fix(featureDev): Prevent crash on large repos or unsupported encoding during file collection

[jsalles]

Problem

There are currently 2 problems in featureDev when doing collection of repository files:

• Trying to decode binary files might crash the extension host, and
• too large repos might make the extension run out of available memory.

Solution

This CR fixes both problems by:

• Adding a maximum allowed repo size of 200mb for collection
• Changing the behavior of the TextDecoder to replace the unsupported characters instead of throwing TypeError

Users who try to use featureDev on a codebase larger than 200mb will see a message asking for a different workspace to be selected.

License

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[jpinkney-aws]

Just one more thing, are you able to add a changelog? I'm sure theres some users that would like to know this is fixed 😄

[jsalles]

Just one more thing, are you able to add a changelog? I'm sure theres some users that would like to know this is fixed 😄

Yes, definitely! Just added a changelog with the bugfix :)

[justinmk3]

the current approach used to zip files does not "stream" the zip creation so will also run into memory constraints.

please look at #4769 , can you rewrite this feature on top of that (after it lands)?

[grimalk]

What's the ETA of this being ready? Presume we want both these in the same wave of the toolkit?

[jsalles]

Yeah I'm happy to change our implementation to use the stream after it's merged.

[justinmk3]

What's the ETA of this being ready?

The zip streaming won't be ready today, but hopefully next week. @ctlai95