« first day (571 days earlier)      last day (878 days later) » 

01:37
I wonder if we could expand the ROM maker to also include the RAM initialized with the data section
We probably could, actually
Is there any performance advantage to having the rodata actually be in the ROM
I guess if you need to jump back there?
No it would be more like referencing it.
That could actually be the unused addressing mode: read from position in the ROM
That would be signifcant hardware changes I suspect
Reading from ROM kinda doesn't work with the current hardware
01:43
@quartata And software changes too (specifically, in the QFTASM interpreter).
It would only be worth it if ROM access is faster
Other we shant even bother with an rodata section
actually wait I'm dumb there is no unused space since there's also just immediate in addition to ABC
so it would mean a new instruction or an extra bit which is definitely not worth it
Nevermind
At any rate the potential for preinitialized RAM is reason enough for me not to make this backend emit moves for every word in the data section....
So if we make a new assembler what should the data section text format look like is the question
I suspect GCC does AT&T syntax for its default
 
13 hours later…
Anonymous
15:21
@quartata Do you have a repo anywhere with your work on the backend?
I mean locally yes but not on GH yet. I wanted to get that up today or tomorrow
part of the problem is that I'm not sure I want to upload all of GCC, it's huge
if I just upload my part I'd need to include a build script or something that retrieves the rest
I could use a submodule I guess
Anonymous
A submodule seems like the best way
15:37
I propose that the third "program" we do (after Pong) is a simulation of Game of Life.
Anonymous
@El'endiaStarman Gotta loop it back around :P
2meta
 
2 hours later…
Anonymous
17:14
just_enough_meta.jpg

« first day (571 days earlier)      last day (878 days later) »