I’m delighted that the current conflict is resolved by the combined updates!
Still, I think part of the motivation of @simon’s idea here was to maximize ease of troubleshooting if/when there is any kind of unexpected behavior. It’s sobering that even @jeremyruston was not able to grok the problem until I could rebuild an internal-core version that replicated the issue. In theory, any issue that requires inspecting the source code — even an issue that turns out to be unrelated to uglify — would be harder to troubleshoot if/when it affects projects that have been compressed this way.
Since uglify is not easily reversible, it seems wiser to me to have the tiddlyhost setup default to a non-uglified default for external-core wikis.
This compromise seems to hit the “sweet spot” of avoiding redundant server-side and bandwidth burdens — especially important for users like me who run multiple tiddlyhost projects, and who also store the backup history (grateful to @simon for making such things possible!) — while minimizing the risk of the code being virtually unparsable to human eyes when that’s what’s needed.
Surely, those who wish to use a compressed / uglified version of the already-lean external core can of course still do so, correct?