Jack C Lipton's
WebSite Maintenance
Shell Scripts

My Main page.

Contacting Me
eMail: cupasoup@softhome.net
ICQ#: 174178749


The above cartoon reflects my attitude a little too well.
Hopefully Scott Adams will not mind my use of it here.

Introduction

I use a Linux system at home (based on SuSE Linux) and have allocated an account just for writing. This account maintains a directory as a workspace for editing (and an RCS repository subdirectory as well) while I'm working on stories. I also have a public_html subdirectory which directly mimics my website contents and where updates are pushed from.

I use the make tool to handle pulling content from the RCS repository and managing the transition to readiness for the web site. A helper script exists to build the makefile from a list of files within the RCS repository and to translate the text files into HTML versions (albeit roughly).

The make file is then used to bring in new copies of the text files for constructing the HTML versions. (I've used the make file structures for so long that it's almost second nature to me. Which worries me, come to think of it. I've not been a developer for a while, though, so "make" is no longer the fastest word I can type.) The makefile contains a list of all of the files to be generated which simplifies matters enormously.

From this list of files I build my index.html file to link to all of my content. The .mkindex.sh script does this important task.

The .txt2html.sh script assumes the presence of SHS headers and also assumes that the input file is using blank lines to seperate the paragraphs.

There will be purists telling everybody that I should've used Perl instead of sh and awk and they're not wrong; But, if there's one thing I've learned about Unix (and it's deviants) is that there's usually at least 17 different ways to get a particular job done- and none of them is really "wrong".

Please note that as of 7/12/2003 I have been converting over to another tool I've written (in C) named fmtr which replaces the .txt2html.sh script.

If anyone is really interested in any of these scripts (or wants to make suggestions for improvement) please either e-mail me or post to alt.sex.stories.d ; I'll "steal" any ideas that sound good. :-)

KatieMcN has written a useful website generator of her own that is *much* more aesthetically pleasing even to me but sheer inertia and thick-headedness keeps me using my own tools. (Of course I'm used to them.)

I have not made any effort to see whether this will all work under CygWin or GNUWin2 environments; I'd have to be more interested in running Windoze at all in order to try. Sorry, I'm undermotivated. I've used Unix deviants for so long there's not much point in losing so much functionality.

Basic Generators

• .mkmk
  This script generates the makefile from the contents of the RCS directory, providing the infrastructure.
  It uses two files of interest:
  • .suppress to not pull specific titles from the RCS directory
  • .poem to enable "poem" formatting for the .txt2html.sh script runs
  
  Additionally, there are two other files recently added:
  • .prevlink provides a pair of names containing the "prev file" linkage for a file, to wit:
    cling01:cling00.html
    
  • .nextlink provides a pair of names containing the "next file" linkage for a file, to wit:
    cling01:cling02.html
    
• fmtr.c
  This does pretty much everything that the .txt2html.sh script below does but is written in C.
• .txt2html.sh
  OBSOLETE! This script takes a file (preferably w/ an SHS header) and generates an HTML formatted version.
  Some caveats:
  • The SHS headers really *are* helpful since I throw any other heading away;
  • Paragraphs are defined as having one (or more) blank lines between them.
  (If you're good w/ AWK you can be far more clever.) Note: I have added logic that invokes the button for Lix's Rating/Review allowing feedback to be directly accessible (avoiding the cut/paste kind of stuff).
• .mkindex.sh
  This script generates a story index file. There's some extraneous stuff you may want to edit in/out etc. This breaks up stories into "Universe"-based sections.
• .mkrevchron.sh
  Like .mkindex.sh above, this script generates a story index file; The big difference is that the resultant file is sorted in reverse chronological order based upon the RCS date and time, so it's far simpler in dropping the "sectioning" that the regular index generator has to cope with.
• .mkspotlite.sh
  This script takes a .spotlites file and generates a "nice" HTML file listing the stories and authors (with links) or recommended stories.

Note: RCS is a "Revision Control System" which allows retention of previously edited text (albeit without it being present in the "current" version) so it acts as an archive containing all of the previous versions of a file. Unlike M$'s .doc file format, though, the extracted copy only contains the "current" version of the file. RCS uses the ci and co commands (check-in to save the latest changes, check-out to grab a current or previous version) to access the contents of an RCS archive file. I've gotten to like using it for C programs and various scripts, and, when I started writing, I used it as well simply because a major change didn't necessitate me making a new file (and keeping track of it); I can always hunt back through the history and pull out an old version if I need to steal the text for something else.

Seriously Moronic Material

I have no idea why I'm letting anyone look at this crap. They ain't all that bad at showing how to generate HTML tables and forms, though, so I guess they're not all bad.

• .mksquick.sh
  This script generates my first pass at a "squick" index.
• .mkaff.sh
  This script generates my first pass at a story affinity profile.
• .mkaff2.sh
  This script generates my second pass at a story affinity profile.
• .squicks
  This file feeds the above.