No such file or directory @ rb_sysopen - tmp/rspec_guard_result with a syntax error

[dima4p]

In the case of syntax error in the code RSpecProcess does not create formatter_tmp_file, so Guard::RSpec::Results.new generates an exception and guard fails.

Sorry, I can not now create PR, but the problem is fixed by writing the line results.rb#11 as follow:
lines = File.readlines(filename) rescue ['Syntax error']

[e2]

Can you at least hint what kind of syntax error causes this?
(Not every syntax errors results in this error).

Also, the fix needs to be handled at a higher level (I have an idea for this, but I need to reproduce this).

I can't accept this fix due to:

1. Other file related errors are hidden
2. Silencing the errors with no details of what failed (Generic rescue statements like this were used before - and they caused extremely hard to debug problems)

I can't reproduce this problem. The syntax errors I come up with just show up as syntax errors.

[dima4p]

yes, in a model
validates :name, presence: tru

[dima4p]

I agree, that higher level is better ;-)

[e2]

I can't reproduce this.

I'm correctly getting: undefined local variable or method 'tru' for #<Class:0x000000062e0b98> (NameError)

It may depend how model classes are loaded, rails version, spring, etc.

[ryan2johnson9]

thanks dima4p for getting me back up and running, even if it is not ideal..
I got this error after changing back to MRI-2.3 from rbx-2.5.7
in case it helps
I'm running

spring-1.7.1
guard-2.13.0
guard-rspec-4.6.5
rspec-3.4.0
rspec-rails-3.4.2
rails-4.1.6

[e2]

Gosh, I'd be happy to fix this quickly (and release a fix) if I could only reproduce this.

Could someone copy a failing project, remove all unnecessary files, and upload it as a repo so I can reproduce this and take a look?

[dima4p]

Please, try to run with Spork. I meet this problem in the projects that use it.

[e2]

@dima4p - could you create a small repo to reproduce the problem so I can work on this?

If someone invests a bit of time into creating and sharing a repo with the issue, I'll invest some time into fixing this and releasing a fixed version.

(I've been debugging similar issues with Zeus and it's just too easy get mixed up - so an example repository is very, very helpful).

[e2]

I've released v4.7.0 with extensive debugging to help resolve the issue.

So I'll close this. If it still occurs:

1. Open a new issue
2. Attach the full info from bundle exec guard -d.

[ryan2johnson9]

Thanks @e2 I will give it a try when I can. I did try to get a simple project going to illustrate the error, but could not reproduce it. Made me think it was some kind of syntactical issue within one of my many specs in the problematic project. I tried systematically commenting out specs and simplifying my guardfile but alas, the error did not go away.

[e2]

With the new guard-rspec release and running under bundle exec guard -d, the output alone should be enough for me to know what's going on.

Hope that helps nails this.

On Sun, May 22, 2016 at 04:28:42PM -0700, Ryan Johnson wrote:

Thanks @e2 I will give it a try when I can. I did try to get a simple project going to illustrate the error, but could not reproduce it. Made me think it was some kind of syntactical issue within one of my many specs in the problematic project. I tried systematically commenting out specs and simplifying my guardfile but ALAS, the error did not go away.

---

You are receiving this because you were mentioned.
Reply to this email directly or view it on GitHub:
#362 (comment)

[ryan2johnson9]

I updated to 4.7.2 and I no longer get the error. Thanks. sorry it took me so long.

[e2]

Glad you have it work. Thanks for reporting.