Bump to OCCT 8.0.1

Four mechanical binding edits; the bound Python surface is unchanged
(sigdiff compared the 7.9.3.1 dump against 8.0.1.1's in both directions
and found no class, member or constructor arity moved), so the app needs
no change.

- NCollection_Utf8String is gone; NCollection_String is the same UTF-8
  type. Python name kept, since the app imports it.
- Standard_Failure lost its Standard_Transient RTTI when it moved to
  deriving std::exception; ExceptionType() replaced DynamicType()->Name()
  and returns the same class names the exception dispatch keys on.
- FindKey gained a size_t overload beside the int one, so the plain
  member pointer is ambiguous; overload_cast picks the int form, which
  keeps the negative-index guard.
- StdPrs moved from TKService to TKV3d, so CMakeLists links both.

The byte-identity fixtures did NOT retire, contrary to the watchlist:
all six round-trip to the same digests under the new kernel, as do the
measurement, history and mesh-count blocks, so tests/data is untouched
and the app's content-addressed BREP payloads stay valid.

Gates: binding suite 83 passed, 138/138 symbols, sigdiff clean both
directions, ASAN clean, app suite 1798 passed / 1 skipped (unchanged),
sweep over 4486 parts with anchors_digest and every status unchanged,
-m perf 16% faster on the 32-feature chain.

CMakeLists.txt

@@ -70,10 +70,13 @@ target_link_libraries(_OCP PRIVATE
   TKDESTEP     # STEPControl
   TKDEIGES     # IGESControl, IGESData
   TKDESTL      # RWStl
-  # Inc 4 — TKService carries the BRep font/text builders.  The builder image
+  # Inc 4 — the BRep font/text builders.  OCCT 8.0 moved the StdPrs package
-  # asserts it links without libGL/libX11, so this costs the runtime image
+  # from TKService to TKV3d (Font_BRepFont is now the typedef of
-  # nothing.
+  # StdPrs_BRepFont, not the other way round), so both are needed.  The builder
+  # image asserts each links without libGL/libX11, so this costs the runtime
+  # image nothing.
   TKService
+  TKV3d
 )
 
 target_compile_definitions(_OCP PRIVATE

Makefile

@@ -1,4 +1,4 @@
-OCCT_VER ?= 7.9.3
+OCCT_VER ?= 8.0.1
 IMG_N    ?= 1
 IMAGE    ?= git.stroblme.de/n3xd/occt-build:$(OCCT_VER)-$(IMG_N)
 # Build cache lives off the root filesystem, which is tight on this host.

README.md

@@ -7,14 +7,16 @@ backend uses — 138 symbols across 47 `OCP.*` modules, not all of OCCT.
 The package installs as a top-level `OCP`, so it is a drop-in replacement for
 `cadquery-ocp-novtk` and the app's 442 import sites stay untouched.
 
-Status: **coverage complete** — all 138 symbols the app imports, across 53
+Status: **in production use** — all 138 symbols the app imports, across 53
-bound modules, published as `7.9.3.1.dev5`. The app's full backend suite passes
+bound modules, published as `8.0.1.1` (OCCT 8.0.1). The app cut over on
-against it (1797 passed / 1 skipped, the same as the stock wheel), a sweep of
+2026-08-10 at `7.9.3.1` and took the kernel bump on 2026-08-11; its full
-the whole project store reproduces every part's geometry exactly (4486 parts;
+backend suite passes against both (1798 passed / 1 skipped, the same as the
-no change in statuses, volume, area, bbox, entity counts, triangles or anchor
+stock wheel it replaced), a sweep of the whole project store reproduces every
-digests), and BREP serialisation stays byte-identical, which the pools and the
+part's geometry exactly (4486 parts; no change in statuses, volume, area,
-content-addressed derive payloads depend on. The app is not swapped yet — that
+bbox, entity counts, triangles or anchor digests), and BREP serialisation
-is the cutover, roadmap 10C.
+stays byte-identical, which the pools and the content-addressed derive
+payloads depend on — across the kernel bump as well, so nothing stored had to
+be rewritten.
 
