You are not logged in.
Pages: 1
Hi, I'm trying to build and install idris with stack. Working through the dependencies was just fine but when I reach the point of building idris, the last thing I see is this:
idris-1.0: build
Package idris uses a custom Cabal build, but does not use a custom-setup stanza
Not using the explicit setup deps approach based on configuration
Strongly recommend fixing the package's cabal fileThe process takes for ever, the last time I killed it after ~9 hours, but funnily, after killing ghc, a whole lot of new output appeared in the pane, it begins like this:
make: Leaving directory '/tmp/stack11833/idris-1.0/rts'
Generating .stack-work/dist/x86_64-linux-ncurses6/Cabal-1.24.2.0/build/autogen/Version_idris.hs for prerelease PRE
Building idris-1.0...
Preprocessing library idris-1.0...
[ 1 of 108] Compiling Version_idris ( src/Version_idris.hs, .stack-work/dist/x86_64-linux-ncurses6/Cabal-1.24.2.0/build/Version_idris.o )
[ 2 of 108] Compiling Util.System ( src/Util/System.hs, .stack-work/dist/x86_64-linux-ncurses6/Cabal-1.24.2.0/build/Util/System.o )I assume it to stem from one of the few child processes of ghc, and I wondered, is it possible, provided I found out from which one exactly, to get to those informations without killing anything? As far as I understand, they somehow got shaded by ghc, if that's possible.
[EDIT] I found out about the
tail -f /proc/<pid>/fd/1command, it works well on ghc (apparently some other process was shading it and ghc actually produced the 'good' output).
Last edited by haysutonirst (2017-04-11 11:56:22)
Offline
Hi, I'm trying to build and install idris with stack.
Can you provide information/links on these? Specifically, idris is already available in [community], why do you need to build it yourself? I'm not familiar with idris or stack: idris is readily googleable, but "stack" not so much. If you can provide some background you'll greatly expand the pool of people who might be able to help you.
Working through the dependencies was just fine
Why not start with the PKGBUILD from the ABS?
I wondered, is it possible ... to get to those informations without killing anything? As far as I understand, they somehow got shaded by ghc, if that's possible.
I doubt they were shaded/hidden. I suspect that part of the build was just never reached until the previous stages exited. When you killed a process, the next in the series started.
It is possible that background processes would not send their output to the terminal, but killing the parent would not likely change where the backgrounded processes output was sent.
However, if I am wrong about all of the above, you can use ps/pgrep to get the pid of the process in question, then find it's standard streams under /proc/ to investigate further. But again, I highly doubt this output is being hidden, it's just not being generated when the earlier stage of the build process is stalled.
"UNIX is simple and coherent" - Dennis Ritchie; "GNU's Not Unix" - Richard Stallman
Offline
Pages: 1