Open
Conversation
1a9b847 to
e08f401
Compare
e08f401 to
f30b412
Compare
MSVC does not honour friend-class access when the accessing code is in an inline constructor body defined inside the friend class's own class definition. Move Tensor::Tensor(Buffer) out-of-line so that both Buffer and Tensor are fully defined at the point of access. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Blank context lines in unified diff must carry a leading space; a bare newline is treated as end-of-patch by git apply, causing the "corrupt patch" error. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MSVC does not honour the 'friend class Tensor' declaration in Buffer, so moving the definition out-of-line (previous approach) had no effect. Instead, add a public inline accessor GetBuilder() to Buffer and call it from Tensor's constructor, eliminating the need for friend-class access entirely. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
vcpkg overrides find_package with a macro that stores \${ARGN} in a
double-quoted string. Textual macro substitution causes the value to
be re-parsed as a string literal, so a Windows path whose components
start with a letter (e.g. \build_bot -> \b) triggers CMake's CMP0010
invalid-escape-sequence error. Normalizing with file(TO_CMAKE_PATH)
converts backslashes to forward slashes before the path reaches the
macro.
No upstream vcpkg issue exists for this as of 2026-02-25.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
With vcpkg support in place for the apps, only two stand out as needing additional treatment: hannk and onnx. These both use FetchContent, which fundamentally confuses the build system with dependency provision. This PR removes FetchContent from hannk, replacing it with structured vcpkg ports that benefit from binary caching (we used to build tensorflow-lite on every PR!), SHA512 validated sources, and faster HTTPS based downloads (rather than a shallow git clone). It also upgrades tensorflow-lite to 2.21.0-rc2, which is the first (and latest) version to have a usable CMake package API.
This will likely require making the buildbots unconditionally use vcpkg when building the apps.