Well, a year ago since the last news, perhaps that is a sign.
I see the bug list has been growing since I left, thanks Trevor!
The main reason I am back working on what is supposed to be a done project :-) is that I still want P5 to self compile and get off needing a host compiler.To do that, it needs tickets #43 (stack direction) and #49 (constants move to top of code), which will feed into ticket #55 (unrelated).
Why do these items need to be done to build a code generating backend? Read ticket #43, but basically trying to match a backend to modern stack down machines when the front end generates stack up was just too much work. Thus P5 needed some modernization here.
Anyways, the plan is to run through these three tickets, then try to go over the bug list, retest and generate a new version.
Please note: it is helpful to me (and moves things along quickly) if bug tickets are generated against the current code version, which is 1.2 as of this writing.
Thanks!
Scott Franco
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Follow up: #43, stack direction change, is working thoughout the acceptance test suite. I have seen failures in the rejection suite, and these will be followed up.
I plan to follow up with a bug scrub. This is pretty much a requirement, since many of these fixes are already in the code and thus I need a merge.
There are a series of "modernization" fixes:
43 Feature addition: Change stack direction to down, heap to up
49 Feature addition: Move constants to top of code
48 Feature addition: Move to pulldown marks
The importance of these cannot be overstated. Basically this feature set aligns P5 with the way modern CPUs work, and thus move P5 closer to being a real tool, not just a history project. The good news is the hardest one, #43, is done, and the others will follow quickly.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Well, a year ago since the last news, perhaps that is a sign.
I see the bug list has been growing since I left, thanks Trevor!
The main reason I am back working on what is supposed to be a done project :-) is that I still want P5 to self compile and get off needing a host compiler.To do that, it needs tickets #43 (stack direction) and #49 (constants move to top of code), which will feed into ticket #55 (unrelated).
Why do these items need to be done to build a code generating backend? Read ticket #43, but basically trying to match a backend to modern stack down machines when the front end generates stack up was just too much work. Thus P5 needed some modernization here.
Anyways, the plan is to run through these three tickets, then try to go over the bug list, retest and generate a new version.
Please note: it is helpful to me (and moves things along quickly) if bug tickets are generated against the current code version, which is 1.2 as of this writing.
Thanks!
Scott Franco
Follow up: #43, stack direction change, is working thoughout the acceptance test suite. I have seen failures in the rejection suite, and these will be followed up.
I plan to follow up with a bug scrub. This is pretty much a requirement, since many of these fixes are already in the code and thus I need a merge.
There are a series of "modernization" fixes:
43 Feature addition: Change stack direction to down, heap to up
49 Feature addition: Move constants to top of code
48 Feature addition: Move to pulldown marks
The importance of these cannot be overstated. Basically this feature set aligns P5 with the way modern CPUs work, and thus move P5 closer to being a real tool, not just a history project. The good news is the hardest one, #43, is done, and the others will follow quickly.