jbeldock

I am attempting to add a customer search to a website that requires me to use my enterprise credentials to log in via SSO.  Typically, I simply enter the target website URL and the site checks if I have an auth token. If I do, I am logged in. If I do not, then I have to sign in via my enterprise credentials.

When I enter this URL into the "Add Customer Search" dialog in Alfred, the URL remains orange, and the save button is disabled. I suspect this is because Alfred is trying to validate the URL somehow.  See this image:

Since the Alfred app doesn't have an auth token for this site, it will be redirected, just as it is when I try to look up the search, so I suspect this is causing the app not to accept the URL.

Any ideas on how to skip the validation step?

 

Problem was the script being hard-coded to

!

~/Library/Caches/com.runningwithcrayons.Alfred-2/Workflow Data

A simple symlink solved the problem:

ln -s com.runningwithcrayons.Alfred-3/ com.runningwithcrayons.Alfred-2

Have started to debug this. I see the following error in the Alfred workflow debugger:

[2016-11-11 16:47:14][ERROR: input.scriptfilter] Code 1: 0:90: execution error: mkdir: /Users/jbeldock/Library/Caches/com.runningwithcrayons.Alfred-2/Workflow Data: No such file or directory (1)

I'm experiencing very odd behavior. If I attempt to select a workflow using a keyword, I get the default search instead, even though the workflow keyword shows up in the Alfred menu and I select it. For example, I use the ens keyword to access the installed "Evernote" workflow. When I select "Search..." from the Alfred Dropdown (see image 1), I very briefly see "Searching..." appear, but then it appears to time out almost instantaneously and then I see the fallback search options (see image 2).

Similarly, if I type a workflow keyword, Alfred very briefly shows the keyword result, but then immediately switches to the fallback search options. I am unable to use the workflow keyword, and whatever i type in {query} is simply fed into the fallback query.

Is there some kind of super-short timeout being applied here? Is there a way to adjust this?