Small things, mostly.
1. Buildproblem. Well-known and annoying. Build fails if debugging is on
(which is the default) because Debugging.h breaks devQuartz.c
compiles. If debugging is switched off on the command line or in
the environment, then devQuartz.c builds, but aquaconsole.c does
not compile, because it needs Debugging.h. I do not want to edit
Apple system files, and not files in the R distribution either.
2. It would be useful if quartz devices could be opened from a
menu.
3. Maybe a primary prompt in the input window would be useful.
4. Input and output windows cannot be resized independently. I
would still vote for two separate input and output windows.
5. Interaction with Tcl/Tk is still problematic. For instance, Tcl/Tk
takes over the menubar, and if you don't use tkStartGUI() the
tk windows don't respond to mouse clicks. For instance, try
using Rcmdr. Also if you use tkStartGUI(), and AquaTk has
taken over, it creates windows (for instance for downloading CRAN
packages) without scrollbars. This is important, because Tcl/Tk
is still the only possible way of building GUI applications
from RAqua.
6. If you want to download from CRAN you select packages in
a window, and the download starts if you close the window.
That's not intuitive, there should be a separate download or go
button.
7. I thought I saw somewhere that RAqua can now open X11
devices. Mine can't.
8. Installation should be in ~/Applications, or in some other
user defined position. My /Applications only has Apple
software.
9. I see that there already is an "R Evaluate" service advertised.
Great. Doesn't work yet, though.
==Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
Editor: Journal of Multivariate Analysis, Journal of Statistical
Software
US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu
homepage: http://gifi.stat.ucla.edu
------------------------------------------------------------------------
-------------------------
No matter where you go, there you are. --- Buckaroo Banzai
http://gifi.stat.ucla.edu/sounds/nomatter.au
Jan de Leeuw <deleeuw@stat.ucla.edu> writes:> 5. Interaction with Tcl/Tk is still problematic. For instance, Tcl/Tk > takes over the menubar, and if you don't use tkStartGUI() the > tk windows don't respond to mouse clicks. For instance, try > using Rcmdr. Also if you use tkStartGUI(), and AquaTk has > taken over, it creates windows (for instance for downloading CRAN > packages) without scrollbars. This is important, because Tcl/Tk > is still the only possible way of building GUI applications > from RAqua.tkStartGUI is just a proof of concept and many parts of it are sketchy. It doesn't create scrollbars on unix either, but you can scroll with the middle button (now how do you do that with a single button mouse?). It would be easy to add the scrollbar, I just don't think this is due to bad interactions with Aqua in particular. Starting yet another console to get Tk windows to respond is just wrong. Presumably, the case is that nothing is running the Tk event loop and that tkStartGUI executes a "wait variable" which runs the loop by itself. It should be fairly easy to wedge a tk event handler into the Aqua loop (disclaimer: I don't have access to a Mac, I'm just assuming that Aqua works like other GUI toolkits and allows you to define idle handlers and the like). -- O__ ---- Peter Dalgaard Blegdamsvej 3 c/ /'_ --- Dept. of Biostatistics 2200 Cph. N (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907
On Mercoled?, ago 13, 2003, at 04:44 Europe/Rome, Jan de Leeuw wrote:> Small things, mostly. > > 1. Buildproblem. Well-known and annoying. Build fails if debugging is > on > (which is the default) because Debugging.h breaks devQuartz.c > compiles. If debugging is switched off on the command line or in > the environment, then devQuartz.c builds, but aquaconsole.c does > not compile, because it needs Debugging.h. I do not want to edit > Apple system files, and not files in the R distribution either.Not a small thing this one :) I'll substitute the code inside aquaconsole & C. that conflicts with Debuggin.h and add then use -D__DEBUGGING__> > 2. It would be useful if quartz devices could be opened from a > menu.ok> > 3. Maybe a primary prompt in the input window would be useful. > > 4. Input and output windows cannot be resized independently. I > would still vote for two separate input and output windows. > > 5. Interaction with Tcl/Tk is still problematic. For instance, Tcl/Tk > takes over the menubar, and if you don't use tkStartGUI() the > tk windows don't respond to mouse clicks. For instance, try > using Rcmdr. Also if you use tkStartGUI(), and AquaTk has > taken over, it creates windows (for instance for downloading CRAN > packages) without scrollbars. This is important, because Tcl/Tk > is still the only possible way of building GUI applications > from RAqua. >> 6. If you want to download from CRAN you select packages in > a window, and the download starts if you close the window.it starts only if you check the item "install", otherwise the R session continues giving you a prompt.