 Start with [docs/design.md](docs/design.md) for the decisions,
 [docs/building.md](docs/building.md) to build one, and
@@ -24,7 +26,7 @@ phase plan lives in the app repo at `docs-private/reference/roadmap.md`
 
 ## Why
 
-`cadquery-ocp` lags OCCT (it wraps 7.9.3; OCCT 8.0 shipped 2026-05), builds
+`cadquery-ocp` lags OCCT (it still wraps 7.9.3; we are on 8.0.1), builds
 Windows and macOS wheels we never use, and until recently forced a 638 MB VTK
 dependency into the image. So this exists for version velocity, footprint, and
 two defects that a binding we control prevents by construction:
@@ -54,7 +56,7 @@ production host (4 cores, and a kernel build is multi-hour). Wheels are built
 here on a dev box and published to the Gitea package registry.
 
 ```bash
-make image    # once, ~40 min: compiles OCCT 7.9.3 into the builder image
+make image    # once, ~40 min: compiles OCCT 8.0.1 into the builder image
 make dev      # inner loop: incremental compile + tests
 make wheel    # compile, stubs, auditwheel, self-containment smoke test
 make publish  # -> https://git.stroblme.de/api/packages/N3XD/pypi

docs/building.md

@@ -6,7 +6,7 @@ and (for publishing) `uv`.
 
 ## The builder image
 
-`occt/Dockerfile` compiles OCCT 7.9.3 once inside
+`occt/Dockerfile` compiles OCCT 8.0.1 once inside
 `quay.io/pypa/manylinux_2_28_x86_64` (both the base digest and the source
 tarball's sha256 are pinned) and installs it to `/opt/occt`.
 
@@ -66,7 +66,15 @@ cd ../app && .venv/bin/python ../ocp/tools/gen_fixtures.py
 
 The generator asserts stock is idempotent for each fixture before recording its
 digest — otherwise the gate would compare against a moving target. Regenerate
-only when deliberately re-blessing (e.g. the OCCT 8.0 bump).
+only when deliberately re-blessing.
+
+The OCCT 8.0.1 bump was expected to be such an occasion and turned out not to
+be: every fixture round-tripped to the same digest under the new kernel, as did
+the measurement, history and mesh-count blocks, so `tests/data/` was left
+untouched. Do not assume the next bump re-blesses either — run the gate first
+and let it say. Note also that the generator needs a wheel carrying more of
+OCCT than this binding exposes (it is written against the stock wheel), so
+regenerating is not currently possible from the app's own venv.
 
 ## Publishing
 

docs/design.md

@@ -236,26 +236,48 @@ were gated:
   `IsDeleted` maps compared exactly, since that is the substrate the app's
   topological naming is built on.
 
-## OCCT 8.0 watchlist
+## OCCT 8.0 bump — what actually moved
 
-Recorded as each module landed, so the 10E bump is a known quantity:
+Done at 8.0.1 (2026-08-11). Four edits across ~4 000 lines of binding, and the
+bound Python surface came out **identical** — `sigdiff` compared 7.9.3.1's dump
+against 8.0.1.1's in both directions and found no class, member or constructor
+arity changed, so the drop-in contract held without the app being touched.
 
-- `Standard_Failure` derives `std::exception`, and the `Raise`/`Throw`/
+What broke, all of it mechanical:
-  `Instance` static helpers are gone. None of them is bound, so the exception
+
-  table should port unchanged.
+- **`NCollection_Utf8String` is gone.** `NCollection_String`
-- `GeomLProp` is superseded by the new `GeomProp`/`BRepProp` packages, whose
+  (`NCollection_UtfString<char>`) is the same UTF-8 type and is what the font
-  results come back as structs with `IsDefined` flags instead of throwing. Only
+  API takes. The Python name is unchanged, since the app imports it.
-  `GeomLProp_SLProps.Normal` is bound, so this is a one-line move.
+- **`Standard_Failure` lost its `Standard_Transient` RTTI** when it moved to
