Add .ppt (PowerPoint Binary File Format) read+write support

[Mearman]

Add read+write support for .ppt, the PowerPoint Binary File Format, as defined by [MS-PPT]. Unlike Word/Excel's stream-of-records-in-one-stream layout, PowerPoint's binary format is a tree of nested records (the "Current User" stream plus the PowerPoint Document stream's record hierarchy) inside an [MS-CFB] compound file — archive-codec's existing CFB reader already gets a codec to the raw stream bytes; parsing the record tree itself (and mapping it to/from the shared document-schema.js content schema, matching how ooxml.js's pptx support and odf.js's odp support both target the same schema) is the actual work.

Part of #85.

[Mearman]

Opened #846 (draft) with packages/ppt-codec — a read path only, so this issue stays open.

Shipped: the compound-file stream layer via archive-codec, the record-tree walk, the whole "live record" resolution ([MS-PPT] 2.1.2 Part 1: CurrentUserAtom → the UserEditAtom chain → the persist directory, later entries superseding earlier), DocumentAtom's slide size, the font collection, SlideListWithTextContainer, each SlideContainer's OfficeArt shape tree with anchors resolved through nested group coordinate systems, both text-body spellings with OutlineTextRefAtom indirection for placeholders, and StyleTextPropAtom's paragraph and character runs (indent level, alignment, bold/italic/underline/shadow/emboss, typeface reference, size, literal sRGB colour). Public surface is readPpt / readPptContent / readPptStreams, producing the same ContentSlide/ContentShape/ContentParagraph/ContentRun vocabulary as the pptx and odp readers.

Still open on this issue:

• The write path — nothing of it exists yet, and it is the larger half of the work: producing a valid persist directory and user edit, not just records.
• Speaker notes (the notes list and its NotesContainer persist objects).
• Document metadata — lives in the compound file's SummaryInformation stream ([MS-OSHARED]), outside every [MS-PPT] record.
• Master and layout style inheritance (TextMasterStyleAtom), and scheme colours (SlideSchemeColorSchemeAtom), both of which the current reader reports as "absent" rather than resolving.
• Per-shape text insets from OfficeArtFOPT.
• Images, tables, and OLE embeddings (ExObjListContainer, ExOleObjStg).
• Hyperlinks, bullets, paragraph spacing and margins — parsed past correctly today, not surfaced.
• Encryption — detected and refused by name, not implemented.
• Wiring into documents.js's conversion registry — deliberately separate, since it adds a dependency edge at the top of the graph.

The package README carries the same list in full, so it does not go stale here.

[Mearman]

Partial write support landed in #877: a genuine [MS-PPT] structure writer producing real presentation bytes via archive-codec's CFB writer, with every record's byte layout the direct inverse of the equivalent read function's own field order (verified by writing then reading back through this package's own independently-built reader, 38 new tests, including a Workers-isolate round trip).

Covered: one slide per input slide, each a plain text box, with per-run character formatting (bold/italic/underline, font family via a real FontCollectionContainer, size, literal RGB colour), paragraph alignment and list/indent level, and a real single-edit persist layer (CurrentUserAtom -> UserEditAtom -> PersistDirectoryAtom).

Still open: images/tables/OLE embeddings, grouped/rotated shapes, masters/layouts/scheme colours, speaker notes, document metadata, hyperlinks/bullets/spacing/margins, list numbering identity, strike (no MS-PPT bit exists for it), and animations/transitions/comments -- documented exhaustively in packages/ppt-codec/README.md's "What it does not write yet" section.

[Mearman]

ppt-codec is now reachable through documents.js's own conversion engine: convertDocument("ppt", "pdf", ...)/pdfToPpt/pptToPdf and every same-variant presentation bridge documents.js already supports (PR #881). document-cli/the web UI gained read-only PDF-preview support for the same reason. Still open: the codec's own remaining write scope (text-box slides only), and no live TUI editor.

[Mearman]

ppt-codec now reads and writes real document metadata via the same shared [MS-OLEPS] SummaryInformation support (PR #887).