FWIW the issue should be resolved in R 3.4.1 - the default in Tcl/Tk is to build without Xft so for R 3.4.1 we have now re-build Tcl/Tk 8.6.6 binary with explicitly enabled Xft.
Cheers, Simon > On May 16, 2017, at 7:30 PM, Simon Urbanek <simon.urba...@r-project.org> > wrote: > > >> On May 16, 2017, at 12:20 PM, peter dalgaard <pda...@gmail.com> wrote: >> >> Mystery solved after looking into the actual .pkg files. They simply contain >> a full Tcl/Tk which installs into /usr/local, which obviously affects >> anything that links it dynamically. I somehow thought that they would >> install somewhere separate. (Stop grinning, Simon!) >> > > I'm not sure I understand the "they" - is it us? The difference in R 3.4.0 is > that we are using everything from the system (and XQuartz) whereas before we > were building dependencies ourselves before. > > >> Anyways, the way forward must be to get Tcl/Tk compiled exactly as they were >> in 3.3.x. >> > > R before 3.4.0 we used Tcl/Tk 8.6.0 for OS X Mavericks from 2013, for R 3.4.0 > we are using Tcl/Tk 8.6.6 for El Capitan. > > FWIW the config flags were > --disable-corefoundation --disable-framework --disable-aqua --enable-64bit > --with-tcl=/usr/local/lib --with-x --x-includes=/opt/X11/include > --x-libraries=/opt/X11/lib CC=clang CXX=clang++ > > I have no idea what you are talking about here so if anyone has suggestions > to test, let me know. > > Cheers, > Simon > > > >> -pd >> >> >> >> >>> On 16 May 2017, at 15:18 , peter dalgaard <pda...@gmail.com> wrote: >>> >>> Double-checked: Indeed, installing 3.3-branch-mavericks.pkg enables xft on >>> my local builds of R-patched and R-devel. Hmmmm...... >>> >>> -pd >>> >>>> On 16 May 2017, at 14:55 , peter dalgaard <pda...@gmail.com> wrote: >>>> >>>> Curiouser and curiouser, now my local build of R-devel stopped using xft >>>> _without any recompile_ I swear that it worked before I installed >>>> R-3.4.0.pkg.... >>>> >>>> I have both /opt/local/ and /usr/local variants of Tcl/Tk 8.6 and only the >>>> former has wish8.6 understanding the tk::pkgconfig stuff, so there is >>>> plenty of room for confusion. The latter seems to be from Simon's site; >>>> the other one could be a MacPorts relic... >>>> >>>> -pd >>>> >>>>> On 16 May 2017, at 13:28 , peter dalgaard <pda...@gmail.com> wrote: >>>>> >>>>> It's a build issue relating to Tcl/Tk, not specifically Rcmdr (Rcmdr does >>>>> something slightly weird[*] with the fonts too, but this seems to be >>>>> orthogonal to the ugly-font issue). >>>>> >>>>> To see the issue quickly, try >>>>> >>>>> library(tcltk) >>>>> .Tcl("::tk::pkgconfig get fontsystem") >>>>> >>>>> With CRAN 3.3.3, I get >>>>> >>>>> <Tcl> xft >>>>> >>>>> and likewise with a local nightly build of R-devel, but CRAN 3.4.0 says >>>>> >>>>>> .Tcl("::tk::pkgconfig get fontsystem") >>>>> Error in .Tcl("::tk::pkgconfig get fontsystem") : >>>>> could not find function ".Tcl" >>>>> >>>>> as does a local build of R-patched (???!!!). >>>>> >>>>> This is pretty weird. I'll see if I can dig up some more about where the >>>>> difference is. >>>>> >>>>> -pd >>>>> >>>>> >>>>> >>>>> [*] Try library(tcltk); demo(tkfaq) then library(Rcmdr) and watch the >>>>> font change in the tkfaq window. >>>>> >>>>> >>>>>> On 15 May 2017, at 23:17 , Fox, John <j...@mcmaster.ca> wrote: >>>>>> >>>>>> Dear Jose, >>>>>> >>>>>> It's a bit difficult to respond to your latest messages because you sent >>>>>> so many of them. >>>>>> >>>>>> I'm not entirely sure what's producing your font problems -- perhaps >>>>>> Simon will have some additional ideas -- but you should be aware that >>>>>> you can't load the Rcmdr package -- or, more generally, the tcltk >>>>>> package -- without XQuartz (or another X11 system). Normally, XQuartz >>>>>> should start automatically when the Rcmdr (or tcltk) package is loaded. >>>>>> That is, you should not have to start it separately. >>>>>> >>>>>> It may help to know that the Rcmdr package selects its default dialog >>>>>> font by querying the default Tk font, via the command >>>>>> >>>>>> tclvalue(.Tcl("font actual TkDefaultFont -family")) >>>>>> >>>>>> and sets the font for commands and output by querying the default Tk >>>>>> fixed-width font: >>>>>> >>>>>> tclvalue(.Tcl("font actual TkFixedFont -family")) >>>>>> >>>>>> On my Macs, these are respectively the helvetica and courier fonts. >>>>>> >>>>>> In any event, I'm glad that you seem to have solved your problem. >>>>>> >>>>>> John >>>>>> >>>>>>> -----Original Message----- >>>>>>> From: Jose G Conde Santiago [mailto:jose.con...