search

Gert-dev / php-ide-serenata

Atom IDE package that integrates the Serenata server to provide PHP code assistance
https://serenata.gitlab.io/
Other
275 stars 22 forks source link

Add busy-signal package dependency #502

Closed edcrouch closed 4 years ago

edcrouch commented 4 years ago

I was having the same issue as https://github.com/Gert-dev/php-ide-serenata/issues/490 with no packages installed other than those in Core, and the two dependencies listed for this package.

I experienced the issue through the "Test my setup" button on the post-installation pop-up that directs the user to install the Serenata server. After installing the busy-signal package and retrying the config test, the error went away.

I added the busy-signal package as a dependency, and then installed my fork of the php-ide-serenata package.

I was able to verify that the busy-signal package was installed along side the other package dependencies, and that the configuration test ran successfully with this change.

Gert-dev commented 4 years ago

Thanks for taking the time to investigate this. However, I think the busy-signal service should already be exposed by the atom-ide-ui package.

I think the API exposed there is compatible with the one from the busy-signal package, which would explain why it also works.

Perhaps the issue on the "Test my setup" popup is a timing issue with the service registration not having been called (yet).

edcrouch commented 4 years ago

Ah interesting, I didn't even think to look there. I usually use Visual Studio Code so I'm not familiar with the major Atom packages.

The busy signal service included with the atom-ide-ui package does work just fine in practice. I opened a large php project and can confirm that the indicator is running in the bottom bar.

I also tried:

  1. Installing busy-signal
  2. Running the config check (successfully)
  3. Uninstalling busy-signal
  4. Closing and reopening Atom

  5. Running the config check again (also successful)

    After all of that, the config check did not produce an error, which I thought was interesting. But as you said, perhaps it's some timing issue.

Anyway, thank you for taking the time to look into it as well, it looks like everything is working fine aside from the odd error condition. I'll go ahead and close the pull request.