Another year down. I remember thinking about how the years after 2000 would be (during my school days). A couple of days ago 4 of my relatives appeared online (all at once!), which was quite surprising to me. I had not seen them online for long (some like ‘ever’).

The absence of any messing with code at home is troubling me. Last week I tried out trigger. It did not work out-of-the-box. First it failed to start complaining about needing OpenGL 1.1 or below (Actually it said it wont for OpenGL > 1.1). After hard wiring that one came the terrain problem – segfault. It took some time to figure out that glDrawRangeElements was failing but glDrawElements worked. The most of last weekend was with playing that one – The physics is good (actually (too?) real). I got really worked up with one of the levels (the ‘Pulp: Cars can fly’ available at trigger homepage).

This week I wanted to try chaning the sun thingy for flightgear. I went right into the source code and found that no related code is there !! Most of it has been moved to Simgear. Now flight gear source just has flight sim stuff (kinda cool). The biggest discovery was the nasal thingy in Flightgear. FGFS has a virtual machine now (just like all those pro games – ID FPS family, Ultima (usecode)) and is in Simgear.

Cant see any way out. The bios does not have anything remotely resembling AMDK7NOW! which the driver is looking for. (*sigh*).

Emacs: I’ve split the emacs configuration into various files which are loaded by the .emacs. If your car were emacs : here.

Also met GNU ID Utils while trying to find how the people work with crossrefernce in source trees (that of linux kernel, say). I use etags at work, but it gives false jumps at times. I guess I’ve to look at the command line options a bit closer.

There is an emacs interface (a bit thin one, though) to call the gid program (to query the ID database) giving it any string to look up. The way I see it, its effectively indexing the find - grep combination. More reading on the ID utils remain. If it is capable of differentiating definition and reference of various language entities it would be cool.

Just found this variable that should fix the problem. Quote
the help on mml-insert-mime-headers-always:

mml-insert-mime-headers-always's value is t

Documentation:
If non-nil, always put Content-Type: text/plain at top of empty parts.
It is necessary to work against a bug in certain clients.

You can customize this variable.

Defined in `mml'.

Looks like, in emacs, solutions just wait there to be discovered.

Recent Emacs thingies

I started using gnus at work.

Started using emacs-snapshot-gtk package (from Debian). And a win32 build of emacs 22.0.50 at work.

Wrote some elisp to send multipart alternative (for html mail and plain text). This
needs some fixing though; It does not put the content-type for the plain text part and
many mail clients would not would not show the plain-text content.

Learned about putting window configuration in registers (I could not stop telling
people about it for a couple of hours). Now I don’t use frames (at-least I for the reason
I had before this piece of learning).

I am all sold out on Emacs.

HTML Mail

BTW, do not think that I am for HTML mail. Its just that the M$ outlook and many other clients (and Gmail) show plain text in variable width fonts (when I want fixed width font to show off my ascii-art signature). They say HTML mail is rude.

Recent findings

Just found that the ‘netiquette guidelines‘ is an RFC ! and that (mail) clients are supposed to
assume the ‘default’ Content-Type (which is text/plain with us-ascii encoding) as per
RFC 2045 (Sec 5.2 – let me know if I did not read it enough). I wonder why that Thunder-bird was not showing the plain text alternative in my gnus made mail. The only problem that is apparent is the absence of Content-Type.

OT: Scott Adam’s blog is here.