234 entries in 0.593s
spyked:
http://btcbase.org/log/2018-03-28#1790159 <-- I checked on my system using ave1's instructions and it indeed uses
adacore's gcc. however: gcc uses its collect2 utility for linking (for some link-time optimization stuff -- check using gcc -v source.c), which in turn uses the *system* ld, i.e. /usr/bin/ld. that's why there's a difference in the link-time behaviour
☝︎ a111: Logged on 2018-03-28 20:41 phf: hmm, i wonder if
adacore's gprbuild uses its own gcc or global one
phf: hmm, i wonder if
adacore's gprbuild uses its own gcc or global one
☟︎☟︎ mod6: for what its worth, at the bottom of my paste, i've denoted that im using gcc 4.9.4 and
adacore 16
phf: the issue was already reported by someone else, but at the time the suggested fix was to put a bunch of C level annotations (some combination of static inlines), which i didn't think was an adequate solution, given that i don't understand why it does or doesn't work. but ascii's explanation makes sense, though i can't reproduce the issue on any of the machines i have with
adacore's gnat (freebsd, osx, debian)
diana_coman: I'm using gprbuild 2016 with gcc 4.9.4, both from
adacore spyked:
http://btcbase.org/log/2018-03-01#1786626 <-- I ran it on Debian/
Adacore 2016 -- system ld is version 2.29.1, and I expect
Adacore uses that instead of its own, otherwise I can't explain the behaviour. I'll also give it a try on a Gentoo with same
adacore. static inline should be safe from the pov of linkage, but yeah, there's a more general problem there that might benefit from a config.h/system.h or something similar. that or a
☝︎ diana_coman: but again,
adacore gprbuild, I can't vouch for any other version atm
mircea_popescu: diana_coman so basically a rut is shaping up here, "use gpr from
adacore -- gnat / others dun actually work" ?
diana_coman: as all my non-
adacore environments proved a mess
diana_coman: gprbuild from
adacore works however fine to build aggregate libs as far as I can tell
mod6: anyway, on the mule, i never did get
adacore running properly. this lappy is the only box i got
adacore '16 running on thus far.
a111: Logged on 2018-01-22 14:23 caaddr: GNATMAKE 4.9.2 is the answer to the now redundant question. I'll use
adacore instead. I had avoided this because it contains precompiled binaries, with no independent reproducible build certification
a111: Logged on 2018-01-22 14:22 asciilifeform: this would seem to create the unpleasant situation of having just 1 adatron. but it is not clear to me that there ever were 2. there was only
adacore and broken-
adacore (aka gcc-gnat)
caaddr: I was able to migrate to
Adacore 2016 okay, for what it's worth
a111: Logged on 2018-01-22 14:19 asciilifeform: and also i think at this point i will declare gpl-gnat to be a work of wreckers. it has zero upsides over
adacore's, and a million breakages , large and small.
ave1: Yes on restrict.adc plus it's the inline pragma that did not work for earlier FFA (
adacore 2014).
a111: Logged on 2018-01-22 16:04 ave1: re:
http://btcbase.org/log/2018-01-22#1773970, it seems that gcc simply lags behind and a releases version will never be updated / fixed to newer
adacore releases.
AdaCore 2014 also cannot build FFA from the box
ave1: nobody outside of
adacore seems to work on gnat GCC, so yes single version...
a111: Logged on 2018-01-22 14:19 asciilifeform: and also i think at this point i will declare gpl-gnat to be a work of wreckers. it has zero upsides over
adacore's, and a million breakages , large and small.
caaddr: success building Ch1 with
Adacore 2016
caaddr: thanks mod6. I'm installing
Adacore now
caaddr: GNATMAKE 4.9.2 is the answer to the now redundant question. I'll use
adacore instead. I had avoided this because it contains precompiled binaries, with no independent reproducible build certification
☟︎ diana_coman: from my own experience with
adacore and with gnat, I tend to agree with asciilifeform's evaluation there ^^
diana_coman: aha; tbh you are best off using
adacore's tools
ave1: the
adacore diff also includes previous 4.9.4 patches for musl
ave1: Note that this builds
adacore 2016 release (also build gprbuild)
a111: Logged on 2017-11-29 01:41 asciilifeform: funnily enough
adacore itself publishes a great many cut-down runtimes for various embedded boxes, e.g.
