Calling all experts

Do you have a niche area you like to think you're pretty top-notch in?  Might be a library, a user application, etc.  You're one of the top few people in the world who could be pulled for proficient work in this area, no matter how silly it is.

For me, I like to think I am one of the few people with a strong grasp on the automation aspects of the IRC protocol.  Sounds like a joke, but you've got to consider that despite numerous RFCs, IRC really has no protocol at all.  A lot of IRC libraries don't get it right, and for good reasons... IRC is meant for human consumption, and is poorly suited to parsing and state tracking.  A simple example of its many inconsistencies: you join a RFC1459-strict server and join a channel, where bob has +ov.  You only see the +o when you join, but never see bob's +v.  I think expressing this in the IRC library (that it is unknown whether bob has +v) is a great way to deal with mode ambiguity when writing IRC bots, but it seems I'm the only one.

So yeah, it's an area you're a pedant in, probably.  A self-absorbed ass nugget.

>>40
Where do I get the .tpu to make signatures out of
Once you've identified the version of the libraries in use, you should use FLAIR (they're utilities which should be distributed separately from IDA, but officially supported) to create the signatures, to actually parse TPU files, you need to use the tool located in pascal\bin\* , read the text file for actual instructions, it should support TP 4.0-7.0 and Delphi 1.0. There's a bunch of other helpful scripts in the pascal dir that you might want to look at. The tool should output a PAT file, which you can then use to generate FLIRT signatures, like you would with anything else (sigmake). It's pretty simple stuff, and you would have found out how to do this, if you would have read the documentation.

P.S.: IDA already should have FLIRT signatures for Turbo Pascal 5.5-6, it's called tpdos.sig, so next time, look more carefully at what the program provides before looking to make your own.

P.P.S.: AFAIK, Turbo Pascal's compiler was quite primitive, and a decompiler for it was feasible, if you like wasting your time, you could write one.