-- `StdPrs_BRepFont` and `StdPrs_BRepTextBuilder` become deprecated typedefs of
+  deriving `std::exception`, so `DynamicType()->Name()` no longer compiles. The
-  `Font_BRepFont`/`Font_BRepTextBuilder`. They are real classes in 7.9.3 and are
+  virtual `ExceptionType()` replaced it and returns the same class names, which
-  bound as such; 8.0 makes it a rename behind the same `OCP.StdPrs` names.
+  is what the exception dispatch keys on — and what `cad_pool` marshals home.
-- Methods that returned handles through an out-parameter return them by value.
+- **The `Size()` → `size_t` migration added index overloads.** `FindKey` on the
-  `BRep_Tool.Surface`/`Curve` and the adaptors are the ones this binding
+  indexed maps now has both an `int` and a `size_t` form, so a plain member
-  touches.
+  pointer is ambiguous; `nb::overload_cast<Standard_Integer>` picks the `int`
-- The global math wrappers are deprecated in favour of `std::`. Nothing here
+  one, which keeps the negative-index guard.
-  binds them.
+- **`StdPrs` moved from TKService to TKV3d**, so `CMakeLists.txt` links both.
-- The byte-identity fixtures retire at the bump: a different kernel legitimately
+
-  writes different BREP bytes. The measurement and history blocks stay.
+Three watchlist predictions were wrong, recorded here because the reasoning
+behind them was plausible:
+
+- `StdPrs_BRepFont`/`StdPrs_BRepTextBuilder` did **not** become typedefs of
+  `Font_BRepFont`/`Font_BRepTextBuilder`. The aliasing runs the other way — they
+  are still the real classes and the `Font_*` names are the typedefs.
+- `GeomLProp` was **not** superseded by `GeomProp`/`BRepProp`. It became a
+  template (`GeomLProp_SLProps` is now an alias for
+  `GeomLProp_SLPropsBase<handle<Geom_Surface>>`) and kept the constructor and
+  accessors this binding uses, so the module compiled unchanged.
+- **The byte-identity fixtures did not retire.** A different kernel was assumed
+  to write different BREP bytes; 8.0.1 does not. All six fixtures round-trip to
+  the same digests, and the measurement, history and mesh-count blocks match
+  too, so `tests/data/` was left alone and no re-bless happened. That also means
+  the app's content-addressed BREP payloads stay valid across the bump.
+
+Unchanged as predicted: the handle model (`occt_handle.h` needed nothing, ASAN
+clean), the exception *table* (the removed `Raise`/`Throw`/`Instance` helpers
+were never bound), the deprecated out-parameter handle returns this binding
+still uses, and the global math wrappers it never bound.
 
 ## Open questions
 

occt/Dockerfile

@@ -16,8 +16,8 @@ FROM quay.io/pypa/manylinux_2_28_x86_64@sha256:f854c50adf7b7a325bc4794316f3758d3
 RUN dnf install -y freetype-devel fontconfig-devel ninja-build ccache valgrind \
     && dnf clean all
 
-ARG OCCT_TAG=V7_9_3
+ARG OCCT_TAG=V8_0_1
-ARG OCCT_SHA256=5ecf094ec6b12d5413dfb851d8c3590c354058aee556e32e408bdfbf8c357d57
+ARG OCCT_SHA256=0d6913eae4bcc09a3653ceced6dda1aec11c35a1513d4c06762c9b002092c68a
 
 RUN curl -fsSL -o /tmp/occt.tar.gz \
       "https://github.com/Open-Cascade-SAS/OCCT/archive/refs/tags/${OCCT_TAG}.tar.gz" \
@@ -43,7 +43,6 @@ RUN cmake -G Ninja -S /src -B /build \
       -DBUILD_MODULE_DataExchange=ON \
       -DBUILD_MODULE_ApplicationFramework=ON \
       -DBUILD_MODULE_Draw=OFF \
-      -DBUILD_MODULE_DETools=OFF \
       -DUSE_FREETYPE=ON \
       -DUSE_VTK=OFF -DUSE_TK=OFF -DUSE_TCL=OFF \
       -DUSE_XLIB=OFF -DUSE_OPENGL=OFF -DUSE_GLES2=OFF \

pyproject.toml

@@ -9,7 +9,7 @@ name = "n3xd-ocp"
 # .devN suffix: the registry never allows republishing a version, and its
 # cleanup rule collects only the dev ones.  CMake asserts the prefix matches
 # the OCCT it found.
-version = "7.9.3.1"
