extended <EN> command and used it to optimize "lexer.test..." macros

 * EN may now be used for matching file names (similar to fnmatch(3)).
   This is used to check the current buffers file extension in the
   lexer configuration macros instead of using expensive Q-Register
   manipulations.
   This halves the overall startup time - it is now acceptable even
   with the current amount of lexer configurations.
 * EN may now be used for checking file types.
   session.tes has been simplified.
 * BREAKS macro portability (EN now has 2 string arguments).
 * The Globber class has been extended to allow filtering of
   glob results by file type.

1 files changed, 7 insertions, 8 deletions

diff --git a/lib/session.tes b/lib/session.tes
index 697c350..09bd647 100644
--- a/lib/session.tes
+++ b/lib/session.tes
@@ -26,12 +26,11 @@ EU[session.path]Q[$SCITECOCONFIG]/.teco_session
 @[session.load]{
   EJ<-EF>
   ! now we have only one unnamed buffer !
-  EQ.x ENQ[session.path]
-  Z"=
-    1EB
-  |
-    :EMQ[session.path]
-  '
+  [_
+    1:EN*Q[session.path]"S
+      :EMQ[session.path]
+    '
+  ]_
 }
 
 ! Call this to set up the session path unique to the
@@ -59,12 +58,12 @@ EU[session.path]Q[$SCITECOCONFIG]/.teco_session
   Requires the SVN client v1.7 or later in $PATH. !
 @[session.svn]{
   ! if there is no SVN working copy, session.path is left as it is !
-  :EG.isvn info --xml"S [*
+  :EG.isvn info --xml"S [*[_
     EQ.i
     :S<wcroot-abspath>"S
       .(S</.-2)X[session.path] :[session.path]/.teco_session
     '
-  ]* '
+  ]_]* '
 }
 
 ! Call to set up the session path unique to the