Re: Memory Leak?

[Robert Holmgren holmgren@xxxxxxxx]

** Reply to note from xywrite@xxxxxxxx Tue, 6 Apr 1999 07:48:03 -0500

> Your objection to my ci string makes sense, but I've used the left-arrow
>  character in ci as long as I've used XyWrite.

Hmmm. Well, let's go back to your original command. That's the issue.
To be honest, I am totally confused by your differing descriptions. The
LF wildcard (255-192-153, or func NN{25}) is a _DOWN_-pointing arrow! And
when you say "you have to account for DOS's use of carriage return plus
linefeed" -- I haven't a clue what you mean, or what we have to account
for. LF is 10 (0Ah), CR is 13 (0Dh), and CR/LF is 13/10 (0D0Ah).

I like the consistency and reliability of the "official" wildcard for
13+10 lineends. It's the one on func WC, a *red* left-pointing arrow.
Either func WC or a 3-byte Ascii-10 in the KBD file will put this
character on the CMline. The wildcard itself is the 3-byte sequence
255-192-155 (you could probably also put 255-192-155 in the KBD file).

The truth is, this isn't a real wildcard. Neither are the red guillemet
"wildcards", nor LF, nor CR. They stand unambiguously for a single
character, or combination of characters, so there's no "joker is wild"
aspect to these cards; and XyWrite *can* (usually) handle an unbalanced
number of them, or the unprecedented introduction of them on the right
side of a search/replace.

The core difficulty with imaginatively rendered "transliterations" of
actual commands is that nobody knows what anyone else is talking about.
Not really; not _exactly_. It's a problem. You say "LF" -- I don't
know what it means. Now (apparently replacing previous references to
"LF") you say "left-arrow" -- I don't know what that means either. You
say "255-192-155" (or the hex equivalent), then I know what you mean, and
I can reproduce it; XYWWWEB.U2 contains a multitude of tools for dealing
with precise data, or generating precise code -- ANY sequence of bytes, in
any document.

Here's my interpretation of your command
 ci /{qst}WWR{note}][wW][w<]/{qst}[w<]MDRV[w>][wW][w<]M
DNM[w>][wC]{note}[w<]/[XC_][cr|lf]
-nd
XPLeNCODE

You can save this msg to a file, CAll it into XyWrite, issue
"decode", and read (or DeFine and then RUNCODE) the
result of the above code.

Now, even if my interpretation is not 100%, it must be fairly close; and
-- the above command works just fine; on my machine, it executes perfectly
-- no errors, everything properly replaced.

Why garbage is entering your document is the question. It's possible that
your EDITOR.EXE is corrupt. That happens to me about once every six
months, and I just swap it out for a fresh copy. Worth trying.

My experience is, that if you obey the rules, XyWrite obeys the rules; but
that Xy4 can get extremely stressed when you push its parsing capabilities
to the edge. For example, if your separator character "/" happens to
occur in the Wild-W text, it can go bananas. Complicated CI[A]
expressions are fiendishly difficult to parse, and hence to trap, under
XPL; and the relevance of that observation is, that XyWrite's internal
logic closely mimics or parallels the way XPL works -- which is the horse
and which is the cart is
an academic question, but clearly they are similar processes, maybe even
identical internally. So "bug" is obviously a possibility; but somehow, I
don't think so, simply for the reason that it would have happened to me
long long ago.

What else? Yesterday you (I think it was you) wrote, with generous
humanitarianism:

> That aside, there has to be room in the world for people to learn from
> their mistakes. I think that applies to a "virus" that sends itself
> without the knowledge or will of its "victim".

There IS room. Plenty. We suffer a whole lot of idiocy and boring blah
blah blah around here. Tons. But this is different; this carries a real
cost. What kind of logic is it that its MY "responsibility" to ensure
that people haven't sent me dangerous stuff? Is it my responsibility to
ensure that I don't receive letter bombs? Nonsense. My responsibility is
to not SEND that kind of stuff to the list -- not the other way around!
And if a participant is incapable of knowing whether they're sending them
or not, of running a decent Email client, or of guaranteeing their own
benign behavior, then they should be GONE, A.S.A.P. I think that I have
the right to trust people on a subscription list; and if they're
untrustworthy, then the hell with them. This isn't a charity. Any other
closed list would dump them without ceremony; it amazes me that here, in
this ineffably polite world of literate small talk, people just keep
chattering away as if nothing happened. (OK, I admit: Maybe I'm also
expressing some frustration at sitting down and wasting a couple hours on
two weeks' worth of msgs, and finding NOTHING of interest -- nothing new,
nothing challenging, nothing substantial at all. Just a lot of idle chit
chat. Generally it seems to me that the people with the least to offer do
the most insistent talking. And I know that I am far from alone in
feeling that this list has become a huge waste of time over the last two
years or so.)

What possible difference does it make that the Happy worm just happens to
be more or less innocuous? (Not that the descriptions I've read of it
seem all that innocuous! I mean, overwriting your socket layer is not
fun and games.) Would you feel differently if you had to reinstall your
OS? I'll bet you'd be all fired up! Is the person who sent it (whoever
that person is) more or less of an idiot, or more or less dangerous to you
and me, just because he witlessly lucked out this time? That's a weak
argument for toleration.

Bye.


-----------
Robert Holmgren
holmgren@xxxxxxxx
-----------