clasp push hangs if .claspignore file is present

[dandv]

This is odd given the basic level of functionality/breakage, but clasp push, which normally takes 5-10 seconds, hangs for over a minute if .claspignore is present. Whether it's empty, or it contains one line denoting a particular file (e.g. local-stuff.js, a file that does exist), the behavior is the same.

Expected Behavior

clasp push works fine:

$ ./node_modules/.bin/clasp push
⠋ Pushing files…
└─ /home/dandv/...
[...]
Pushed 10 files.

Actual Behavior

clasp push freezes/hangs:

$ ./node_modules/.bin/clasp push
⠹ Pushing files…





^C^C^C^C^C^C^C^C
^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C^C
⠸ Pushing files…

The ASCII spinner freezes for a while, then resumes spinning. I've run out of patience after 90+ seconds.

Steps to Reproduce the Problem

1. Create file .claspignore, even empty.
2. Run ./node_modules/.bin/clasp push

Specifications

• Node version (node -v): v18.18.0
• Version (clasp -v): 2.4.2
• OS (Mac/Linux/Windows): Fedora 38

[iRupam]

@dandv Hi, I've tried just now. It's working for me as expected. Anyways, is your issue fixed?

[zawadsoon]

I observed this issue as well while investigating this one: #968

I believe the process does not hang or freeze; it simply requires more time because every single file in your node_modules/ directory is iterated to mark files as ignored. If you have a large number of dependencies, this can take a considerable amount of time. In my case, I installed some development dependencies to format my files.

Try deleting the node_modules/ and .git/ directories, and then run clasp status. You should see a significant performance improvement.

I believe everyone who tries to do something serious using clasp struggles with this performance issue.

If I find the time, I will attempt to open a PR to optimize this, as I have been using clasp extensively recently. I have an idea that ignored directories could be skipped completely unless a specific flag is provided by the user.