+version = "8.0.1.1"
 description = "nanobind bindings for the OpenCASCADE geometry kernel (drop-in OCP)"
 readme = "README.md"
 requires-python = ">=3.12"

src/common/occt_exceptions.cpp

@@ -90,9 +90,12 @@ void register_occt_exceptions(nb::module_ &root) {
             } catch (const Standard_Failure &e) {
                 // Dispatch on the dynamic OCCT type so the Python type name
                 // matches what the kernel actually threw; anything we did not
-                // bind falls back to the RuntimeError-derived base.
+                // bind falls back to the RuntimeError-derived base.  OCCT 8.0
+                // dropped the Standard_Transient RTTI here (Standard_Failure
+                // now derives std::exception); ExceptionType() is the virtual
+                // that replaced it and returns the same class names.
                 PyObject *type = g_base;
-                const char *name = e.DynamicType()->Name();
+                const char *name = e.ExceptionType();
                 if (name) {
                     auto it = g_exc_types.find(name);
                     if (it != g_exc_types.end())

src/modules/mod_GeomLProp.cpp

@@ -4,9 +4,9 @@
     One use: text emboss reads the surface normal at the anchor's UV to decide
     which way the glyph solid grows out of a curved face.
 
-    OCCT 8.0 supersedes this package with GeomProp/BRepProp (results come back
+    OCCT 8.0 rebuilt this as a template — GeomLProp_SLProps is now an alias for
-    as structs carrying IsDefined flags instead of throwing).  Only Normal()
+    GeomLProp_SLPropsBase<handle<Geom_Surface>> — but kept the constructor and
-    is bound, so the port is a one-line move — see design.md's 8.0 watchlist.
+    the accessors we bind, so the surface below is unchanged.
 */
 
 #include "../common/occt_module.h"

src/modules/mod_NCollection.cpp

@@ -1,9 +1,13 @@
 /*
     OCP.NCollection — the UTF-8 string the font API takes.
 
-    NCollection_Utf8String is the only NCollection type the app names directly:
+    This is the only NCollection type the app names directly: it wraps the font
-    it wraps the font path and the text to emboss.  Constructed from a Python
+    path and the text to emboss.  Constructed from a Python str, which nanobind
-    str, which nanobind hands over as UTF-8 already.
+    hands over as UTF-8 already.
+
+    OCCT 8.0 dropped the NCollection_Utf8String alias; NCollection_String is the
+    same type (NCollection_UtfString<char>) and is what the font API takes.  The
+    Python name stays as it was — it is what the app imports.
 */
 
 #include "../common/occt_module.h"
@@ -13,20 +17,20 @@
 void register_NCollection(nb::module_ &root) {
     nb::module_ m = ocp_submodule(root, "NCollection");
 
-    nb::class_<NCollection_Utf8String>(m, "NCollection_Utf8String")
+    nb::class_<NCollection_String>(m, "NCollection_Utf8String")
         .def(nb::init<>())
         // Upstream also offers an explicit-length overload; the app always
         // passes a whole Python str, which arrives NUL-terminated, so only the
         // one-argument form is bound.
         .def(
             "__init__",
-            [](NCollection_Utf8String *self, const char *text) {
+            [](NCollection_String *self, const char *text) {
-                new (self) NCollection_Utf8String(text);
+                new (self) NCollection_String(text);
             },
             "theCopy"_a)
-        .def("ToCString", &NCollection_Utf8String::ToCString)
+        .def("ToCString", &NCollection_String::ToCString)