> That's not intuitive, there should be a separate download or go > button. > > 7. I thought I saw somewhere that RAqua can now open X11 > devices. Mine can't.you should start QuartzX and use x11(display=":0")> > 8. Installation should be in ~/Applications, or in some other > user defined position. My /Applications only has Apple > software.We can manage this through the installer probably. The problem is that Apple bundles seem to be too much rigid, I will think about it.> > 9. I see that there already is an "R Evaluate" service advertised. > Great. Doesn't work yet, though.emh, never implemented! It is strange you see anything as services.> ==> Jan de Leeuw; Professor and Chair, UCLA Department of Statistics; > Editor: Journal of Multivariate Analysis, Journal of Statistical > Software > US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554 > phone (310)-825-9550; fax (310)-206-5658; email: > deleeuw@stat.ucla.edu > homepage: http://gifi.stat.ucla.edu > > ----------------------------------------------------------------------- > -------------------------- > No matter where you go, there you are. --- Buckaroo Banzai > http://gifi.stat.ucla.edu/sounds/nomatter.au > > ______________________________________________ > R-devel@stat.math.ethz.ch mailing list > https://www.stat.math.ethz.ch/mailman/listinfo/r-devel >
Dear Jan and Peter, Given Peter's explanation, setting options(Rcmdr=list(tkwait=TRUE)) before loading the Rcmdr package might allow the Commander to work (although it will disable the R command prompt while the Commander window is open). Not a good solution, of course. Regards, John At 09:05 AM 8/13/2003 +0200, Peter Dalgaard BSA wrote:>Jan de Leeuw <deleeuw@stat.ucla.edu> writes: > > > 5. Interaction with Tcl/Tk is still problematic. For instance, Tcl/Tk > > takes over the menubar, and if you don't use tkStartGUI() the > > tk windows don't respond to mouse clicks. For instance, try > > using Rcmdr. Also if you use tkStartGUI(), and AquaTk has > > taken over, it creates windows (for instance for downloading CRAN > > packages) without scrollbars. This is important, because Tcl/Tk > > is still the only possible way of building GUI applications > > from RAqua. > >tkStartGUI is just a proof of concept and many parts of it are >sketchy. It doesn't create scrollbars on unix either, but you can >scroll with the middle button (now how do you do that with a single >button mouse?). It would be easy to add the scrollbar, I just don't >think this is due to bad interactions with Aqua in particular. > >Starting yet another console to get Tk windows to respond is just >wrong. Presumably, the case is that nothing is running the Tk event >loop and that tkStartGUI executes a "wait variable" which runs the >loop by itself. It should be fairly easy to wedge a tk event handler >into the Aqua loop (disclaimer: I don't have access to a Mac, I'm just >assuming that Aqua works like other GUI toolkits and allows you to >define idle handlers and the like). > >-- > O__ ---- Peter Dalgaard Blegdamsvej 3 > c/ /'_ --- Dept. of Biostatistics 2200 Cph. N > (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 >~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk) FAX: (+45) 35327907 > >______________________________________________ >R-devel@stat.math.ethz.ch mailing list >https://www.stat.math.ethz.ch/mailman/listinfo/r-devel----------------------------------------------------- John Fox Department of Sociology McMaster University Hamilton, Ontario, Canada L8S 4M4 email: jfox@mcmaster.ca phone: 905-525-9140x23604 web: www.socsci.mcmaster.ca/jfox
By disabling the require_action statements in aquaconsole.c
and by setting CPPFLAGS to -D__DEBUGGING__ the compile
actually happens. But StartR does not work, because of
undefined symbols.
Fatal error: The AQUA shared library could not be loaded.
The error was dlcompat: dyld:
/Applications/StartR.app/RAqua.app/Contents/MacOS/RAqua Undefined
symbols:
_check_noerr
_require_noerr
Oh yes, and another small gripe: double clicking RStart opens
the R console behind all other open windows. Unobtrusive, to
be sure.
==Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
Editor: Journal of Multivariate Analysis, Journal of Statistical
Software
US mail: 9432 Boelter Hall, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550; fax (310)-206-5658; email: deleeuw@stat.ucla.edu
homepage: http://gifi.stat.ucla.edu
------------------------------------------------------------------------
-------------------------
No matter where you go, there you are. --- Buckaroo Banzai
http://gifi.stat.ucla.edu/sounds/nomatter.au