Error: Process exited with no output, code: 0

[jonstaryuk]

Starting from v1.9.2, I get this error when javac exits with no errors/warnings. (If javac warns about something, the linter still works correctly.) I looked at sb-exec and the base linter package but couldn't figure out where exactly the problem came from...

Error: Process exited with no output, code: 0
    at ChildProcess.<anonymous> (/Users/jonathan/.atom/packages/linter-javac/node_modules/atom-linter/node_modules/sb-exec/lib/index.js:51:20)
    at emitTwo (events.js:87:13)
    at ChildProcess.emit (events.js:172:7)
    at maybeClose (internal/child_process.js:818:16)
    at Process.ChildProcess._handle.onexit (internal/child_process.js:211:5)

[florianb]

Thanks for reporting @jonstaryuk - i will investigate this issue. Sorry for the inconveniences.

[florianb]

@jonstaryuk i couldn't find a general reason for that issue. It seems like your javac-call doesn't return anything, but i don't know why. It would be helpful if you could run the lint with "Verbose Logging" set to true until the issue appears (please remind to turn it off afterwards) and copy & paste the console output (Window: Toggle Dev Tools) here.

I am assuming you're using Java 1.7/.8?

However - it seems we have currently a general issue in the package, too. I guess i will rewrite it the next week. 😩

[jonstaryuk]

Yes, javac version 1.8.0_65.

Manually running javac -Xlint:all *.java produces no output if there are no problems with the code. Is that behavior not normal?

Logs on Atom launch:

linter-javac: requiring modules finished.
linter-javac: providing linter, examining javac-callability.

Lint with no errors/warnings (throws error):

linter-javac: starting to lint.
linter-javac: start searching java-files with " /Users/jonathan/Code/class/cpe103/lab9 " as search-directory.
linter-javac: collected the following arguments: -Xlint:all
linter-javac: adding 2 files to the javac-arguments (from " /Users/jonathan/Code/class/cpe103/lab9/DirectedGraph.java " to " /Users/jonathan/Code/class/cpe103/lab9/DirectedGraphDriver.java ").
linter-javac: calling javac with 5 arguments by invoking " javac ". The approximated command length is 138 characters long, the last argument is: /Users/jonathan/Code/class/cpe103/lab9/DirectedGraphDriver.java

Lint with errors/warnings (works correctly):

linter-javac: starting to lint.
linter-javac: start searching java-files with " /Users/jonathan/Code/class/cpe103/lab9 " as search-directory.
linter-javac: collected the following arguments: -Xlint:all
linter-javac: adding 2 files to the javac-arguments (from " /Users/jonathan/Code/class/cpe103/lab9/DirectedGraph.java " to " /Users/jonathan/Code/class/cpe103/lab9/DirectedGraphDriver.java ").
linter-javac: calling javac with 3 arguments by invoking " javac ". The approximated command length is 132 characters long, the last argument is: /Users/jonathan/Code/class/cpe103/lab9/DirectedGraphDriver.java
linter-javac: parsing:
 /Users/jonathan/Code/class/cpe103/lab9/DirectedGraphDriver.java:23: error: incompatible types: int cannot be converted to String
          graph.addEdge(v1, v2);
                            ^
Note: Some messages have been simplified; recompile with -Xdiags:verbose to get full output
1 error
linter-javac: detecting languages
linter-javac: detected the following language-code: en
linter-javac: returning 1 linter-messages.

(oops, didn't mean to click close)

[florianb]

No worries 😄 -- thanks for the detailed response!

Okay as far as i understood the issue this seems to belong to the fact that this package expect javac to respond to stderr. And the exception is thrown because stderr didn't contain anything.

Since javac behaves that way (so this is normal) i need to find out why your javac-call complains about the missing stderr-output and mine not.

[winstliu]

I can confirm this on two different Windows computers - one running Atom 1.9.0-dev-d71f6e5/Windows 10 and the other Atom 1.7.4/Windows 7. Both are using linter-javac 1.9.3, though like @jonstaryuk said this exception started happening around 1.9.0 or so. The 1.9.0-dev laptop is using Java 1.8.0_92 while the 1.7.4 one is using Java 6 or 7 (not too sure; I can check tomorrow).

Happy to provide any other information and debug if necessary!

[florianb]

@50Wliu thank you very much - it seems like this is a Windows-related bug. I guess i need to set AppVeyor up. I will need some more time - please be patient. :)

[jskenney]

@florianb I am having the same issue with ubuntu 16.04 and java version "1.8.0_91"

[florianb]

@jskenney - thanks a lot, these are bad news. Since it seems there are several people affected, i will try to release a hotfix. 🐞

[Arcanemagus]

FYI @florianb this should be as simple as setting the allowEmptyStderr flag to true on the exec() options (and making sure the code doesn't blow up with an empty output 😛).

[florianb]

Thanks @Arcanemagus - i already thought about enabling allowEmptyStderr, but i currently can't test that behavior. I would prefer to know what's happening.

May it be that we all will face that issue as soon as steelbrain's exec is rolled out to everybody? I thought greenkeeper would also update the package-dependencies to the explicit latest versions.