« on: September 15, 2015, 08:51 PM »
I think it was OpenVMS Language Sensitive Editor had its own pattern matching scheme. A percent sign matched any character is the only one I remember. I wonder if they used regex internally or rolled their own.
The (Open)VMS Language Sensitive Editor (LSE), along with EVE (Extensible VMS Editor) and my own KED (Keypad Editor), all rolled their own pattern matching mini-languages/syntax, as these all were developed well before (or at least in parallel with) regular expression (regexp) syntax became common or even popular. LSE, EVE and KED were each written in a very interesting (alas, proprietary for VMS) text processing language called TPU (unimaginatively, Text Processing Utility).
TPU was (still is) a very powerful, high-level, text manipulation programming language, with data structures and operations to handle unlimited text buffers, arbitrarily-sized display windows, select (and other) regions and ranges, all necessary primitive operations on characters and strings (up/low-case, trims, concatenation, substrings, etc), and a very powerful and general string/pattern matching sub-language (functions).
Developing an application (like a specific text editor) in TPU is much like working with a dynamic interpreted language like Ruby or Python; however, TPU programs (scripts) could be either run via the interpreter, or could be "compiled" into a byte-code like binary file for execution speed. TPU code could "write" or generate TPU code, which it then could directly evaluate, giving the same incredible runtime power that you find in Ruby or Lisp today.
I've not actually done it, but I'm pretty sure that a complete RegExp implementation could be built in TPU. Again, as a tool of its time and its place, the syntax of RegExps hadn't actually occurred to developers then...