Skip to content

Update llama.cpp to v0.4.0 #31

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
mcharytoniuk merged 11 commits into from
Sep 15, 2026
Merged

Update llama.cpp to v0.4.0 #31

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
934d4cb
Update llama.cpp to v0.4.0
mcharytoniuk Sep 14, 2026
924232a
Make lazy mode errors portable across targets
mcharytoniuk Sep 14, 2026
b58954f
Pass tokenize bitmaps as a const pointer array
mcharytoniuk Sep 14, 2026
6b97335
Group the lazy mode builder with the other mode builders and import i…
mcharytoniuk Sep 14, 2026
1dc303e
Name llama.cpp failures after the library instead of calling them ven…
mcharytoniuk Sep 15, 2026
e5c7e35
Surface C++ exceptions from buffer-backed bitmap loading through a sh…
mcharytoniuk Sep 15, 2026
3dfd1e7
let the tool-call opener win over a reasoning close and treat an abse…
mcharytoniuk Aug 27, 2026
0f35d9b
Redesign streaming markers and optional KV cache APIs
mcharytoniuk Aug 28, 2026
49200f5
Expose the streaming marker API publicly and give role candidates a n…
mcharytoniuk Sep 15, 2026
e0d1158
Cover the role-to-section mapping and duplicate role rejection
mcharytoniuk Sep 15, 2026
93cce06
Preserve section across ambiguous closing markers
mcharytoniuk Sep 15, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions Cargo.lock
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 2 additions & 0 deletions Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,6 +12,7 @@ members = [
"llama-cpp-log-decoder",
"llama-cpp-test-harness",
"llama-cpp-test-harness-macros",
"llama-cpp-wrapper-error-fixture",
"llama-cpp-wrapper-sources",
]

Expand Down Expand Up @@ -42,6 +43,7 @@ llama-cpp-gbnf = { path = "llama-cpp-gbnf", version = "=0.13.0" }
llama-cpp-log-decoder = { path = "llama-cpp-log-decoder", version = "=0.13.0" }
llama-cpp-test-harness = { path = "llama-cpp-test-harness", version = "=0.13.0" }
llama-cpp-test-harness-macros = { path = "llama-cpp-test-harness-macros", version = "=0.13.0" }
llama-cpp-wrapper-error-fixture = { path = "llama-cpp-wrapper-error-fixture" }
llama-cpp-wrapper-sources = { path = "llama-cpp-wrapper-sources", version = "=0.13.0" }
llguidance = "=1.7.0"
log = "=0.4.29"
Expand Down
2 changes: 2 additions & 0 deletions llama-cpp-bindings-sys/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,8 @@ include = [

"/llama.cpp/convert_hf_to_gguf.py",
"/llama.cpp/common/build-info.cpp.in",
"/llama.cpp/ggml/src/ggml-version.h.in",
"/llama.cpp/src/llama-version.h.in",

"/llama.cpp/include/llama.h",
"/llama.cpp/include/llama-cpp.h",
Expand Down
2 changes: 1 addition & 1 deletion llama-cpp-bindings-sys/llama.cpp
Open in desktop
Submodule llama.cpp updated 756 files
2 changes: 1 addition & 1 deletion llama-cpp-bindings-sys/src/lib.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@
clippy::doc_markdown,
clippy::pub_underscore_fields,
clippy::use_self,
reason = "bindgen writes this file from the vendored headers; its shape is not ours to change"
reason = "bindgen writes this file from the llama.cpp headers; its shape is not ours to change"
)]

include!(concat!(env!("OUT_DIR"), "/bindings.rs"));
6 changes: 3 additions & 3 deletions llama-cpp-bindings-sys/wrapper_chat_apply.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,18 +87,18 @@ extern "C" auto llama_rs_apply_chat_template(

