Add a failing test for version conflicts

[magicmark]

Summary
This is a failing test to highlight the issue the issue here #3780 https://github.com/mxmul/yarn_flat_regression

tl;dr:

a => b => left-pad@>=0.0.1
a => left-pad@1.0.0

yarn fails when doing --flat with the following:

info Unable to find a suitable version for "left-pad", please choose one by typing one of the numbers below:
  1) "left-pad@>=0.0.1" which resolved to "1.1.3"
  2) "left-pad@1.0.0" which resolved to "1.0.0"

(yarn version: 0.27.1)

(this seems to be a regression introduced in 0.27.0, this worked in 0.26.1)

Test plan
This is a test only change.

---

This test is currently failing because it's prompting for a resolution, with the following stacktrace

(node:23637) PromiseRejectionHandledWarning: Promise rejection was handled asynchronously (rejection id: 2)
 FAIL  __tests__/commands/install/integration.js (264.799s)
  ● unbound transitive dependencies should not conflict with top level dependency

    Error: Can't answer a question unless a user TTY
      
      at ConsoleReporter.select (src/reporters/console/console-reporter.js:322:29)
      at src/cli/commands/install.js:932:182
      at next (native)
      at step (node_modules/babel-runtime/helpers/asyncToGenerator.js:17:30)
      at node_modules/babel-runtime/helpers/asyncToGenerator.js:35:14
      at Promise.F (node_modules/core-js/library/modules/_export.js:35:28)
      at node_modules/babel-runtime/helpers/asyncToGenerator.js:14:12
      at Install.flatten (src/cli/commands/install.js:933:2215)
      at src/cli/commands/install.js:922:3570
      at next (native)
      Console output:
       warning package.json: No license field
      warning No license field
      [1/4] Resolving packages...
      
      at __tests__/commands/_helpers.js:144:13
      at throw (native)
      at step (node_modules/babel-runtime/helpers/asyncToGenerator.js:17:30)
      at node_modules/babel-runtime/helpers/asyncToGenerator.js:30:13

[bestander]

did you mean 1.0.1?

[magicmark]

Don't think so

I think >=0.0.1 and 1.0.0 should not conflict?

This test shows that they do conflict.

(looks like >=0.0.1 resolves to a hard pin for whatever the latest version is?)

This behaviour was not present prior to 0.27.0

[bestander]

I see, thanks.
This is related to this task in general then #3778.
Let's make it passing for now and link that it should be reversed when #3778 is resolved.

Do you want to have a go at implementing the post processing after resolution?

[magicmark]

@bestander Skipped the test, added comment linking to issue.

I'll have a go at #3778 :)

[bestander]

Awesome!

…

On Sun, Jul 2, 2017 at 10:19 AM Mark Larah ***@***.***> wrote: @bestander <https://github.com/bestander> Skipped the test, added comment linking to issue. I'll have a go at #3778 <#3778> :) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3779 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACBdWO0zurtbtqml4vzAVkbRQeojzVHMks5sJ9E3gaJpZM4OLEIW> .

[bestander]

Merged this, ping me on a fix :)