technomancy / grenchman Goto Github PK

Download and run of OSX binary on 10.8.5 produces dynamic link error:

grench
dyld: Library not loaded: /usr/local/opt/libffi/lib/libffi.6.dylib
Referenced from: /usr/local/bin/grench
Reason: image not found
Trace/BPT trap

Note there is no /usr/local/opt directory on my system (default installation)

uname -a
Darwin twelf 12.5.0 Darwin Kernel Version 12.5.0: Mon Jul 29 16:33:49 PDT 2013; root:xnu-2050.48.11~1/RELEASE_X86_64 x86_64

otool -L grench
grench:
/usr/local/opt/libffi/lib/libffi.6.dylib (compatibility version 7.0.0, current version 7.1.0)
/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 159.1.0)
/usr/lib/libedit.3.dylib (compatibility version 2.0.0, current version 3.0.0)

mdfind -name libffi
/System/Library/Tcl/8.5/Ffidl0.6.1/libFfidl0.6.1.dylib
/System/Library/Tcl/8.4/Ffidl0.6.1/libFfidl0.6.1.dylib
/usr/lib/libffi.dylib # appears to be version 1.0.0

If a form is incomplete, we need to read more input.

We might be able to just send it to the server and rely on its parser to avoid reinventing the reader. Just read another line if it returns an "EOF while reading".

Trickier because defmulti dispatch can't be changed dynamically.

Outputs get stuck behind inputs when we send stdin or interrupt messages.

The upshot of which is that the Wheezy binary now links to a nonexistent libffi, but the Ubuntu binary runs just fine.
the version in Jessie/Sid is 3.1-2:
https://packages.debian.org/jessie/libffi6
And the version the ubuntu package was last tested against apparently was 3.0.13
http://packages.ubuntu.com/raring/libffi6
which has been bumped up to 3.1-ish? which is close enough I'd think to say it works there too:
http://packages.ubuntu.com/trusty/libffi6

Loading up libreadline during the build process instead of at runtime with Dl.dlopen is supposedly better and allows for Mac compatibility since it doesn't hard-code "libreadline.so" into the source, but it breaks compatibility with Ubuntu.

Waiting for clarification as to what's going happening on this thread on the ctypes mailing list: http://lists.ocaml.org/pipermail/ctypes/2013-October/000013.html

http://p.hagelb.org/mystery.gif

I'm trying to install grenchman on OSX ML:

$ wget https://grenchman.s3.amazonaws.com/downloads/grench-0.2.0-mac
$ chmod +x grench-0.2.0-mac
$ mv grench-0.2.0-mac /usr/local/bin/grench
$ grench
dyld: Library not loaded: /usr/local/opt/libffi/lib/libffi.6.dylib
  Referenced from: /usr/local/bin/grench
  Reason: image not found
[1]    20224 trace trap  grench

Am I missing something?

Thanks :)

https://github.com/diml/lambda-term

Pros:

• More advanced repl features
• The other repl enhancement requests could either be done more easily with existing lambda-term features, or contributed to lambda-term for the good of all of us

Cons:

• Adds a bunch of dependencies (though it removes the dependency on libreadline)

At the time of this writing all of the download links on http://leiningen.org/grench.html are unavailable.

Fetching https://grenchman.s3.amazonaws.com/downloads/grench-0.2.0-debian-7:

<Error>
  <Code>AllAccessDisabled</Code>
  <Message>All access to this object has been disabled</Message>
  <RequestId>66A583042E4636FA</RequestId>
  <HostId>lztFuukq6sLsdTuGGIXW7MmCxfjAoC27qbaw1n+yDy/6z9LdUAlWIBIfdnHlFtqkCuqAPpj9qG0=</HostId>
</Error>

This should help with shebangs.

~/.inputrc setting 'set editing-mode vi' ignored, even if it seems to be working for other programs, i.e. bash.

Mac OS X: 10.11.6 (El Capitan)
grenchman 0.2.0
Leiningen 2.6.1 on Java 1.8.0_05

I suppose the project is no longer actively developed but I'd like to submit the idea, just in case.

For scripting (e.g. on CI server) it would be great to be able to get exit error codes on evaluation exceptions. At the moment it looks like grench always returns 0:

ᐅ grench eval '(throw (ex-info "oops" {}))'
clojure.lang.ExceptionInfo: oops
 at clojure.core$ex_info.invokeStatic (core.clj:4725)
 ...

ᐅ echo $?                                                              
0

This is what I tried:

$ lein new app hello
$ cd hello
$ lein run
Hello, World!
$ lein repl :headless
$ LEIN_REPL_PORT=`cat .nrepl-port` grench lein run # in another terminal session
  Unknown status: 19:namespace-not-found
  Unknown status: 5:error

It seems like setting the ns to leiningen.core.main is giving a namespace-not-found error, but I have no idea why.

When starting a process like grench lein cljsbuild auto foo it doesn't seem possible to kill this with Control-C.

$ lein trampoline repl :headless 
nREPL server started on port 38457 on host 127.0.0.1
$ lein version
Leiningen 2.3.0 on Java 1.7.0_25 OpenJDK 64-Bit Server VM

so far so good

$ ~/bin/grench-0.2.0-debian-7 eval '(println "hi")' 
Couldn't read port from .nrepl-port or $GRENCH_PORT.
$ GRENCH_PORT=38457 ~/bin/grench-0.2.0-debian-7 eval '(println "hi")'
hi

ok

$ ~/bin/grench-0.2.0-debian-7 lein version 
Couldn't read port from ~/.lein/repl-port or $LEIN_REPL_PORT.

If Leiningen is not running, launch `lein repl :headless' from outside a
project directory and try again.
$ LEIN_REPL_PORT=38457 ~/bin/grench-0.2.0-debian-7 lein version
  Unknown status: 19:namespace-not-found
  Unknown status: 5:error

o_O

So, to sum it up

• Why wouldn't ~/.lein/repl-port be created automatically? Can I just do that? Why no hint which format?
• Why is .nrepl-port not there? And why is that in the current dir? Same questions as above as well
• If I set LEIN_REPL_PORT, why does that not work?

Upgrading to leiningen 2.3.3 did not help, btw.

If we are to re-use a JVM for Leiningen itself, we'll get unpredictable results unless we isolate each project's set of plugins in their own classloader. This will need to be implemented inside Leiningen.

using the build instructions on ubuntu 17.10 64bit

ocamlbuild -use-ocamlfind -lflags -cclib,-lreadline grench.native
+ ocamlfind ocamlc -c -g -annot -thread -package core -package async -package ctypes.foreign -o nrepl.cmo nrepl.ml
File "nrepl.ml", line 1, characters 5-14:
Warning 3: deprecated: module Async.Std
[since 2017-02] use Async. The Std sub-module is no longer needed
File "nrepl.ml", line 2, characters 5-13:
Warning 3: deprecated: module Core.Std
[since 2017-02] Use Core. The Std sub-module is no longer needed
File "nrepl.ml", line 11, characters 21-61:
Warning 3: deprecated: module Core.Std
[since 2017-02] Use Core. The Std sub-module is no longer needed
File "nrepl.ml", line 76, characters 6-13:
Error: This pattern matches values of type 'a option
       but a pattern was expected which matches values of type
         equal:(string -> string -> bool) -> string option
Command exited with code 2.
Compilation unsuccessful after building 8 targets (6 cached) in 00:00:00

It's pretty easy to confuse it by running grench help copying, etc.

According to the Cake guys, this is really hard to get right. Maybe better to let users handle this for now.

According to the readline docs, we need to set the rl_completion_entry_function variable to a function pointer, but the ctypes docs don't really cover how that works.

Or is this a non-starter because nrepl doesn't have authentication? Or is it there and I'm just really bad at reading OCaml?

When I execute grench-0.2.0-mac (placed on the $PATH as executable), I get the following error:

(user@air) ~/bin $ ./grench-0.2.0-mac
dyld: Library not loaded: /usr/local/opt/libffi/lib/libffi.6.dylib
Referenced from: /Users/user/bin/./grench-0.2.0-mac
Reason: image not found
Trace/BPT trap: 5

I get an error when I launch, for instance grench lein help. The conf below, this is the error I get:

Exception in thread "nREPL-worker-0" java.lang.Error: java.net.SocketException: Socket closed
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1148)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
    at java.lang.Thread.run(Thread.java:745)
Caused by: java.net.SocketException: Socket closed
    at java.net.SocketOutputStream.socketWrite(SocketOutputStream.java:116)
    at java.net.SocketOutputStream.write(SocketOutputStream.java:153)
    at java.io.BufferedOutputStream.flushBuffer(BufferedOutputStream.java:82)
    at java.io.BufferedOutputStream.flush(BufferedOutputStream.java:140)
    at clojure.tools.nrepl.transport$bencode$fn__243.invoke(transport.clj:103)
    at clojure.tools.nrepl.transport.FnTransport.send(transport.clj:28)
    at clojure.tools.nrepl.middleware.render_values$wrap_renderer$reify__1138.send(render_values.clj:30)
    at cider.nrepl.middleware.track_state$make_transport$reify__15743.send(track_state.clj:186)
    at clojure.tools.nrepl.middleware.interruptible_eval$interruptible_eval$fn__679$fn__682.invoke(interruptible_eval.clj:223)
    at clojure.tools.nrepl.middleware.interruptible_eval$run_next$fn__674.invoke(interruptible_eval.clj:190)
    at clojure.lang.AFn.run(AFn.java:22)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
    ... 2 more

It looks like a nrepl middleware issue. If I remove my profile.clj the error changes in:

Unknown status: 19:namespace-not-found

Can I do something to debug?

My conf:

Leiningen 2.5.3 on Java 1.8.0_66 Java HotSpot(TM) 64-Bit Server VM
nREPL server started on port 34379 on host 127.0.0.1 - nrepl://127.0.0.1:34379
REPL-y 0.3.7, nREPL 0.2.11
Clojure 1.7.0

Basic implementation exists in the Repl module. Needs:

• proper prompt loop displaying namespace
• readline-ish support
• interrupt

Completion can be done later.

This should give you a rudimentary repl, but it hangs when run inside a project:

$ grench run -m clojure.main/main -r

Turns out the need-input status is never sent over nrepl.