diff options
| author | jan.nijtmans <nijtmans@users.sourceforge.net> | 2016-03-23 09:46:40 (GMT) |
|---|---|---|
| committer | jan.nijtmans <nijtmans@users.sourceforge.net> | 2016-03-23 09:46:40 (GMT) |
| commit | fcc11a1aff2224c6ee48e44c28412af1a2af2e3c (patch) | |
| tree | ec81cf655e874570eeb665168ad3c6f2587c103f | |
| parent | b21ca892017fbab11d710b506f9d403584841308 (diff) | |
| parent | 35c8885bad4032db82374bb2f4b1db1ac0839f88 (diff) | |
| download | tcl-fcc11a1aff2224c6ee48e44c28412af1a2af2e3c.zip tcl-fcc11a1aff2224c6ee48e44c28412af1a2af2e3c.tar.gz tcl-fcc11a1aff2224c6ee48e44c28412af1a2af2e3c.tar.bz2 | |
Fix [f1253530cdd83e66]: tcl_wordchars and tcl_nonwordchars are unknown at startup.
For trunk this will not be the complete story, but for core-8-6-branch and core-8-5-branch it will at least make tcl_wordchars and tcl_nonwordchars function as documented. That's the least polish we should do.
See: [https://groups.google.com/forum/#!topic/comp.lang.tcl/ZZ_WwfQdmoE].
People like Eric Brunel, who want the most logical behavior for any Tcl release can simply do:
set ::tcl_wordchars {\w}
set ::tcl_nonwordchars {\W}
| -rw-r--r-- | library/word.tcl | 16 |
1 files changed, 12 insertions, 4 deletions
diff --git a/library/word.tcl b/library/word.tcl index b8f34a5..3e4bc3a 100644 --- a/library/word.tcl +++ b/library/word.tcl @@ -15,12 +15,20 @@ if {$::tcl_platform(platform) eq "windows"} { # Windows style - any but a unicode space char - set ::tcl_wordchars {\S} - set ::tcl_nonwordchars {\s} + if {![info exists ::tcl_wordchars]} { + set ::tcl_wordchars {\S} + } + if {![info exists ::tcl_nonwordchars]} { + set ::tcl_nonwordchars {\s} + } } else { # Motif style - any unicode word char (number, letter, or underscore) - set ::tcl_wordchars {\w} - set ::tcl_nonwordchars {\W} + if {![info exists ::tcl_wordchars]} { + set ::tcl_wordchars {\w} + } + if {![info exists ::tcl_nonwordchars]} { + set ::tcl_nonwordchars {\W} + } } # Arrange for caches of the real matcher REs to be kept, which enables the REs |