2018-04-17 2:14 GMT+02:00 Steve Landers:
> I am wondering if anyone has presented arguments against Rolf’s alternative> proposal of making Tcl_UniChar 32 bit. Especially if Androwish does build> with 32-bit Tcl_UniChar by default.>> If not, then I suggest a vote on TIP 389 is premature.
Well, just have a look at the changes made in Androwish to make this
work. My first objection to this (for 8.7) is the binary incompatibilty.
At this moment, Tcl 8.6 supports TCL_UTF_MAX=3, and it
kind of tolerates TCL_UTF_MAX=4 and TCL_UTF_MAX=6.
After TIP #389, TCL_UTF_MAX=4 will be the official supported
configuration while TCL_UTF_MAX=6 is tolerated. I am in
favor of backporting as much as reasonable from androwish
to make TCL_UTF_MAX=6 an official supported configuration too.
I already created an "androwish" branch for that in core.tcl/tk/tcl.
Unfortunately, this branch contains a lot more changes than
needed for TCL_UTF_MAX=6. Still a lot of work to do.
> There is even an up to now legal (or at least working) use case of> Tcl_UtfToUniChar and friends, that will break after this core change:> converting an UTF-8 string to a Tcl_UniChar array for random index> access (or simple looping over the characters).
This is simply not true. Please show me a test-case demonstrating
the problem.
> The tip has rumors about problems with that ("a lot of additional fixes> are needed to make it work right"). Can be someone please a bit more> specfic about the unsolved problems with that approach?
So, have a look at the "androwish" branch.
Nothing in TIP #389 stands in the way of further progress in
the direction of TCL_UTF_MAX=6.
Regards,
Jan Nijtmans
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Tcl-Core mailing list
Tcl-...@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/tcl-core