Source release bundles third-party doc-site assets not accounted for in LICENSE

[nssalian]

Feature Request / Improvement

The source release tarball bundles third-party assets under site/docs/assets/ that aren't listed in the root LICENSE and whose license texts aren't included:

File	Origin	License in file
stylesheets/bootstrap-grid.css	Bootstrap Grid v5.3.0-alpha3	copyright + MIT link, no MIT text
stylesheets/termynal.css	termynal (Ines Montani)	@license MIT, no MIT text
javascript/termynal.js	termynal (Ines Montani)	@license MIT, no MIT text
javascript/lottie-player.js	Lottie player (minified bundle)	no banner; embeds BSD-3-Clause + Google/Microsoft copyrights

(stylesheets/bootstrap-buttons.css looks Bootstrap-derived but has no header, so its origin is worth confirming.)

This isn't a regression. The same assets ship in 1.9.0/1.10.0/1.11.0. It was raised on the 1.12.0 RC0 vote. They got in because dev/source-release.sh uses git archive, and .gitattributes export-ignores /build, /examples, /docs but not site/.

Reproduced:

git archive dec392570a88e45ac3a417a487dd72f9bdb629d5 --worktree-attributes --prefix apache-iceberg-1.12.0/ -o /tmp/repro.tar.gz
tar tzf /tmp/repro.tar.gz | grep site/docs/assets/

Possible fixes:

• add /site/docs/** export-ignore to .gitattributes (these aren't needed to build from source)
• or include the license texts and add root LICENSE entries

Refs: https://infra.apache.org/licensing-howto.html . Prior discussion thread

Query engine

Other (source release / build tooling)

Willingness to contribute

• I can contribute this improvement/feature independently
• I would be willing to contribute this improvement/feature with guidance from the Iceberg community
• I cannot contribute this improvement/feature at this time

[RussellSpitzer]

I guess just ignore? I'm not quite sure on this because I'm pretty sure we should be doing licensing for everything we ship (including docs) but it seems like we are doing that already.

[kevinjqliu]

I think we should just ignore those files and not package them. We currently do the same for the files under docs/**

iceberg/.gitattributes

Lines 30 to 31 in e45573b

	/docs export-ignore
	/docs/** export-ignore

Lets just ignore everything under site/**

And I think this might mean we need a new RC

[nssalian]

Should be an easy fix. I'll put up a PR. I'll wait for the vote to close so we can check if more fixes are needed before cutting another RC