| 14:50 |
asciilifeform |
trinque: re http://logs.nosuchlabs.com/log/ossasepia/2020-04-05#1023752 -- worked in my test in '18 |
| 14:50 |
snsabot |
(ossasepia) 2020-04-05 trinque: is over here feeding ave1's gcc to itself. appears this hadn't been done yet, as naive attempt breaks. |
| 14:52 |
trinque |
asciilifeform: ah cool. figured that was table-stakes for compiler build script. |
| 14:53 |
trinque |
will post the breakage in a bit if I can't push through; seems to relate to linker |
| |
~ 3 hours 17 minutes ~ |
| 18:11 |
trinque |
asciilifeform: http://paste.deedbot.org/?id=aiMy << here's what I'm seeing when I use the gcc built by ave1's script as the host gcc on another run. |
| 18:11 |
trinque |
that's the behavior I expect out of said ld, that it's refusing to dynamically link. |
| 18:12 |
trinque |
what I don't know is why it wants to use liblto_plugin.so in the first place. |
| 18:12 |
trinque |
given that I added flags -fno-lto -fno-use-linker-plugin |
| 18:15 |
trinque |
http://ave1.org/2020/living-in-the-free-world/ << sounds like ave1 is preoccupied with an encounter with the real atm. |
| 18:15 |
trinque |
which, hell, I get it. |
| 18:15 |
asciilifeform |
hm 1st time i see this particular barf |
| 18:15 |
asciilifeform |
trinque: what type of gcc were you building with (and how obtained it?) |
| 18:15 |
asciilifeform |
i'ma guess it was gnu's gnat ? |
| 18:16 |
trinque |
originally used adacore gnat-gpl 2016 |
| 18:16 |
asciilifeform |
hmm then same thing i built with, odd |
| 18:16 |
trinque |
produced a gcc at /opt/gnat2016-musl |
| 18:16 |
trinque |
am now rebuilding with /opt/gnat2016-musl as host gcc |
| 18:16 |
trinque |
I suspect that ave1's never bothered to produce a static *gcc* |
| 18:17 |
trinque |
only tested that gcc produced static items other than itself. |
| 18:17 |
trinque |
would be glad to find I'm wrong |
| 18:17 |
asciilifeform |
is this happening on a musltronic linux ? |
| 18:17 |
asciilifeform |
( i have not attempted there ) |
| 18:17 |
trinque |
it's sitting on a "devuan" |
| 18:17 |
trinque |
systemd-less debian |
| 18:18 |
asciilifeform |
ought ping BingoBoingo , iirc he has been using devuan for some time |
| 18:18 |
asciilifeform |
( no idea whether tried gnat there tho ) |
| 18:20 |
trinque |
hey BingoBoingo ftr you oughta be able to self-voice. |
| 18:21 |
trinque |
lmk if not |
| 18:22 |
asciilifeform |
btw are kernel-headers installed ? iirc gnat depends on'em for somereason |
| 18:23 |
trinque |
yep, and got a fully functional gnat out the end of the first run. |
| 18:23 |
asciilifeform |
interestingly, ac_nonexistent.h dun appear to exist on my boxen. anywhere. |
| 18:23 |
trinque |
http://paste.deedbot.org/?id=gUX7 |
| 18:23 |
trinque |
yeah that's a sanity check |
| 18:23 |
trinque |
it wants the header to fail. |
| 18:24 |
asciilifeform |
nuttery |
| 18:25 |
trinque |
yup, pile of fractal contempt |
| 18:27 |
asciilifeform |
1 thing i had to do when building, was to remove all paths to native gcc from shell envir |
| 18:27 |
asciilifeform |
(leaving only adacore's gnat) |
| 18:27 |
trinque |
yeah, distro's gcc is literally uninstalled |
| 18:27 |
asciilifeform |
otherwise got header barf |
| 18:27 |
asciilifeform |
ah |
| 18:28 |
asciilifeform |
what was it you built adacore's 2016 with originally ? |
| 18:28 |
trinque |
anyhow, soon as I've got a static gcc, I'm hauling it into my busybox env, bringing in just the deps necessary to do a rebuild, and I'm not doing this bootstrapping exercise again |
| 18:28 |
trinque |
a devuan gcc-4.9.4 |
| 18:28 |
asciilifeform |
hm then, oughta work, grr |
| 18:29 |
asciilifeform |
trinque: on my ffa box, i in fact went straight to using ave1's binary package, after which built ave1's src with it, have been using ever since |
| 18:29 |
* |
trinque suspects the script is not insulated from host in some undiscovered way |
| 18:29 |
asciilifeform |
^ |
| 18:29 |
trinque |
interesting, I'll pull down his bins and see what they do |
| 18:30 |
asciilifeform |
his had a script that built it w/out problem (on a box w/ adacore's 2016) here |
| 18:31 |
trinque |
ah but that's exactly what I used lol |
| 18:31 |
trinque |
no big deal; I'll get through it |
| 18:31 |
trinque |
this is way smaller than all the fruitless gentoo hairshirting I did |
| 18:31 |
trinque |
but seriously ave1, lick your damned wounds and show up. |
| 18:31 |
asciilifeform |
aite. if you remain stuck, trinque , upload the various tarballs (and put link to the devuan used) and i'ma try to replicate |
| 18:31 |
trinque |
word, thanks |
| 18:32 |
asciilifeform |
i have a machine avail. on which can do this |
| 18:32 |
* |
asciilifeform bbl, meat maneuvers |
| 18:32 |
* |
trinque is 33 today, likely to disappear, get hammered, and chase naked woman around the house |
| |
~ 1 hours 2 minutes ~ |
| 19:34 |
BingoBoingo |
trinque: Ah, self voicing works, ty and happy birthday |
| 19:36 |
trinque |
thanks sir |
| 19:43 |
trinque |
upstack, the script seems out of order. binutils builds before musl, but binutils wants to link with a libc, so what gives? |
| 19:44 |
trinque |
http://paste.deedbot.org/?id=JEbL << current barf, dl_iterate_phdr is defined in libc |
| 19:44 |
trinque |
(trying to build gas) |
| 19:46 |
asciilifeform |
happy bd trinque ! |
| 19:46 |
trinque |
ty asciilifeform |
| 19:54 |
* |
BingoBoingo indeed fell into the GGC4 release of devuan on most machines during the last round of housekeeping. |
| 20:01 |
lobbes |
hbd trinque. enjoy the hammer time |