Bradley Harden

I just took a deeper look at `queue_pkg` and I think there is room for improvement in the implementation. Right now, the string buffer is linear. When trying to push...

I had another realization about the current implementation of `queue_t`. When a queue is initialized, the `string_ptr` length is 0. Here is the line where the queue is resized. ```vhdl...

Sorry for the spam, but I want to jot these ideas down while I have them. We can extend the reformulation of `queue_t` further. Define `element_t` as a `string_ptr`, then...

@LarsAsplund, I now have a working implementation of `queue_t` that uses an array of `string_ptr_t` rather than a single `string`. It also uses a circular buffer, rather than a linear...

No, I did not consider it. It wouldn't be hard to merge them though. All I did is add an integer_vector to use as a stack for old pointer reference...

@LarsAsplund and @kraigher, I have some preliminary performance tests of my new implementation of `queue_t`. My version of `queue_t` uses an `integer_vector_ptr` for the data, where each integer is a...

Let me know what you decide is best. I don't really have an opinion. I can return it to the old approach. It just means more pin aliases. Speaking of,...

One other thing to note. The `v1` API had the pins in `FloatingInput` mode at reset. But that's actually wrong. They are in a floating disabled mode. I don't think...

@jacobrosenthal, wasn't the undocumented IOSET for the SD card, not the display? I'm not sure what's wrong with the display. Do you have any other symptoms? Maybe you could double...

@jacobrosenthal, I returned the `Sets` to their original architecture. Could you take another look at the display problem? I have no way to test it.