boost 1.81 breaks libetonyek and gtv-calc-header-bar.cxx

[Douglas Guptill]

On Tue, Jan 10, 2023 at 12:17:29PM +0200, Ilmari Lauhakangas wrote:
> On 9.1.2023 3.18, Douglas Guptill wrote:
> > On Sun, Jan 08, 2023 at 09:19:52PM +0200, Ilmari Lauhakangas wrote:
> > > On 8.1.2023 16.23, Douglas Guptill wrote:
> > > > On Sun, Jan 08, 2023 at 04:40:12AM +0000, - Etna wrote:
> > > > 
> > > > > Boost 1.81 has been released, but Libreoffice has not been able to
> > > > > be compiled against it for more than a month.
> > > > > 
> > > > > Compilation fails with a whole bunch of duplicate symbol errors and
> > > > > libreoffice compilation cannot proceed.
> > > > 
> > > > I'm glad to hear that news.
> > > > 
> > > > I've been having compile trouble (at least some of which seems to be
> > > > related to boost) for several days, and being new to libreoffice, I am
> > > > wondering just what to do about it.  Repeating the compile doesn't
> > > > solve the problem.  :)
> > > > 
> > > > I have boost-1.78.0 on my system.  I could add this to my autogen.input
> > > >     --with-system-boost
> > > > 
> > > > Does that sound like a reasonable thing to do?
> > > 
> > > If you pull new commits with 'git pull -r' or './g pull -r' and you still
> > > see the problem, you could share the error here.
> > 
> > For sure.
> > 
> > I did that (./g pull -r) tonight.  Almost a repeat of what I did on
> > Jan 4.  My results on Jan 04 and tonight are similar.  There are many
> > warnings related to boost.
> > 
> > There are other warnings that I haven't yet done anything about.  I'm
> > not sure what to do; I thought that master would/should compile
> > without issues, and have been looking for an error in my procedures.
> > 
> > The 'make' finished with many warnings, but no *errors*.
> > After that I ran 'make check' - which reported errors.
> 
> You don't have to worry about the Boost warnings. I will copy the relevant
> test error from the make check run at the end as I'm not sure if others got
> it.

Thank you for the reply.

OK, I don't have to *worry* about the warnings.  But I cannot work
with them.  They make reading the compile output - which I expect to
do while working on LO - impossible for me.

I have done a new build with --with-system-boost, and the boost errors
are gone.

Regards,
Douglas.