tweaks: remove some ineffectual parts from header-line regexes

Adding "[abc]*" does not restrict the recognized header line in any way.

Also, improve the header-line regex for shell scripts, because it should
not match "barunscript" (for example).

syntax/perl.nanorc

@@ -1,7 +1,7 @@
 ## Here is an example for Perl.
 
 syntax perl "\.p[lm]$"
-header "^#!.*perl[-0-9._]*"
+header "^#!.*perl"
 magic "Perl script"
 comment "#"
 

syntax/python.nanorc

@@ -1,7 +1,7 @@
 ## Here is an example for Python.
 
 syntax python "\.py$"
-header "^#!.*python[-0-9._]*"
+header "^#!.*python"
 magic "Python script"
 linter pyflakes
 comment "#"

syntax/ruby.nanorc

@@ -1,7 +1,7 @@
 ## Here is an example for Ruby.
 
 syntax ruby "\.rb$"
-header "^#!.*ruby[-0-9._]*"
+header "^#!.*ruby"
 magic "Ruby script"
 linter ruby -w -c
 comment "#"

syntax/sh.nanorc

@@ -1,7 +1,7 @@
 ## Here is an example for Bourne shell scripts.
 
 syntax sh "(\.sh|\.(a|ba|c|da|k|mk|pdk|tc|z)(sh|shrc|sh_profile)|/(etc/|\.)profile)$"
-header "^#!.*(a|ba|c|da|k|mk|pdk|tc|z)?(sh[-0-9_]*|openrc-run|runscript)"
+header "^#!.*/(((a|ba|c|da|k|mk|pdk|tc|z)?sh)|openrc-run|runscript)"
 magic "(POSIX|Bourne-Again) shell script.*text"
 linter dash -n
 comment "#"