unsloth: repin #144 to 586b15ef, fixing a brace dropped in the rebase - #164
Merged
Conversation
Run 33385736675 failed to compile on every backend that builds src/, which is
all of them: 'qualified-id in declaration before (' repeated from
llama-model-loader.cpp:1114 and 'expected } at end of input'.
Cause was mine. Rebasing #144 onto b10709 produced a false conflict where
upstream's lazy_read::add() and this branch's borrow_shared_tensor() landed at
the same offset. I resolved it keep-both, but add() ended at 'return true;' with
its closing brace on the far side of the conflict, so concatenating the sides
left add() unterminated and the shared brace closed borrow_shared_tensor
instead. Every member function after it then parsed as nested.
The brace is restored in the commit that introduced the problem rather than
patched on top. Verified with g++ -fsyntax-only on the merged tree, which is the
check that would have caught this before the run: llama-model-loader.cpp,
llama.cpp, llama-model.cpp and models/qwen4exp.cpp all parse clean.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
Member
|
Run 33385736675 failed across the build matrix because the pinned #144 revision, This repins #144 to its current head, |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What broke
Run 33385736675 failed to compile on every backend that builds
src/, which is all of them: CPU, Vulkan, macOS, ROCm and CUDA-legacy.Resolve tagsucceeded, so the 14-pin chain merges. The failure is a syntax error in the merged source.Cause
Mine. Rebasing #144 onto
b10709produced a false conflict: upstream'slazy_read::add()and this branch'sborrow_shared_tensor()landed at the same offset inllama-model-loader.cpp. I resolved it keep-both, which was right in intent, butadd()ended atreturn true;with its closing brace on the far side of the conflict region. Concatenating the two sides leftadd()unterminated, and the shared trailing brace closedborrow_shared_tensorinstead. Every member function after that parsed as nested, which is exactly whatqualified-id in declaration before '('means.Fix
The brace is restored in the commit that introduced the problem,
a70cf195in the old lineage, rather than patched on top, so no commit in the branch is left unbuildable.Verification
g++ -fsyntax-onlyon the fully merged tree (b10709+ ggml-org#27941 + #144), which is the check that would have caught this before the run rather than after:src/llama-model-loader.cppsrc/llama.cppsrc/llama-model.cppsrc/models/qwen4exp.cppBrace counting alone said 0 delta both before and after the fix on the branch tip, so it is not a sufficient check; the parse is.
Note
A merge that resolves cleanly is not a merge that compiles. The pin preflight only proves the former. Nothing in the pin machinery builds the merged tree, which is why this reached a publish run.