libreoffice is currently FTBFS on i686 in Rawhide. Both recent attempts to build it failed only on that arch:

https://koji.fedoraproject.org/koji/buildinfo?buildID=1104722
https://koji.fedoraproject.org/koji/buildinfo?buildID=1113247

This is an urgent issue as the failure to build on i686 means it is not updated on any other arch either, and it is not installable in Rawhide at all because it needs rebuilding for Python 3.7, the current packages depend on Python 3.6 and thus cannot be installed. This means all Rawhide installs are currently lacking LibreOffice and anyone upgrading to Rawhide loses LibreOffice as part of the upgrade.

The relevant bit of the build log (I think) looks to be this:

[build XHC] nlpsolver as
  adding: com.sun.star.comp.Calc.NLPSolver/Options.xhp (deflated 78%)
  adding: com.sun.star.comp.Calc.NLPSolver/Usage.xhp (deflated 54%)
  adding: com.sun.star.comp.Calc.NLPSolver/Options.xhp (deflated 77%)
  adding: com.sun.star.comp.Calc.NLPSolver/Usage.xhp (deflated 54%)
S=/builddir/build/BUILD/libreoffice-6.0.5.2 && I=$S/instdir && W=$S/workdir &&  mkdir -p $W/Extension/nlpsolver/root/help/as && LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program"   $W/LinkTarget/Executable/HelpLinker -mod help -extlangsrc $W/Extension/nlpsolver/help/as -sty $S/xmlhelp/util/embed.xsl -extlangdest $W/Extension/nlpsolver/root/help/as -idxcaption $S/xmlhelp/util/idxcaption.xsl -idxcontent $S/xmlhelp/util/idxcontent.xsl com.sun.star.comp.Calc.NLPSolver/Options.xhp com.sun.star.comp.Calc.NLPSolver/Usage.xhp && (cd $W/Extension/nlpsolver/help/as && zip  -r $W/Extension/nlpsolver/root/help/as/help.jar com.sun.star.comp.Calc.NLPSolver/Options.xhp com.sun.star.comp.Calc.NLPSolver/Usage.xhp) && LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program"   $W/LinkTarget/Executable/HelpIndexer -lang as -mod help -dir $W/Extension/nlpsolver/root/help/as && rm -fr $W/Extension/nlpsolver/root/help/as/caption $W/Extension/nlpsolver/root/help/as/content && touch $W/Extension/nlpsolver/root/help/as.done
[build XHC] nlpsolver bg
free(): invalid next size (fast)
  adding: com.sun.star.comp.Calc.NLPSolver/Options.xhp (deflated 81%)
  adding: com.sun.star.comp.Calc.NLPSolver/Usage.xhp (deflated 60%)
free(): invalid next size (fast)
/bin/sh: line 1:  5451 Aborted                 (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program" $W/LinkTarget/Executable/HelpIndexer -lang ar -mod help -dir $W/Extension/nlpsolver/root/help/ar
/bin/sh: line 1:  5460 Aborted                 (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program" $W/LinkTarget/Executable/HelpIndexer -lang as -mod help -dir $W/Extension/nlpsolver/root/help/as
make[1]: *** [/builddir/build/BUILD/libreoffice-6.0.5.2/nlpsolver/Extension_nlpsolver.mk:13: /builddir/build/BUILD/libreoffice-6.0.5.2/workdir/Extension/nlpsolver/root/help/as.done] Error 134
make[1]: *** Waiting for unfinished jobs....
make[1]: *** [/builddir/build/BUILD/libreoffice-6.0.5.2/nlpsolver/Extension_nlpsolver.mk:13: /builddir/build/BUILD/libreoffice-6.0.5.2/workdir/Extension/nlpsolver/root/help/ar.done] Error 134
/bin/sh: line 1:  5437 Segmentation fault      (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program" $W/LinkTarget/Executable/HelpIndexer -lang en-US -mod help -dir $W/Extension/nlpsolver/root/help/en-US
make[1]: *** [/builddir/build/BUILD/libreoffice-6.0.5.2/nlpsolver/Extension_nlpsolver.mk:13: /builddir/build/BUILD/libreoffice-6.0.5.2/workdir/Extension/nlpsolver/root/help/en-US.done] Error 139
/bin/sh: line 1:  5443 Segmentation fault      (core dumped) LD_LIBRARY_PATH=${LD_LIBRARY_PATH:+$LD_LIBRARY_PATH:}"$I/program:$I/program" $W/LinkTarget/Executable/HelpIndexer -lang af -mod help -dir $W/Extension/nlpsolver/root/help/af
make[1]: *** [/builddir/build/BUILD/libreoffice-6.0.5.2/nlpsolver/Extension_nlpsolver.mk:13: /builddir/build/BUILD/libreoffice-6.0.5.2/workdir/Extension/nlpsolver/root/help/af.done] Error 139

i.e. we get a couple of "free(): invalid next size (fast)" errors, then a coredump. I don't think Koji preserves the coredump, and I don't think anyone's reproduced it locally and got a copy of it (or the traceback) yet.

Jeff Backus from the x86 SIG is looking into this ATM, CCing him.