I am exploring the possibilities of using *Tkinter* for a serious application mainly devoted to the handling of pure text.
 
As en exercise I have tried to rewrite in *Python* an interesting example that is unusually included in the manual page of *Tk* for the *font* component (*Tcl8.5.7/Tk8.5.7 Documentation > TkCmd > font > example*); this code works fine under *tclsh* and *wish*. What is does is getting a list of the font names available in the machine in which it is running and displaying nicely each of its items with a sample of the corresponding font.
 
I have performed most of my purpose and my *Python* script works. Along the way have learned a lot about *Tkinter/Tk/Tcl*, which was also a desired goal of the exercise. But the script does not work completely well, as I have also hit against an obstacle that I do not know how to overcome and seems to be rather deep. It is the following:
 
Some of the names of the font families returned by the operating system (be it Windows or Linux) are composed of two or more words (like *Times New Roman CE*). Now these multiword arguments are not easily accepted by *Tcl/Tk*, especially if they are referenced by some variable, so that for instance tags cannot be reconfigured this way in order to apply such fonts; or the size of sentences using them cannot be measured, two operations that are essential for the procedure.
 
I have learned that this is a real and inherent difficulty with *Tcl* (Harrison and McLennan, *Effective Tcl/Tk programming*: 34, 66); this is why the original code uses such a strange syntax like
 
    .t tag configure f[incr count] -font [list $family 10]
 
The function *list* of *Tcl* seems to be mainly dedicated to solve this odd problem with the dereferencing of multiword contents of variables (*family* in this case, which holds the name of the font). But apparently this does not work when *Tcl/Tk* are finally moved by *Tkinter* during the execution of the script or at least I have been unable to find a way to make it work in due form.
 
Which does not mean much, as I am not a programmer or, at most, I am a very unseasoned one. Can somebody who has had the same problem and solved it succesfully from *Tkinter* offer me (and perhaps us) his insight and describe the essentials of how to do it?
 
Many thanks in advance!

Many thanks to you both, Kevin and Michael for your quick and sensible responses. My problems are more in the line suggested by Michael, although his hint has not solved them, at least as I have tried to implement it.

In order for you to have all the information together, I transcribe the original Tcl code and my *porting* of it to *tkinter*. I commented it for my own use, in the process of conjecturing and confirming what the original code was intended to perform.

If you run the original code with *tclsh* or *wish* and then my code with *Idle*, you will see very clearly where the problem lies.

Kind regards

Francisco

# original Tcl code
# offered as example in the *font manual page* of TkDocs.
# Runs without problems with *tclsh* and *wish*

# pack [text .t -wrap none] -fill both -expand 1
# set count 0
# set tabwidth 0
# foreach family [lsort -dictionary [font families]] {
#    .t tag configure f[incr count] -font [list $family 10]
#    .t insert end ${family}:\t {} \
#            "This is a simple sampler\n" f$count
#    set w [font measure [.t cget -font] ${family}:]
#    if {$w+5 > $tabwidth} {
#        set tabwidth [expr {$w+5}]
#        .t configure -tabs $tabwidth
#    }
# }


from tkinter import *
from tkinter import font


# creation of the main window
master = Tk()

# creation of a text widget
tw = Text( master )
tw.config( wrap = NONE )
tw.pack( fill = BOTH, expand = 1 )

tw.insert( END, "Display of available fonts:\n\n" )

count = 0
tabwidth = 0

# identification of the font being used
f1 = tw.cget( 'font' )

# a measurement function used later
# is a method of class *tkinter.font.Font*,
# so that it is good to have the required instance of it
fo = font.Font( font = ( f1, 10 ) )

# let us get the ordered list of all
# available font families
sampler = sorted( list( font.families() ) )

# handling and display of the elements
for f in sampler :
    # the counter is updated
    count += 1
    # display of the provided font name plus
    # a separator and a tabulator
    tw.insert( END, f + ":\t" )
    # creation of an individual name for the tag
    # that will be applied to the sample for this item
    ftag = f + str( count )

    # my (vain) efforts to solve the multiword problem
    if ' ' in f :
#        fm = "('" + f + "', '10')"   # as for Michael's hint
        fm = '[list ' + f + ']'
#        fm = f.replace( ' ', '_' )
#        fm = f.replace( ' ', '' )
        tw.tag_config( ftag, font = ( fm, 10 ) )
#        tw.tag_config( ftag, font = fm )
    else :
        # single word names work as expected
        tw.tag_config( ftag, font = ( f, 10 ) )
    # and the sample text:
    tw.insert( END, "This is a simple sampler\n", ftag )
   
    # nice columnar disposition by
    # a clever configuration of the tabulator
    w = fo.measure( f + ':' )
    if ( w + 5 ) > tabwidth :
        tabwidth = w + 5
        # this equalizes all the lines as required
        tw.config( tabs = tabwidth )

if __name__ == '__main__' :
    master.mainloop()