EMBR-12903 fix(licenses): sort discovered license files for deterministic cross-platform notices

[overbalance]

What problem is this solving?

generate-license-file picks whichever license file glob returns first when a package ships more than one (e.g. import-in-the-middle ships both LICENSE and a LICENSE-3rdparty.csv dependency manifest). glob's order and casing differ across filesystems, so THIRD_PARTY_NOTICES.txt regenerated differently on macOS vs Linux CI and tripped the licenses:check drift gate. It also meant the committed notices listed the CSV manifest for import-in-the-middle instead of its actual license.

Short description of the changes

• Patch generate-license-file to sort candidate license files case-insensitively before picking, so the same file wins on every platform (upstream: fix: sort discovered license files for deterministic cross-platform output TobyAndToby/generate-license-file#743). A plain .sort() isn't enough here: glob yields LICENSE on Linux but license on macOS, so code-unit ordering still diverges.
• Apply the patch via patch-package. Since .npmrc sets ignore-scripts=true (no postinstall), the patch runs from the root licenses script, which licenses:check / validate / CI all funnel through.
• Regenerate packages/web-sdk/THIRD_PARTY_NOTICES.txt: import-in-the-middle now correctly resolves to its Apache license.

Chosen over depending on the fork directly because its package lives in a subdirectory of a private bun monorepo and isn't published to npm.

How was this tested?

• npm run licenses:check passes with no drift; macOS output matches the CI Linux blob byte-for-byte.
• npm run lint passes.

Checklist

• Covered by the existing licenses:check drift gate
• No changes to public API

[github-actions]

Chrome DevTools Protocol Tracing (Script: 121.39ms, Heap: 13.54MB)

	Number of Requests	Size of Requests	Script Duration	Task Duration	Heap Used Size
Requests	+13 requests	+41.70 KB
Page Loaded			+21.88 ms	+23.67 ms	+1.17 MB
Generate 100 fetch requests			+20.73 ms	+89.06 ms	+1.68 MB
Generate 100 XHR requests			+44.14 ms	+108.11 ms	+2.91 MB
Click 100 buttons and generate 100 logs			+26.05 ms	+50.66 ms	+2.25 MB
Throw a 100 exceptions			-1.08 ms	+26.43 ms	+2.29 MB
End Session			+9.67 ms	+37.03 ms	+3.24 MB
Total	+13 requests	+41.70 KB	+121.39 ms	+334.96 ms	+13.54 MB

Lighthouse (Script Eval: 135.59ms)

	Difference	Description
Total Blocking Time	+4 ms	Difference in Total Blocking Time: Sum of all time periods between FCP and Time to Interactive, when task length exceeded 50ms, expressed in milliseconds. Learn more about the Total Blocking Time metric.
Main Thread Time	+111.56 ms	Difference in Main Thread Time: Consider reducing the time spent parsing, compiling and executing JS. You may find delivering smaller JS payloads helps with this. Learn how to minimize main-thread work
Script Evaluation Time	+135.59 ms	Difference in Script Evaluation Time: Consider reducing the time spent parsing, compiling, and executing JS. You may find delivering smaller JS payloads helps with this. Learn how to reduce Javascript execution time.

Platform Tests (vite-7 es2015 gzip: 69.66KB)

vite-6 Platform Tests

	Total Uncompressed Size	Total Gzip Size
vite-6 - es2015	+203.00 KB	+69.70 KB

vite-7 Platform Tests

	Total Uncompressed Size	Total Gzip Size
vite-7 - es2015	+203.00 KB	+69.66 KB

webpack-5 Platform Tests

	Total Uncompressed Size	Total Gzip Size
webpack-5 - es2015	0 KB	0 KB