IRC channel logs
2020-05-01.log
back to list of logs
<OriansJ>mihi: no the PDF refers to the knight native one as well (it would be the bios x86 one as well if I ever got that done); hence why I am hoping bauen1 makes their kernel buildable by M2-Planet; then the steps to GCC could be done on it prior to the needing to build a Linux kernel or a coreboot/libreboot payload/firmware. <fossy>OriansJ: we can get to linux before gcc <fossy>Tinycc can compile linux 2.4 <OriansJ>certainly something to consider depending on if bauen1's kernel can run GCC <OriansJ>plasma41: in theory but again one would want to minimize the need for a posix to build a posix; as only M2-Planet can build C programs while running on bare metal <OriansJ>so either M2-Planet needs to grow enough to build xv6 or someone would need to strip xv6 down enough for M2-Planet to build it directly <OriansJ>I honestly am more than happy regardless what someone else does; if they save me the work of having to do it all myself <OriansJ>fossy: I could give you full access to the mescc-tools github repo, provided you agree to limit your commits to only changing the contents of the kaem directory. As this would address dddddd's issue. <OriansJ>and that most of your development is done off the main branch <OriansJ>would that be a compromise you would be willing to agree to fossy? <OriansJ>or we could just break kaem off of mescc-tools. <Hagfish>that seems to be a free(?) alternative to JIRA <Hagfish>it might be a good way for people here to collaborate and track progress, if you wanted to check it out <OriansJ>Hagfish: I can't even get people to respond to direct messages or questions. I am up to my neck dealing with a CMS audit, COVID-19 security change requests out my ass and my 2 Month old son. How the hell am I supposed to get people to actually do anything? <OriansJ>we only have janneke, deesix, fossy, markjenkins, gio and myself; actually writing bootstrapping code. All of which appears to be done in when time is available fashion, along the lines which look fun. <OriansJ>I am literally just trying to find reasonable compromises such that we actually work together and help each other reach the collective goal. <fossy>OriansJ: yes, either is acceptable <fossy>I would only change kaem if you went integrating or split off is fine too <OriansJ>thank you fossy. I'll be sewing together a mescc-tools and M2-Planet release tonight <Hagfish>hey OriansJ. sorry, i didn't mean to suggest you needed to "get people to actually do anything" <Hagfish>i just meant that there's a lot of great streams of work being done here, and it would be nice if people could see at a glance where each part is at <Hagfish>if people don't think that the overhead of keeping their tickets up to date is worth it, then yeah, there's no point trying to force them <Hagfish>my personal project right now is teaching a friend Python, but if i was working on an actual software project with him, i'd definitely want some sort of online todo list or something to help coordinate that <OriansJ>Hagfish: sorry for sounding a little short; I guess the sleep deprevation and having to be at work in 6 hours isn't doing my poor communication skills any help. <Hagfish>heh, i know that feeling, don't worry :) <Hagfish>that's very comprehensive and impressive <OriansJ>download and view in org-mode for best results <OriansJ>thus anyone would be able to read the repo and know the state of everything (approximately) <OriansJ>as well as how things evolved in the bootstrap over time as issues are discovered and successes mount <Hagfish>i might be "over-fitting" with my mental model, but i feel like the current pieces of work (and likely next pieces of work) could be represented as a dozen or so "cards"/tasks on a dashboard <Hagfish>there could be columns for "todo", "in progress", "done", and "released" <Hagfish>maybe a separate code review stage for people who want that before release <Hagfish>personally i like having that sort of "paper trail", and being able to offload the brain space of keeping track of the tasks into an external source like a dashboard <Hagfish>i know other people's minds work differently though :) <OriansJ>well if we are just talking finishing the GCC bootstrap in as few steps as possible, we only have one task remaining <OriansJ>mes-m2 is trivial to build with M2-Planet <Hagfish>aren't there shell and POSIX tasks that are still in progress, and maybe a placeholder for hardware stuff, and porting maybe <OriansJ>everything above MesCC is *done ish* <OriansJ>Hagfish: kaem-minimal.hex0 is 737bytes and has been done for a while <OriansJ>The POSIX only matters if we wish to solve the x86 bootstrap down to metal <OriansJ>and isn't technically required to say we bootstrapped GCC from hex0 <Hagfish>the bootstrap path involves tcc, right? what would the tcc run on? <OriansJ>but yes there is bauen1's POSIX work <OriansJ>everything above M2-Planet requires a POSIX <Hagfish>i see what you mean about "gcc from hex0", i guess you are using the usual definition of bootstrapping there <OriansJ>hence why I need bauen1 to agree to the goal of being buildable by M2-Planet, otherwise I will have to write a POSIX on my own. <OriansJ>again the steps outstanding depends entirely on your definitions <Hagfish>is bauen1 not targeting that? or is it too early to decide? <OriansJ>Hagfish: rigt now it appears bauen1 is more interested in targeting TCC <OriansJ>but with minor restrictions, it should be possible to convert from TCC to M2-Planet <Hagfish>i think a task like that could live on the same board for tracking "bootstrapping" progress <OriansJ>but again, that will be alot of work to be done; depending upon what assumptions are made. <Hagfish>i suppose the collaboration here is a bit too loose to have everyone agreeing on one definite path <OriansJ>Hagfish: well there are so very few of us here <OriansJ>Most just watch, even when documentation and other tasks could really help <Hagfish>i think the document you linked to above is probably a good way of onboarding people, even if it is a bit daunting at first <OriansJ>getting the pieces packaged for distros is another piece only janneke and vagrantc seem to be doing. <Hagfish>maybe people just feel that their skills don't overlap with any of the current areas of development <OriansJ>Hagfish: we are more than happy to teach anyone who wants to learn <Hagfish>hmm, ongoing work with moving targets (like packaging and documentation) are maybe hard to track as individual tasks <OriansJ>I'll teach assembly, C, FORTH and/or scheme to anyone willing to pick up an editor and program in the direction of completing this bootstrap <Hagfish>if someone knew those languages, most of the teaching would be explaining what the tasks actually require <Hagfish>or maybe it would be obvious to someone with those skills <OriansJ>Hagfish: good programming is simply doing reasonable steps in an organized fashion to produce a desired result. <OriansJ>Just understanding what you are doing is 90% of the work <OriansJ>The actual coding is only 10% of the effort <OriansJ>mescc-tools v1.0.0 has been released <OriansJ>M2-Planet v1.6.0 is now released (depending upon mescc-tools v1.0.0) <OriansJ>the issue with a seperate kaem repo has been resolved