https://bitbucket.org/tkoskine/embedded-arm-gnat-rts/src . BUT they are not usable: 1) there is -- quite deliberately -- not one targeting conventional userland linux 2) none of them support exception handling, which wouldn't be a problem except that ALL BOUNDS CHECKS ARE EXCEPTIONTRONIC
a111: Logged on 2018-01-07 08:39 diana_coman:
http://btcbase.org/log/2018-01-07#1766177 <- I switched to
adacore at some point, when I had enough of all sorts of weird trouble essentially; never had that specific type of problem though
mod6: i use that
adacore version fwiw
mod6: ya. it's in the logs. i can dig it up if needed. but i believe my problem to be environmental and somehow to related to
adacore.
a111: Logged on 2017-12-11 21:58 mod6: this was supposidly because im running 2015
adacore. which i can buy. but have had even less success with
Adacore 2016 and 2017. not to say that I won't get the incantaiton correct at somepoint, but for now, I can't build it successfully.
mod6: this was supposidly because im running 2015
adacore. which i can buy. but have had even less success with
Adacore 2016 and 2017. not to say that I won't get the incantaiton correct at somepoint, but for now, I can't build it successfully.
☟︎ a111: Logged on 2017-11-18 18:10 ben_vulpes: but the very very funny event was the script from the
adacore binstaller simply fails, and instructed me to check an empty log file
a111: Logged on 2017-11-27 20:42 asciilifeform: i'd naively expect that
adacore's item would be the satanic one, and gcc's fork -- the functional. but so far in no case was this the finding.
mod6: and
adacore's is supposed to be canonical im guessing?
a111: Logged on 2017-11-27 20:40 asciilifeform: currently asciilifeform is barfing over the fact that
adacore's gnat is happy to build static 'independent library' .a while gpl gcc's gnat barfs
a111: Logged on 2017-06-06 20:38 asciilifeform: diana_coman: the example was written on a box with
adacore's gnat; the stock gnu one is stricter, doesn't permit Foo'Image -- instead you gotta FooTypeName'Image(Foo)
ben_vulpes: but the very very funny event was the script from the
adacore binstaller simply fails, and instructed me to check an empty log file
☟︎ ben_vulpes: tried both binturd from libre.
adacore and a portage ebuild
ave1: asciilifeform: for
http://btcbase.org/log/2017-07-21#1688362, it looks like you are building with
adacore 2016 edition, could you try with the 2015 edition (no need to do a real install, just unpack it and set path to the bin dir will be enough)
☝︎ ave1: it will build
adacore's gpl 2015 by default, cd into dir, run ./ada-build.sh /absolute/build/dir
a111: Logged on 2017-06-06 20:56 diana_coman: neah, it says it can't even see it then; will try another time with the
adacore 4.9
a111: Logged on 2017-07-19 12:54 ave1: (but the 4.9 gcc from
adacore does not match any 4.9.* release from FSF)
ave1: p.s I already have a musl build for x86_64 (based on musl_cross) (works on
adacore stuff and gcc-4.9.4), but is needs some more work
shinohai: ave1: Can you link me to the
adacore sources that worked for you?
ave1: and gcc 4.9.4 does so, but the
adacore code has problems
ave1: (but the 4.9 gcc from
adacore does not match any 4.9.* release from FSF)
☟︎ ave1: I can build a gnat using the sources from
adacore and all is fine
ave1: Turns out the version of gnat in gcc 4.9.* is based on GNAT GPL 2014 (from
adacore)
mod6: ok, so i dropped on a binary install of
adacore onto my gentoo environment here just to test...
mod6: im more referring to the massive tarball that you get from
adacore mod6: <+asciilifeform> mod6: the official
adacore gnat installs in 3minutes, but with obvious caveat (you live with binarolade) << yeah, i see they do have bins, but was going to custom handroll my own, to not only get one sanely built for my own environment, but will also allow me to set whatever ./configure flags I might need for optimizations.
mod6: My thinking is that, if i need to be able to use all the switches ad denoted in your makefile (since my mine doesn't seem to allow for that), I may need the
adacore configuration.
mod6: <+asciilifeform> mod6: not the problem. my thing builds on both
adacore's and trad gnu gnat. << yeah not sure what the problem is. it's a binary installation of gnat via apt-get, so that, i'm sure is part of the issue. however, it works better than the first ada configuration that I had, which, didn't work at all.