So is it something inside of QT that's spinning in a loop when it ought to be blocked waiting for an event from the operating system instead? I wonder if the best solution involves patching QT? Does it do that high cpu usage thing when compiled with QT under other operating systems, or only on Linux? Should --enable-qt imply --enable-qtpipes or not? When qt but not qtpipes perhaps it ought to work like it did before with that define commented off?
I tried to compile with --disable-qt and it fails with errors concerning redefinition of operator new and operator delete. I wish I had more time to really learn the TeXmacs source and to try and help. I'm in the middle of lawsuits pro per using TeXmacs for court document production. They look great. I hope it helps my case. On Tue, Jun 9, 2015, 16:31 Philippe Joyez <texmacs.5.pjo...@spamgourmet.com> wrote: > Hola Miguel, > > On my linux box compilation still fails even after your correction. The > problem is different now, though: several undefined reference to > `remove_notifier(socket_notifier)', `add_notifier(socket_notifier)' , > `perform_select()' > > I could fix that by "./configure --enable-qtpipes", but then the bin that > is > created (still) has high CPU load when idle which I thought r9720 was > aiming > to fix (and your trick of commenting #define QT_CPU_FIX 1 in > src/src/Kernel/Abstractions/basic.hpp doesn't work since it breaks > compilation) > > So presently unless one accepts to run with this high power consumption, > it's not really possible to test Joris' latest improvements in linux... > too bad. > > Philippe > > > > > > _______________________________________________ > Texmacs-dev mailing list > Texmacs-dev@gnu.org > https://lists.gnu.org/mailman/listinfo/texmacs-dev >
_______________________________________________ Texmacs-dev mailing list Texmacs-dev@gnu.org https://lists.gnu.org/mailman/listinfo/texmacs-dev
