>> lately, and it is a challenge. Just parsing tcl, with its
>> various circles of quoting hell, can be tricky. There are
> I assume this means you are trying to reimplement tcl parsing rather
> than use the already developed Tcl parsing code available in either
> the tcl core or the Jacl project.
That's right. As a first try, I tried to construct a grammar from the
manpages and implement a recursive descent parser. It works for most
Tcl/Tk code, and I learned a lot about Tcl quoting and evaluation.
To iron out the rest of the bugs, I think I will have to dive into
the Tcl core and emulate the parser more directly. Perhaps tcl, like perl,
is best parsed with the interpreter code itself :)
>> Couple
>> this with the fact that tcl/tk is quite the moving target
>> these days,
> This one statement has me puzzled . Has there been some change in the
> past 4 years or more to Tcl syntax? Or are you referring to the fact
> that Tcl continues to get new commands implemented? Or something else
> that I am missing completely?
There have been few changes in the Tcl syntax recently (namespaces come to
mind), but there have been lots of changes in details, mostly additional
commands and additional options to commands. Thus the parser may not need
to change much, but the translation backend needs to be updated and tested.
For 8.3, there are 148 Tcl and Tk .n manpages, vs. 116 in 7.4/4.0 four
years ago.
-Mark
---------------------------------------------------------------------------
To unsubscribe from the Visual Tcl mailing list, please send a message
to [EMAIL PROTECTED] with "unsubscribe vtcl [EMAIL PROTECTED]" in the
message body (where [EMAIL PROTECTED] is your e-mail address).
