ifcviewer: v16 zstd-compressed sidecars (~10x smaller over the wire)

The .ifcview data is hugely redundant (repeated double instance matrices,
patterned indices) — measured 12x zstd whole-file. Server Content-Encoding
can't be used (it breaks HTTP Range), so compress PER-CHUNK into the format.

Format (v16): geometry becomes per-chunk zstd(vertices)+zstd(indices) frames —
each independently Range-fetchable, so streaming is intact — and the critical +
deferred metadata blocks are single zstd frames. SidecarChunk carries the
compressed blob offsets/sizes; applyStreamedChunk (render/upload) is UNCHANGED —
decompression slots into the fetch. Full readSidecar (test/tooling) reconstructs
by decompress+scatter. zstd: desktop links libzstd (also compresses at bake);
the web build (Emscripten has no zstd port) FetchContent's the pinned zstd
source and compiles its decompress-only subset for wasm — no vendored blob,
same version as desktop. New SidecarCompress wraps it (compress guarded off
under Emscripten). Both stream paths — desktop StreamingThread worker + sync
fallback (readChunkGeometryCompressed) and web beginWebChunkLoad — decompress;
readSidecarMetadataOnly / the web bootstrap / loadDeferredMetadataWeb decompress
the metadata blocks. streamingByteProgress reports COMPRESSED bytes. MEASURED: a
752 MB v15 federation → 75 MB v16 (10x; per-file 6.7-15.3x); PP-PLP 118→15 MB,
loads 13/13 chunks on web, 0 errors.

Three fixes found while testing big federations on a real server:
- Web-streamed race: streaming_from_web was set in the deferred-header callback
  (a round-trip after the model+chunks exist), so driveStreamingLoads could take
  the sync fopen path meanwhile → "failed to read/decompress chunk 0". Now set
  immediately after applyCachedModel.
- OOM abort on 18 models: the pool grew unbounded until an alloc failed, but on
  web that's an uncatchable bad_alloc abort. Cap total pool capacity
  (setMaxTotalCapacity, 3 GB) so it stops before the heap ceiling, and raise
  MAXIMUM_MEMORY 2→4 GB (wasm32 max) for headroom.
- Web never evicted (grow-or-block only). At the hard budget, fall through to the
  LRU/priority evictor so a big federation stays navigable (highest-contribution
  chunks win) instead of freezing with holes.

113/113 desktop + 6/6 web smoke pass. No back-compat: regenerate sidecars
(desktop bakes v16; scratch conv tool migrates v15→v16).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

src/ifcviewer/ModelGpuData.h

@@ -177,6 +177,14 @@ struct ModelGpuData {
         // is recovered by walking mesh_ids and the model's MeshInfo[].
         uint64_t vertex_byte_size             = 0;
         uint64_t index_count                  = 0;
+        // v16: where this chunk's two zstd frames live in the file's geometry
+        // section (offsets relative to model.geometry_section_offset) and their
+        // compressed sizes. The raw sizes are vertex_byte_size / index_count*4.
+        // A per-chunk load fetches [off, +comp) and decompresses.
+        uint64_t v_comp_off                   = 0;
+        uint64_t v_comp_size                  = 0;
+        uint64_t i_comp_off                   = 0;
+        uint64_t i_comp_size                  = 0;
         // Of `index_count`, how many are LOD1 indices. LOD0 indices occupy
         // chunk-local u32 offsets [0, index_count - lod1_index_count); LOD1
         // indices occupy [index_count - lod1_index_count, index_count). 0
@@ -286,8 +294,9 @@ struct ModelGpuData {
     // streaming path: chunks may be non-resident and need byte-range reads
     // from this file. Empty path = legacy non-streaming load.
     std::string streaming_file_path;
-    uint64_t    streaming_vertex_section_offset = 0;
-    uint64_t    streaming_index_section_offset  = 0;
+    // v16: file offset of the compressed geometry section. A chunk's blobs are
+    // at geometry_section_offset + chunk.{v_comp_off,i_comp_off}.
+    uint64_t    geometry_section_offset = 0;
     // Web only: chunk byte ranges come from the JS-side source — a picked File
     // (Blob.slice) or a remote URL (HTTP Range) — read asynchronously, not via
     // a synchronous fopen on streaming_file_path. Set by loadSidecarMetadataWeb
@@ -308,8 +317,11 @@ struct ModelGpuData {
     // it; deferred_meta_loaded latches so it fetches at most once.
     std::vector<PackedElementInfo> elements;
     std::string                    string_table;
-    uint64_t    deferred_meta_offset = 0;
-    uint64_t    deferred_meta_bytes  = 0;
+    // v16: the deferred block is a single zstd frame at deferred_comp_offset of
+    // deferred_comp_size bytes, expanding to deferred_raw_size.
+    uint64_t    deferred_comp_offset = 0;
+    uint64_t    deferred_comp_size   = 0;
+    uint64_t    deferred_raw_size    = 0;
     bool        deferred_meta_loaded = false;
     // applyCachedModel rebases instance object_ids by this base to keep them
     // globally unique across models; deferred elements carry the sidecar's