@upr.edu] >>>>>>> Sent: Monday, May 15, 2017 3:18 PM >>>>>>> To: Simon Urbanek <simon.urba...@r-project.org>; Fox, John >>>>>>> <j...@mcmaster.ca>; r-sig-mac@r-project.org >>>>>>> Subject: Re: [R-SIG-Mac] R commander fonts on R 3.4.0 >>>>>>> >>>>>>> When I run R without XQuartz, the text from the help screen renders >>>>>>> well. I think the problem seems to be related to XQuartz. >>>>>>> >>>>>>> José >>>>>>> >>>>>>> >>>>>>> José G. Conde, MD, MPH >>>>>>> Professor, School of Medicine >>>>>>> Director, CentIT2 >>>>>>> UPR Medical Sciences Campus >>>>>>> >>>>>>> Tel (787) 763-9401 Fax (787) 758-5206 >>>>>>> >>>>>>> Email: jose.con...@upr.edu >>>>>>> >>>>>>> URL: http://rcmi.rcm.upr.edu >>>>>>> >>>>>>>> On May 15, 2017, at 2:16 PM, Jose G Conde Santiago >>>>>>> <jose.con...@upr.edu> wrote: >>>>>>>> >>>>>>>> … and the same I get the same issue with R “help” screens in my iMac. >>>>>>> I cannot reproduce this in the macBook. >>>>>>>> >>>>>>>> José >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> <R help .png> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> José G. Conde, MD, MPH >>>>>>>> Catedrático, Escuela de Medicina >>>>>>>> Director, CentIT2 >>>>>>>> Recinto de Ciencias Médicas, UPR >>>>>>>> >>>>>>>> Tel (787) 763-9401 Fax (787) 758-5206 >>>>>>>> >>>>>>>> Correo electrónico: jose.con...@upr.edu >>>>>>>> >>>>>>>> URL: http://rcmi.rcm.upr.edu >>>>>>>> >>>>>>>>> On May 15, 2017, at 2:14 PM, Jose G Conde Santiago >>>>>>> <jose.con...@upr.edu> wrote: >>>>>>>>> >>>>>>>>> This is the actual resolutions. >>>>>>>>> >>>>>>>>> José >>>>>>>>> >>>>>>>>> <Fonts 20 points.png> >>>>>>>>> José G. Conde, MD, MPH >>>>>>>>> Professor, School of Medicine >>>>>>>>> Director, CentIT2 >>>>>>>>> UPR Medical Sciences Campus >>>>>>>>> >>>>>>>>> Tel (787) 763-9401 Fax (787) 758-5206 >>>>>>>>> >>>>>>>>> Email: jose.con...@upr.edu >>>>>>>>> >>>>>>>>> URL: http://rcmi.rcm.upr.edu >>>>>>>>> >>>>>>>>>> On May 15, 2017, at 12:31 PM, Jose G Conde Santiago >>>>>>> <jose.con...@upr.edu> wrote: >>>>>>>>>> >>>>>>>>>> <Fonts 20 points.png> >>>>>>>>> >>>>>>>> >>>>>> >>>>>> _______________________________________________ >>>>>> R-SIG-Mac mailing list >>>>>> R-SIG-Mac@r-project.org >>>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac >>>>> >>>>> -- >>>>> Peter Dalgaard, Professor, >>>>> Center for Statistics, Copenhagen Business School >>>>> Solbjerg Plads 3, 2000 Frederiksberg, Denmark >>>>> Phone: (+45)38153501 >>>>> Office: A 4.23 >>>>> Email: pd....@cbs.dk Priv: pda...@gmail.com >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>> >>>> -- >>>> Peter Dalgaard, Professor, >>>> Center for Statistics, Copenhagen Business School >>>> Solbjerg Plads 3, 2000 Frederiksberg, Denmark >>>> Phone: (+45)38153501 >>>> Office: A 4.23 >>>> Email: pd....@cbs.dk Priv: pda...@gmail.com >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>> >>> -- >>> Peter Dalgaard, Professor, >>> Center for Statistics, Copenhagen Business School >>> Solbjerg Plads 3, 2000 Frederiksberg, Denmark >>> Phone: (+45)38153501 >>> Office: A 4.23 >>> Email: pd....@cbs.dk Priv: pda...@gmail.com >>> >>> >>> >>> >>> >>> >>> >>> >>> >> >> -- >> Peter Dalgaard, Professor, >> Center for Statistics, Copenhagen Business School >> Solbjerg Plads 3, 2000 Frederiksberg, Denmark >> Phone: (+45)38153501 >> Office: A 4.23 >> Email: pd....@cbs.dk Priv: pda...@gmail.com >> >> >> >> >> >> >> >> >> > > _______________________________________________ > R-SIG-Mac mailing list > R-SIG-Mac@r-project.org > https://stat.ethz.ch/mailman/listinfo/r-sig-mac _______________________________________________ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac