Trying to build from master dies with many errors #239

edubkendo opened this Issue Dec 21, 2013 · 12 comments


None yet
3 participants

Cannot find instance method read(java.lang.String) on org.mirah.jvm.mirrors.generics.TypeInvoker

rake aborted!

Command failed with status (1): [jruby -S mirahc -N --dest build/mirrors --...]

The output is too much to put here, and I'm not sure what's relevant so I'll link to a gist:

It looks like the first warnings begin here:

First error here:

I think this one is where the actual problem is:

It's not finding methods on some of the classes in Generics/mirrors I think.

For the record:

eric@edub:~/git/community/mirah (master) $ bundle install
Fetching gem metadata from
Resolving dependencies..........
Installing rake (10.1.1)
Using bitescript (0.1.4) from (at master)
Installing columnize (0.3.6)
Using mirah ( from source at /home/eric/git/community/mirah
Installing ruby-debug-base (0.10.4)
Installing ruby-debug (0.10.4)
Installing test-unit (2.5.5)
Using bundler (1.3.5)
Your bundle is complete!
Use bundle show [gemname] to see where a bundled gem is installed.


baroquebobcat commented Dec 21, 2013

Hmm. I haven't tried building w/ JRuby 1.7.9 yet, maybe that's the issue. I'll check.


baroquebobcat commented Dec 21, 2013

It doesn't look like that's the problem. It's weird, you should have had to compile a number of other things before compiling the mirrors, but it looks like it only tried to compile mirrors. Have you had this checkout of mirah for a while?

Can you try

rake clean clobber

rake by itself should compile everything and run the test suite. Let me know how that goes.

It was checked out last night, but I had tried a few times trying to figure out what was going on. I thought I'd deleted that dir and re-cloned it before trying again to post the issue.


So, weird thing, I made a fresh clone into a new location. Got the same errors, but then, after trying rake clean clobber, rake I looked back through the first stacktrace and saw it had originally built the gem, it just did some other stuff after that caused the error and caused the build to fail. So I ran rake gem again with the same results, but did get the gem installed. Appears to be working ok, although I haven't tried much yet. Is it missing something?

I gisted all the stacktraces and output here:

You can see where the gem builds here and here

So, dunno what's going on exactly but I do seem to have a working install now, unless, like I said, it needs something that second part of the build was supposed to provide.


baroquebobcat commented Dec 23, 2013

Hmm. It looks like rake gem by itself is no longer creating all the necessary jars.


baroquebobcat commented Dec 24, 2013

I found the issue. rake gem was depending on the wrong bootstrap task. I think I fixed it. Closing.

Thanks for reporting the issue.


KeenS commented Dec 27, 2013

I have the same problem and it seems the errors haven't resolved.

I'm also still dying on this, specifically on the step calling:

jruby -S mirahc -N --dest build/mirrors --classpath javalib/mirah-parser.jar:javalib/mirah-bootstrap.jar:javalib/mirah-compiler.jar src/org/mirah/jvm/mirrors/ src/org/mirah/jvm/model/

When I look in build/mirrors after the failure, the one compiled class is MirrorObjectExtensions$Extension1.class.
I looked in that file and found this:

package org.mirah.jvm.mirrors

import mirah.lang.ast.*

class MirrorObjectExtensions

  # This is needed by the Mirror type system,
  # but it breaks the ruby type system.
  macro def self.[],
                    true, true, @call.position)

Could that be what's causing things to break down here? I'm just shooting in the dark here.


baroquebobcat commented Jan 1, 2014

I think I figured it out. Basically, the task dependencies were in the wrong order and so the jars were being created at the wrong time. 0bb156e

Try it, and I'll close if it works.

@baroquebobcat baroquebobcat reopened this Jan 1, 2014

Still seeing the same issues. Fresh checkout, even tried a fresh install of JRuby to see if something was getting cached. Output here:


baroquebobcat commented Jan 2, 2014

Fun. At least it's a slightly different set of issues. By doing some diffing off of your gist, I think I have a hypothesis as to why it's still failing. The only big difference I see now is the order in which the files are being compiled.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment