Prefer Optional::orElseThrow over Optional::get

[schlosna]

Before this PR

#2946 (comment)

After this PR

==COMMIT_MSG==
Prefer Optional::orElseThrow over Optional::get
==COMMIT_MSG==

Possible downsides?

[changelog-app]

Generate changelog in changelog/@unreleased

What do the change types mean?

• feature: A new feature of the service.
• improvement: An incremental improvement in the functionality or operation of the service.
• fix: Remedies the incorrect behaviour of a component of the service in a backwards-compatible way.
• break: Has the potential to break consumers of this service's API, inclusive of both Palantir services
  and external consumers of the service's API (e.g. customer-written software or integrations).
• deprecation: Advertises the intention to remove service functionality without any change to the
  operation of the service itself.
• manualTask: Requires the possibility of manual intervention (running a script, eyeballing configuration,
  performing database surgery, ...) at the time of upgrade for it to succeed.
• migration: A fully automatic upgrade migration task with no engineer input required.

Note: only one type should be chosen.

How are new versions calculated?

• ❗The break and manual task changelog types will result in a major release!
• 🐛 The fix changelog type will result in a minor release in most cases, and a patch release version for patch branches. This behaviour is configurable in autorelease.
• ✨ All others will result in a minor version release.

Type

• Feature
• Improvement
• Fix
• Break
• Deprecation
• Manual task
• Migration

Description

Prefer Optional::orElseThrow over Optional::get

Check the box to generate changelog(s)

• Generate changelog entry

[schlosna]

Open question: Do we also want a more generic PreferOptionalOrElseThrow check to rewrite Optional::get & opt.get() to Optional::orElseThrow & opt.orElseThrow()

[pkoenig10]

I'd be in favor of this. Especially to fix consumers that have already applied StreamFlatMapOptional.

[pkoenig10]

I'm in favor of this, but would like @carterkozak to weigh in.

[carterkozak]

I suppose I'd mentally modeled Optional#get as "I've verified that the optional is present, and would like to extract the value", and Optional#orElseThrow as "Extract the value, but throwing is reasonable if the optional is empty", and sort of discounted using orElseThrow without args in favor of providing a more specific SafeLoggable exception. I'd never put much thought into this distinction, and I'm not sure it makes sense to have two ways of doing basically the same thing. It's worth verifying if Optional support in NullAway guards get() and orElseThrow() in the same way, or only checks get() but considers orElseThrow() as a guard.

[pkoenig10]

I've come to prefer orElseThrow() primarily for readability. With get(), it can be non-obvious that a particular piece of code can throw. This is especially valuable if:

• The code that checks if the value is present is, for whatever reason, further away from the code that extracts the value.
• The get() is called directly rather than with a method reference, so it's less obvious that the type of the value is an Optional. Particularly since we have a number of other types with get() methods, like Conjure aliases.

[pkoenig10]

orElseThrow() was introduced in JDK-8140281 for exactly this reason. The author ultimately plans to deprecated get() in JDK-8160606 - although I suspect that will never happen.

[carterkozak]

Given JDK-8160606, I agree. I suspect you're right that get may not actually be removed, but it seems reasonable to align with the JDKs design goals.

Unfortunately it seems that NullAway:CheckOptionalEmptiness only protects Optional.get(), not Optional.orElseThrow(). Only a few projects use this option, since we don't enable it by default anymore after some projects hit bugs: #2396
Probably worth looking into re-enabling, though. We've upgraded NullAway several times since that issue.