<rekado>(still stuck with a segfault in the compiled hsc of ghc 4.08.2) <clemens3>oriansj: so which guile version is supported, I have 3.0.7. Is it too new or too old? ***codething is now known as tanvirroshid
<acrow>What guix package provides the 'jar' executable?, icedtea? <acrow>What provides the 'jar' executable: icedtea:jdk. <oriansj>clemens3: I am thinking too new as the manual still references Guile version 2.0.13 or later, including 2.2.x and 3.0.x <clemens3>oriansj: ok, i see what older versions i have around, thanks.. ***ChanServ sets mode: +o rekado_
<rekado_>another lead for Haskell bootstrapping: hbc. <rekado_>early versions of ghc should be buildable with hbc <rekado_>ghc-4’s configure script is so silly. Checks for a bunch of existing Haskell compilers in the environment and then doesn’t do anything with them. <rekado_>also annoying: the build never fails even when make prints errors. It just marches on and then fails because it tries to invoke “hsc”, which couldn’t be linked. <rekado_>part 2 is covered in a bunch of unordered and increasingly scattered notes; this is an attempt to extract *something* useful out of this whole ordeal. <rekado_>I’m using generated C files for GHC 4 to build it as intended <rekado_>but it’s juts 3MB, which is better than a massive GHC 7 blob. <rekado_>unfortunately, the actual compiler segfaults almost immediately <rekado_>and the debug info is terrible (building with debug symbols fails because of how hacky all that GHC code is) <rekado_>new gdb error: “Backtrace stopped: previous frame inner to this frame (corrupt stack?)” <civodul>or exciting! depending on the viewpoint <rekado_>with the rediscovery of the HBC sources I’ll have another path to explore if this GHC 4 stuff turns out to be beyond my ability to fix.