a mapfile for windows, no?
> 
> build on windows works fine, because it does not use map files at all:
> 
> https://hg.openjdk.java.net/jdk8u/jdk8u/file/tip/common/autoconf/flags.m4#l207

OK, thanks. It's just strange to call a make macro
(SetupNativeCompilation) with a file that doesn't exist.

> > How about special case Linux instead and let other platforms use the
> > old, pre JDK-8250627, mapfile-vers and use the one adding
> > Java_jdk_internal_platform_cgroupv1_Metrics_isUseContainerSupport on
> > Linux like so, perhaps?
> > 
> > https://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-8255003/01/webrev/
> 
> Actually, only gcc (except macos) and solaris studio toolchains use
> mapfiles, so this option also looks good.

Thanks for the review! If somebody could test it on Solaris (together
with the one-liner fix for JDK-8251840) i'll propose this patch for
approval.

Thanks,
Severin