Doors CSE 8 is the first and currently only assembly shell for the TI-84 Plus C Silver Edition, TI's flagship color z80 calculator. Built on the Doors CS 7 codebase, it lets you run any kind of TI-BASIC or assembly program, run and edit archived programs, use extra functions in TI-BASIC programs, and offers myriad additional features to make the TI-84+CSE more powerful. The most recent big update to Doors CSE was Doors CSE 8.1 in July: it brought new Hybrid BASIC functions, the ability to list and run Applications, and many bugfixes. Doors CSE 8.1.1 was subsequently released in August, and now Doors CSE 8.1.2 is available. These new updates bring the following fixes:

• Doors CSE 8.1.1: The InsertLine Celtic 2 CSE function no longer has the potential to insert invalid data when used on or past the last line of a file.
  
• Doors CSE 8.1.2: The ColorPixel ASM routine now reads the color for the drawn pixel correctly. In addition, in a major boon to TI-BASIC math programmers, all real() calls in normal TI-BASIC programs will correctly return the real part of the argument, regardless of the value entered. Previously, certain small positive integer values were incorrectly interpreted as Hybrid BASIC function calls. This change ensures that Doors CSE will not interfere with any normal non-Hybrid TI-BASIC program, including those that work with complex numbers.

Unless additional bugs or compatibility issues are discovered, the next Doors CSE release will be Doors CSE 8.2. There have already been many requests for additional Hybrid BASIC functions, and any functions that fit and are deemed widely useable will be included in Doors CSE 8.2. In addition, it is likely that CALCnet 2.2 will be included in the next Doors CSE release.

Download
Doors CSE 8.1.2

This is fantastic news. Will you be back-porting this to DCS 7?

Yes, I will be. However, the fix is a little more complicated for xLIB non-C, because of the differences in which single-argument real() commands exist. In xLIBC (and therefore Doors CSE , it doesn't make programmatic sense to call a single-argument Hybrid real() function before the first multi-argument real(). Therefore, the first real() command that has more than one argument is used as an indication that the program uses Hybrid libraries, at which point Doors CSE starts processing all real() commands as potentially xLIBC functions. In xLIB (and therefore Doors CS 7), there are single-argument real() functions that make sense to call before any multi-argument real()s. I suspect I should try my fix as-is, and look through the existing xLIB-using programs to see if any of them use these single-argument xLIB functions before multi-argument functions.

I really should get the CSE, but alas, my money just ran out leaving me with a measly $10 for emergencies. The ColorPixel routine intrigues my friends who want to learn z80 asm with their new calcs, so congrats on that too!

Is there anything set when a program is run? I'm asking because would it be easier to have something like that happen, so it knows when to run the xlib commands, so you could enable/disable real( from working based on whether a program has been run, and disabled upon DCS regaining control?

Nice to see the problem being addressed for those that want to use that real( function for what it was intended for

I'm not sure that I quite follow what you're suggesting there. The problem is not that real() outside of programs gets intercepted: that never happens. Neither Doors CS nor Doors CSE catch real(), det(), identity(), or sum() on the homescreen or elsewhere outside programs. The problem is programs that use real() to work with complex numbers.