-        .def("Size", &NCollection_Utf8String::Size)
+        .def("Size", &NCollection_String::Size)
-        .def("Length", &NCollection_Utf8String::Length)
+        .def("Length", &NCollection_String::Length)
-        .def("IsEmpty", &NCollection_Utf8String::IsEmpty)
+        .def("IsEmpty", &NCollection_String::IsEmpty)
-        .def("__str__", &NCollection_Utf8String::ToCString);
+        .def("__str__", &NCollection_String::ToCString);
 }

src/modules/mod_StdPrs.cpp

@@ -5,11 +5,12 @@
     compound of face outlines, holes already subtracted, which it then extrudes
     into or out of the model.
 
-    These live in TKService, and both classes are real in 7.9.3 rather than the
+    Both are real classes, and the aliasing runs the opposite way from what the
-    deprecated typedefs of Font_BRepFont/Font_BRepTextBuilder that OCCT 8.0
+    8.0 watchlist expected: Font_BRepFont/Font_BRepTextBuilder are the typedefs
-    turns them into — so the 8.0 port is a rename here and nothing else.  The
+    of these, not the reverse.  What 8.0 did move is the package — StdPrs went
-    builder image already asserts TKService links without libGL or libX11, so
+    from TKService to TKV3d, which is why CMakeLists links both.  The builder
-    binding them costs the runtime image nothing.
+    image asserts neither pulls in libGL or libX11, so binding them costs the
+    runtime image nothing.
 
     Rendering a glyph is FreeType work with no Python in it, so it releases the
     GIL.

src/modules/mod_TopTools.cpp

@@ -62,8 +62,13 @@ void register_TopTools(nb::module_ &root) {
              },
              "theKey"_a)
         .def("Contains", &TopTools_IndexedMapOfShape::Contains, "theKey"_a)
-        .def("FindKey", &TopTools_IndexedMapOfShape::FindKey, "theIndex"_a,
+        // OCCT 8.0 added size_t index overloads beside the int ones, so the
-             OCP_RETURN_COPY)
+        // plain member pointer is ambiguous; the int form keeps the negative
+        // index guarded rather than wrapping it into a huge size_t.
+        .def("FindKey",
+             nb::overload_cast<Standard_Integer>(
+                 &TopTools_IndexedMapOfShape::FindKey, nb::const_),
+             "theIndex"_a, OCP_RETURN_COPY)
         .def("FindIndex", &TopTools_IndexedMapOfShape::FindIndex, "theKey"_a)
         .def("Extent", &TopTools_IndexedMapOfShape::Extent)
         .def("Size", &TopTools_IndexedMapOfShape::Size)
@@ -78,7 +83,10 @@ void register_TopTools(nb::module_ &root) {
         .def(nb::init<>())
         .def("Contains", &TopTools_IndexedDataMapOfShapeListOfShape::Contains,
              "theKey"_a)
-        .def("FindKey", &TopTools_IndexedDataMapOfShapeListOfShape::FindKey,
+        .def("FindKey",
+             nb::overload_cast<Standard_Integer>(
+                 &TopTools_IndexedDataMapOfShapeListOfShape::FindKey,
+                 nb::const_),
              "theIndex"_a, OCP_RETURN_COPY)
         .def("FindIndex", &TopTools_IndexedDataMapOfShapeListOfShape::FindIndex,
              "theKey"_a)

tools/parity_venv.sh

@@ -57,8 +57,9 @@ else
 fi
 
 # Everything below runs the venv's interpreter directly.  `uv run` would re-sync
-# the environment against the app's manifest first, which still asks for
+# the environment against the app's manifest first, reinstalling the pinned
-# cadquery-ocp-novtk — reinstalling it on top of the swap.
+# version on top of the swap — which across a kernel bump means silently
+# testing the old OCCT while believing you are on the new one.
 echo "--- sanity ---"
 # Asserted, not printed: __occt_version__ exists only on our wheel, so this is
 # also the check that the swap landed in the environment we meant.