Travis tests failing all of a sudden

[gibsramen]

Relevant to #106

Looks like something has gone wrong with the Travis build. After submitting the aforementioned PR, I noticed that the Travis build was failing with the following error:

ImportError (/usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by /home/travis/miniconda/envs/test_env/lib/python3.6/site-packages/scipy/fft/_pocketfft/pypocketfft.cpython-36m-x86_64-linux-gnu.so))

@gwarmstrong suggested making a small change to the current (passing) Songbird master branch which I did here. This change should have no effect on the Travis build, yet it fails with the same error. (Test branch Travis build).

EDIT: Current passing build

[gwarmstrong]

Found this relevant thread

[mortonjt]

Sorry, I'm quite swamped and can't look at this for the next couple of weeks. But thanks for uncovering this

…

On Tue, Jan 21, 2020, 7:17 PM George Armstrong ***@***.***> wrote: Found this <https://askubuntu.com/questions/575505/glibcxx-3-4-20-not-found-how-to-fix-this-error> relevant thread — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#107?email_source=notifications&email_token=AA75VXPT6SZLGOIPBAUHFEDQ66GADA5CNFSM4KJ5LT62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEJRYRYQ#issuecomment-576948450>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXPT7AHC4O7JYSRJJLLQ66GADANCNFSM4KJ5LT6Q> .

[gibsramen]

No problem - completely understand. Will work on our end to diagnose/fix the problem(s) and minimize effort on your part when you get the chance.

[fedarko]

Just to update this thread with the progress @gibsramen and @gwarmstrong have made -- looks like there are (at least) two three solutions that solve this problem:

1. sudo apt-get installing / upgrading a few GCC/C++-related packages

   • see George's commit here
   • solution described here

2. pip uninstalling and then reinstalling scipy

   • see Gibs' commit here
   • solution described here

3. bumping up the minimum tensorflow versions

   • see my PR here

Since any of these solutions works, I don't think which we use matters that much. Mostly I just wanted to document these solutions in a visible place here, since I imagine other people will run into these sorts of problems eventually -- it's likely worth updating the README FAQs at some point... :)

[fedarko]

This has been closed now that #108 has been merged in. Thanks @gwarmstrong!

[mortonjt]

Thanks!

…

On Wed, Feb 5, 2020, 5:57 PM Marcus Fedarko ***@***.***> wrote: Closed #107 <#107>. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#107?email_source=notifications&email_token=AA75VXKWJEM6BHIL3PTL2DTRBM743A5CNFSM4KJ5LT62YY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOWODA5TY#event-3011907279>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AA75VXJLWHCMCP6Z7H4CJM3RBM743ANCNFSM4KJ5LT6Q> .