ECD alphabetization

This is more of a quality-of-life issue than an actual problem needing a solution (in case the solution is not trivial to execute) but ideally the ECD should be in alphabetical order. Right not it looks like it just scrapes the document for extended components in line order and then builds the section out of that, rather than re-ordering. This can be mitigated to some extent by having a single body text section for SFRs and then using tags to indicate those that are non-mandatory (rather than having them in appendices off by themselves at the end) but this trick doesn't work if you have a Module with multiple Base-PPs and the Additional SFRs sections of those Base-PPs define their own extended components that would be out of order w/r/t the rest of the list.

No file, test, or entry point is named. Start by tracing how the transforms scrape extended components and assemble the ECD section, including modules with multiple Base-PPs. Done means the resulting ECD is alphabetized consistently across those cases without relying on a single body-text section workaround.