I resolved a few bugs and QoL issues in sporadic coding sessions over the last week or so, and I'm nearly ready to publish a beta. My remaining task before that beta is to expand the test harness to exercise all of the supported equations and operations, plus proper application of PEMDAS.

As promised, a beta for your testing, review, comments, and criticism! I've identified a bug with sqrt() already, and I welcome others: https://www.cemetech.net/downloads/files/2555/x3231

If u try to put x and y in the same trig function it gets VERY angry.

Can you share a sample equation that fails?

*bump* I see Z=cos(X+Y) emits a lot of "MAX EVAL ITERS" errors and ultimately produces an empty graph: is this what you were talking about, the CE guy? This appears related to the bug I identified with sqrt() (in this case, a "symbolic trig not integrated" assertion in the code is triggered, which gives me a good place to start investigating).

Is there any way to / plan to integrate the stats graphing functions of the CE? ( linreg and lines of best fit type of thing)

That's a very good question. I don't currently have any plans to integrate those tools, but if there are some clear curricular reasons why that would be an important addition, I'd certainly consider it. You've inspired me to, for example, consider if I should make the TI EOS variable values (e.g. A-Z and theta, the stats vars) accessible inside of gCAS2 as it computes graph values.

Is it possible to make the actual cube-graph, because I only see the the shapes that I graphed not the actual cube.

Do you mean the bounding cube? Press [WINDOW] to toggle the bounding box and axes on and off.

Thank You!

I'm so excited to download this on my calculator! I've actually been waiting for a fast ce 3d grapher since 2018!! So far, when I tested this on CEMU, f(x, y) = ln(x^2+y^2) returns an error (probably due undefined point at the origin). f(x, y) = e^(x^2+y^2) returns a flat graph. I'm not sure if this is a bug or if I'm not setting my bounds correctly (I just left everything default except changing x and y domains from -1 to 1).

Thanks for testing!

Good guess, but it actually turned out to be the trig (and other) function problem I was previously investigation. This has been repaired:



It seems like I just never set the value of e: this graph doesn't even work on the decade-old Casio Prizm version! Fixed (although predictably, it yields extremely large Z values):



These fixes and others can be found in an updated Beta 2 of Graph3CE. Please do continue testing and let me know if there's anything else you find!

Graph3CE 2.0 Beta 2

Thanks for the fixes! I tested the second beta, and f(x, y)=sqrt(x^2+y^2) is creating a minor issue. The final graph is fine, but the loading screen seems to show an error.

Thanks for that bug report! I had some overzealous debug printing remaining that I have now removed.

Inputting too big of a string causes the calculator to crash

Thanks for this bug report! I may have reproduced it by typing a very long equation and then hitting [GRAPH], which causes the calculator to freeze computing the graph, but didn't explicitly crash. Does this match your behavior?

I've also discovered a handful of minor bugs in my own testing:

• Rapidly pressing keys sometimes causes a ghost key to register
  
• Quitting via [2nd][MODE] still shows the OS MODE menu.
  
• The values of floating-point input fields after editing do not match the entered value exactly

Edit: I was able to reproduce typing even when the text entry buffer is full, although I didn't get it to crash. Thanks again for this report.



Edit #2: That buffer overrun issue has now been repaired, and the issue with showing the OS MODE menu after quitting via MODE has now been resolved as well (by the slightly blunt instrument of adding while(kb_AnyKey()) {} at exit).

Edit #3: Resolved issue "The values of floating-point input fields after editing do not match the entered value exactly" by modifying gCAS2's _g_atof() function. Instead of composing the fractional part by repeated addition of increasingly tiny fractional values, it composes the fractional part as an integer, then divides it a single time.