After floating around with different free dynamic
host services, I finally registered a name with my own country's TLD
(Top Level Domain), so the new domain for this site is 'v01d.com.ar'.
Hopefully, google will crawl this site (as it didn't with my previous
*.ig3.net domains).
I've been wanting to put my programs into
mirrors as this host is not 100% available (it may be either down or
not reachable because of DNS propagation delays). So added libaiml
(just that for now) into ibiblio.org (but it may not be available yet). I'm going to add this URL to freshmeat's listing also.
I'm in the process of getting a new domain for
this site (altought omicron.ig3.net will still be used), which will
probably be v01d.com.ar. I hope google likes that better.
I've been getting a lot of good feedback about
libaiml. Some were questions, and some were compilation problems.
All issues reported I've already fixed for the upcoming release, which
will be a bugfix-only release. This are the changes from 0.6 to the
future 0.6.1 so far:
- FIX: if glob() returned read error, libaiml would return
'unknown error' (this could happen if an unreadable/unexistant
directory is inside the pattern of files to load)
- made XML error reporting more detailed (file, line and column)
- FIX: test_app would exit if the input was only spaces instead of asking for input again
- FIX:
aiml_parser would crash if the file wasn't found (actually if the
xmlCtxt wouldn't be created) (this would trigger if a <learn> tag
pointed to a non-existant file)
- FIX: <learn> tag wasn't working
- FIX:
patterns (from <pattern>, <topic> and <that>) weren't
cleaned for newlines and characters that shouldn't be in the
grapmhaster's nodes (this could be noticed if there were newlines and
such inside this tags in .aiml files)
- FIX: missing virtual destructor on cInterpreterCallbacks
And keep the feedback + bug reports coming!
I've been working on Imp on my spare time and I'm advancing fast. Perl is the exact right language for this kind of program.
So far I've been able to get 90/5% of the Core completed (it is
missing some little details and I have to think some others. As of the
modules, that will go on a little slower, but I have a Logger module
and a Local/Remote Management System (LRMS) module working. With these
two I can control and test the bot all I want.
I'm taking snapshots (well, crond does) of the repository hosting
Imp, so the sources are available to see in progress. I will probably
make a Developers Release or Release Candidates when I feel that the
Core is completed/stable enough to start worrying about plugins.
Maybe because a teacher during a not-so-good glass
talked bad about Perl and other languages without any objective
arguments, or maybe because another teacher once told that Perl had a
builting design that allowed to consider data as "tainted" which was a
great security feature or maybe because it is a language that
incorporates Regular Expressions (which I love for their power of
expression) straight into the language, I've got very curious about
Perl.
I've always heard about Perl having horrible and obfuscated syntax. I
don't think that's the case. Perl, IMHO, is a language that can satisfy
a lot of different coding styles and ways of thinking programs.
I've started reading O'Reillys' books ("Learning Perl" and "Perl
Programming", aka "The Camel Book") and I still like it =b.
So, considering that Perl is a great language for doing stuff like my
bot does (or even libaiml does, but I wouldn't go that far already),
I'm thinking on porting imp to Perl. I would retain the general idea
and design (modules and such). The core can be a lot more secure and a
lot more simple. Imp modules could be just Perl modules that my bot
would load (using "eval" as I saw other Perl Bots did, but I have to
investigate about that). I think I would enjoy also using the "tainted
data" feature for this.
Maybe libaiml could be written in perl also, but for now I feel that a
C++ aiml interpreter is something that needs to be somewhere (that was
my motivation in the first place).
I've been using subversion for some time (1 year
or more) and I can't stop loving it. I use it to host repositories for
my assignments at faculty (C++, Haskell, ASM, etc projects) and it
really helped a lot. Therefore I always wanted to use subversion for my
own projects hosted on this site. What I'm going to do is add my
projects to the repository and use them internally. I'll set a crontab
or something to generate tarballs for snapshots, and when things go
stable, I'll make a stable tarball.
What I haven't decided yet is if I should open the svn to the world
(right now it uses plain svnserve autentification, I think ssh would be
a little more secure for a public server, which is not needed right
now). Also, I host a WebSVN interface (which also works great) to the
repositories.
So, for now, there will be version control on my projects but it will
be hidden until I decide on how to open them to the world. It's just
not easy to satisfy my paranoias =b
