Re: LYNX-DEV Re: <pre></pre> <Hx></Hx> BUG

[Filip M Gieszczykiewicz]

You (Chris Gray) wrote:
> On Sun, 5 Apr 1998, Filip M Gieszczykiewicz wrote:
> > You (Jason F. McBrayer) wrote:
> > > FMG> I'm getting screwed because I can't script this crap, and the
> > > FMG> only winner is the happiness of some really anal validator that
> > > FMG> no one else, it seems, uses or cares that it exists.
> 
> You'll have to explain this.  I use scripts to produce most of my HTML
> these days, and for me one of the great advantages of doing things that
> way is that I know my scripts produce valid HTML - I hardly need to worry
> about HTML synax any more.
> 
> If you have a script which consistently turns out bad HTML (such as 
> <pre>...<h1>...</h1>...</pre>) then surely you must be able to make one
> which consistently turns out good HTML (such as <pre>...</pre><h1><tt>...
> </tt></h1><pre>...</pre>, or whatever it is you want).

Over 80% of the articles on our site are written by someone else. That
person sends me updates every few months (These are changes that I can
not "diff" into existing HTML because changes are so small but numerous)
I **USED** to have enough time to: extract all links and format them into a
sed search'n'replace file, load the file into lynx, dump it as text, run
it through my txt2HTML converter, then go by hand and convert it into
my format (so it came out nice and pretty later on), then run it through
my scripts which would extract and build a table of contents [W], then 
through a cross-referencer script to fix all the "for more info see section:
blah blah", then run it through sed to re-insert all the links where they
used to be, segment the files [W] (many are over 500KB... not fun on a slow
link, scripts munch something like that into 25-30Kb chunks and inter-link),
then run it through a fix-up which would convert an additional layer of
cross-references (only valid at this point), tar it, run the glimpse on the
results directory, and upload the new index files. This took more than a
day(s) of hard work... (thanks to linux'es flawless multitasking and VM :-)

[W] means run it through weblint at this stage.

Times have changed. I'm in college now. Time is at a premium.
I get the files from him - often peppered with a &lt; here and
there... (but the > not alted), links in all kinds of odd places...
All the errors relate to bad markup... And can be fixed with a
few hours of hacking... (hours? I need to make sure it doesn't trip
up all the other layes of kludges and hacks ;-) which I do not
have at this moment. Trust me, I _always_ have something better
to do than do what doesn't _need_ to be done just _would be nice_.

We're getting off the topic of this group. Never mind. I'll get
pissed off _enough_ one of these days and fix it RealGood(tm).
Today is my last "check e-mail" day... next week is finals week!

> IOW, your script is broken; fix it.

Script works fine... been working fine for years. New situation:
Garbage in, garbage out :-)

Take care.

P.S. I spent most of today hacking on minicom... the $%^# thing screw
up history buffer... it seems to only "save" only when the screen is
scrolled or completely erased... show me an _application_ that does
that! None. It happily trashes screenfulls of info that I want to kill
it for. If you've ever seen minicom's source, you'd see how much fun
it is to fix something that deals with these "odd" problems. Ie. I'm
not in my _best_ of moods.