/Tcl.jl

Optimized Julia interface to Tcl/Tk

Primary LanguageJuliaOtherNOASSERTION

Julia interface to Tcl/Tk

License Build Status

This package provides an optimized Julia interface to Tcl/Tk.

Features

  • As many Tcl interpreters as needed can be started. At least one is always there and serves as the default interpreter.

  • Reading/writing a Tcl variable is as easy as:

    interp[var]             # read Tcl variable value
interp[var] = val       # set Tcl variable value
interp[var] = nothing   # unset Tcl variable

    where interp is a Tcl interpreter, var is the name of the Tcl variable and val is its value.

  • Consistent conversion between Tcl internal representation of values and Julia types. That is to say, evaluating a Tcl script or getting the value of a Tcl variable not necessarily yields a string. For instance, a Tcl float yields a Julia float, a list of integers yields a Julia vector of integers, a list of lists yields a vector of vectors, and so on. Of course, forcing conversion to strings is still possible (and easy).

  • By avoiding systematic string conversion, faster communication with Tcl/Tk is achieved.

  • Tcl objects can be manipulated directly in Julia and may be converted to Julia values (strings, integers, floats or vectors of these).

  • Scripts can be strings but can also be expressed using a syntax which is closer to Julia. For instance, keywords are converted to Tcl options. Scripts can also be built as efficient lists of Tcl objects. Evaluating a script is done by:

    Tcl.eval(script)         # evaluate Tcl script in initial interpreter
Tcl.eval(interp, script) # evaluate Tcl script with specific interpreter
interp(script)           # idem

  • A number of wrappers are provided to symplify the use of widgets.

  • Julia arrays can be used to set Tk images and conversely. A number of methods are provided to apply pseudo-colormaps or retrieve colorplanes or alpha channel. Temporaries and copies are avoided.

  • Julia functions may be used as Tk callbacks.

Alternatives

There exists another Julia Tk package but with different design choices and some issues I wanted to avoid (for instance, X conflict with PyPlot when using Gtk backend, Qt backend is OK). This is why I started this project. I would be very happy if, eventually, the two projects merge.

Installation

Tcl.jl is not yet an offical Julia package but it is easy to install from the REPL of Julia's package manager[pkg] as follows:

pkg> add https://github.com/emmt/Tcl.jl.git

where pkg> represents the package manager prompt and https protocol has been assumed; if ssh is more suitable for you, then:

pkg> add git@github.com:emmt/Tcl.jl.git

To check whether Tcl package works correctly:

pkg> test Tcl

To update to the last version:

pkg> update Tcl
pkg> build Tcl

and perhaps test again...

If something goes wrong, it may be because you already have an old version of Tcl.jl. Uninstall Tcl.jl as follows:

pkg> rm Tcl
pkg> gc
pkg> add https://github.com/emmt/Tcl.jl.git

before re-installing.

[pkg] To switch from julia REPL to the package manager REPL, just hit the ] key and you should get a ... pkg> prompt. To revert to Julia's REPL, hit the Backspace key at the ... pkg> prompt.