Whole-known-network
Yeah, there are forks of LLVM though, right? Maybe that's the way to go? Maybe not super diverse, but better than nothing?<br><br>I'm hardly an expert so much as an onlooker in such realms, but:<br><br>There's llvm-mos (which has the 6502 and even the friggin HuC6280 [wow!] as targets), so I can't really imagine how deprecating MC68k support would be sane in contrast; albeit maybe it should live on in a fork like llvm-mos?<br><br>I'm just spitballin here, but from my vantage, the MC6800 and MOS6502 are already so interrelated that llvm-mos might be the <i>best</i> place for MC680x0 targets to thrive relative to like, anything that LLVM mainline should consider in active development?<br><br>LLVM didn't get started until 2000 and I am hard pressed to think of anyone who was using MC68k as a daily driver at that point (even weirdos such as I, who still has my Amiga 1200 amidst my prized material possessions). Sure, OpenBSD and some other outliers (Hi NetBSD!) were still targeting Amigas and even some MC880000 systems circa 2000, but IIRC they were doing so with GCC at the time, LLVM was way too <i>new</i> in 2000 (heck remember when TenDRA seemed as if it might be a viable alternative to GCC? It wasn't much later than that, but here I am getting all nostalgic.)<br><br>Meanwhile, I'm still trying to wrap my head around how some have been supposedly been using Zig with llvm-mos (you do you I guess, whoever you are), but then the other day someone else had a YouTube video saying how their preferred Zig program couldn't be built on a dual G5 they bought for $20? (like, wtfh!? IBM still makes newer Power9/10/11 systems I think; though I guess G5s are very long in the tooth, how are folks building software for older PPC systems? I wouldn't know without digging some PowerBooks out of storage and giving them a whirl I guess.).<br><br>That word "valuable" is so loaded; but I certainly understand wanting to move the needle forward even if some seem stuck (or more rarely, doing revitalizing retrodev) in the past.<br><br>But there's also the part of me which remembers: MC68k is <i>so nice</i> at the asm level, like 6502 before it.<br><br>C seems <i>overwrought</i> for those systems, to me at least. C felt really bloated to me on such things back then (I am still dumb founded by how <i>expensive</i> and <i>slow</i> Apple and NeXT MC68k systems seemed relative to alternatives [not Intel, more like CBM and SGI and Sun] in that era); and when I look back at how much simpler C was then relative to LLVM/Clang and such now I just wonder if all software is doomed to have the "Tetsuo turns into a giant baby scene from AKIRA" growth pattern? I really don't think it should be that way, yet here we are anyway. ;-/<br><br>CC: <span class="h-card"><a href="https://social.treehouse.systems/users/dysfun" class="u-url mention">@dysfun@treehouse.systems</a></span> <span class="h-card"><a href="https://mastodon.social/users/mcc" class="u-url mention">@mcc@mastodon.social</a></span><br>
<p><span class="h-card" translate="no"><a href="https://with.iridium.ink/@artemis" class="u-url mention">@<span>artemis</span></a></span> the modern ones are bad in different ways that in my view are wore >.></p>
<p><span class="h-card" translate="no"><a href="https://social.treehouse.systems/@dysfun" class="u-url mention">@<span>dysfun</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> as a more extreme example, take 8051. i heavily use 8051 to this day. do i think LLVM, and by extension all of its users, should pay for my desire to compile C code to 8051? absolutely not!</p><p>this is probably the main reason compiler diversity is important: it allows different groups to care about different things</p>
<p><span class="h-card" translate="no"><a href="https://social.treehouse.systems/@dysfun" class="u-url mention">@<span>dysfun</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> yeah this basically tracks</p><p>i'm not really convinced that supporting m68k is valuable, but it is a fact that LLVM as it is today isn't a good fit for it (and it's also a fact that the degree of churn makes supporting it out-of-tree not a viable long term solution)</p>
<p><span class="h-card" translate="no"><a href="https://gotosocial-dev.svc.0x0a.network/@littlefox" class="u-url mention">@<span>littlefox</span></a></span> <span class="h-card" translate="no"><a href="https://pony.social/@thephd" class="u-url mention">@<span>thephd</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> haha no ISE is the stuff of absolute nightmares</p>
<p><span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> <span class="h-card" translate="no"><a href="https://infosec.exchange/@malwareminigun" class="u-url mention">@<span>malwareminigun</span></a></span> <span class="h-card" translate="no"><a href="https://pony.social/@thephd" class="u-url mention">@<span>thephd</span></a></span> Tcl is like one of the least bad things about the tooling situation. Tcl is kind of cute if you squint. There is absolutely nothing cute about, for example, the "IP" situation</p>
<p><span class="h-card" translate="no"><a href="https://infosec.exchange/@malwareminigun" class="u-url mention">@<span>malwareminigun</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> <span class="h-card" translate="no"><a href="https://pony.social/@thephd" class="u-url mention">@<span>thephd</span></a></span> Let me put it this way: If the HDL tooling situation were merely "Oh, you just have to learn TCL and you're fine", this would be a godsend.</p><p>Rather what happens is you show up, the existing users explain "you just have to learn TCL and you're fine", so you learn TCL, *and it does not help you at all*, and nothing is fine</p>
<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> <span class="h-card" translate="no"><a href="https://pony.social/@thephd" class="u-url mention">@<span>thephd</span></a></span> <span class="h-card" translate="no"><a href="https://mastodon.social/@mcc" class="u-url mention">@<span>mcc</span></a></span> The crunching you just heard is the phrase 'autoconf [...] look[s] like a dream in comparison' and it caused me to clench teeth so hard they're all gone now</p>
<p><span class="h-card" translate="no"><a href="https://hachyderm.io/@natik" class="u-url mention">@<span>natik</span></a></span> yeah. I mean, I can handle disagreement, and I can work towards things I wouldn't personally choose, but if at some point I realize the choices being made are not going, in the long run, to benefit *anybody* involved, I don't have a lot of faith in the ability of leaders to... lead</p>