Fix appveyor

[dignifiedquire]

No description provided.

[dignifiedquire]

No idea why circle ci is unhappy :(

[Kubuxu]

It is happy for me 😄

[whyrusleeping]

@chriscool could we get some advice here? the sharness tests fail on windows CI because of path separator issues (at least, thats what it seems here). What would be a good way to do this?

[chriscool]

As far as I can see the first error is this:

expecting success: 
    echo "Error: failed to take lock at $IPFS_PATH: permission denied" > init_fail_exp &&
    test_cmp init_fail_exp init_fail_out

> diff -u init_fail_exp init_fail_out
--- init_fail_exp   2015-12-29 21:34:41.168675500 +0000
+++ init_fail_out   2015-12-29 21:34:41.168675500 +0000
@@ -1 +1 @@
-Error: failed to take lock at /cygdrive/c/go/src/github.com/ipfs/go-ipfs/test/sharness/trash directory.t0020-init.sh/.badipfs: permission denied
+Error: mkdir /cygdrive/c/go/src/github.com/ipfs/go-ipfs/test/sharness/trash directory.t0020-init.sh/.badipfs: The system cannot find the path specified.

not ok 3 - ipfs init output looks good

As permissions on Windows are working differently than on Linux/MacOS. It could be expected that the error message would not be the same on Windows.

Maybe this could be tested on Windows and either the error message we test against could be changed when testing on Windows or if the test is not relevant it could be skipped on Windows (maybe with a UNIXPERM prerequisite or something).

I don't have access to a Windows machine here. I will try to have a look at other test that are failing though.

[whyrusleeping]

I think we should merge this in for now. It still doesnt pass appveyor tests, but its gets a little farther and we can move forward from there

[dignifiedquire]

👍 for merging. At least actual tests are now failing

[GitCop]

There were the following issues with your Pull Request

• Commit: d2157f5
  • Invalid signoff. Commit message must end with
    License: MIT
    Signed-off-by: .* <.*>
• Commit: d586996
  • Invalid signoff. Commit message must end with
    License: MIT
    Signed-off-by: .* <.*>
• Commit: 596fc03
  • Invalid signoff. Commit message must end with
    License: MIT
    Signed-off-by: .* <.*>

We ask for a few features in the commit message for Open Source licensing hygiene and commit message clarity.
git commit --amend can often help you quickly improve the commit message.
Guidelines and a script are available to help in the long run.
Your feedback on GitCop is welcome on this issue.

---

This message was auto-generated by https://gitcop.com

[dignifiedquire]

@whyrusleeping just rebased and squashed. Can we get this merged now?

[jbenet]

@dignifiedquire is it actually fixed? it says:

Cannot find the tests' local ipfs tool.
Please check test and ipfs tool installation.
Makefile:29: recipe for target 't0010-basic-commands.sh' failed
make[1]: *** [t0010-basic-commands.sh] Error 1
make[1]: Leaving directory '/cygdrive/c/go/src/github.com/ipfs/go-ipfs/test/sharness'
Makefile:59: recipe for target 'test_sharness_expensive' failed
make: *** [test_sharness_expensive] Error 2
Command exited with code 2

• https://ci.appveyor.com/project/ipfsbot/test/build/561

[dignifiedquire]

It was :( but it looks like something went wrong/changed after the rebase.

[dignifiedquire]

@jbenet that was close, luckily I had the fixed version on my other machine. Should be better now

[jbenet]

were these changes contributed to iptb proper?

[dignifiedquire]

Yes, this was just pulled in from iptb

[jbenet]

ah ok!

[dignifiedquire]

ipfs/iptb@b21f95b is the commit I think

[jbenet]

Ok great! this LGTM. down to merge it before fixing the bugs. But before-- the iptb stuff should be updated in the iptb repo (https://github.com/whyrusleeping/iptb) and then update the dep here.

[jbenet]

Thanks @dignifiedquire, @whyrusleeping, @chriscool, @Kubuxu, and whoever else was involved -- it's a big deal to get this working :)

[dignifiedquire]

You have mostly to thank 32C3 dark room hack sessions :D