return LLAMA_RS_APPLY_CHAT_TEMPLATE_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_APPLY_CHAT_TEMPLATE_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_APPLY_CHAT_TEMPLATE_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & ex) {
*out_error = llama_rs_dup_string(std::string(ex.what()));
if (*out_error == nullptr) {
return LLAMA_RS_APPLY_CHAT_TEMPLATE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_APPLY_CHAT_TEMPLATE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_APPLY_CHAT_TEMPLATE_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
*out_error = llama_rs_dup_string(std::string("unknown c++ exception"));
if (*out_error == nullptr) {
return LLAMA_RS_APPLY_CHAT_TEMPLATE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_APPLY_CHAT_TEMPLATE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_APPLY_CHAT_TEMPLATE_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}
4 changes: 2 additions & 2 deletions llama-cpp-bindings-sys/wrapper_chat_apply.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,8 +19,8 @@ typedef enum llama_rs_apply_chat_template_status {
LLAMA_RS_APPLY_CHAT_TEMPLATE_MODEL_HAS_NO_VOCAB,
LLAMA_RS_APPLY_CHAT_TEMPLATE_TEMPLATE_APPLICATION_FAILED,
LLAMA_RS_APPLY_CHAT_TEMPLATE_ERROR_STRING_ALLOCATION_FAILED,
LLAMA_RS_APPLY_CHAT_TEMPLATE_VENDORED_OUT_OF_MEMORY,
LLAMA_RS_APPLY_CHAT_TEMPLATE_VENDORED_THREW_CXX_EXCEPTION,
LLAMA_RS_APPLY_CHAT_TEMPLATE_LLAMA_CPP_OUT_OF_MEMORY,
LLAMA_RS_APPLY_CHAT_TEMPLATE_LLAMA_CPP_THREW_CXX_EXCEPTION,
} llama_rs_apply_chat_template_status;

llama_rs_apply_chat_template_status llama_rs_apply_chat_template(
Expand Down
52 changes: 26 additions & 26 deletions llama-cpp-bindings-sys/wrapper_chat_parse.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,19 +75,19 @@ extern "C" auto llama_rs_chat_parser_create(

return LLAMA_RS_CHAT_PARSER_CREATE_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_CHAT_PARSER_CREATE_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_CHAT_PARSER_CREATE_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & ex) {
*out_error = llama_rs_dup_string(std::string(ex.what()));
if (*out_error == nullptr) {
return LLAMA_RS_CHAT_PARSER_CREATE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_CHAT_PARSER_CREATE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_CHAT_PARSER_CREATE_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
*out_error = llama_rs_dup_string(std::string("unknown c++ exception"));
if (*out_error == nullptr) {
return LLAMA_RS_CHAT_PARSER_CREATE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_CHAT_PARSER_CREATE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_CHAT_PARSER_CREATE_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand All @@ -101,7 +101,7 @@ extern "C" auto llama_rs_chat_parser_free(
const std::unique_ptr<llama_rs_chat_parser> reclaimed(parser);
return LLAMA_RS_CHAT_PARSER_FREE_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_CHAT_PARSER_FREE_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_CHAT_PARSER_FREE_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
Expand Down Expand Up @@ -171,19 +171,19 @@ extern "C" auto llama_rs_parse_chat_message(

return LLAMA_RS_PARSE_CHAT_MESSAGE_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSE_CHAT_MESSAGE_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSE_CHAT_MESSAGE_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & ex) {
*out_error = llama_rs_dup_string(std::string(ex.what()));
if (*out_error == nullptr) {
return LLAMA_RS_PARSE_CHAT_MESSAGE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_PARSE_CHAT_MESSAGE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSE_CHAT_MESSAGE_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
*out_error = llama_rs_dup_string(std::string("unknown c++ exception"));
if (*out_error == nullptr) {
return LLAMA_RS_PARSE_CHAT_MESSAGE_ERROR_STRING_ALLOCATION_FAILED;
}
return LLAMA_RS_PARSE_CHAT_MESSAGE_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSE_CHAT_MESSAGE_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand All @@ -197,7 +197,7 @@ extern "C" auto llama_rs_parsed_chat_free(
const std::unique_ptr<llama_rs_parsed_chat> reclaimed(handle);
return LLAMA_RS_PARSED_CHAT_FREE_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_FREE_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_FREE_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
Expand Down Expand Up @@ -237,23 +237,23 @@ extern "C" auto llama_rs_parsed_chat_tool_call_count(
*out_count = handle->message.tool_calls.size();
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_COUNT_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand Down Expand Up @@ -285,23 +285,23 @@ extern "C" auto llama_rs_parsed_chat_tool_call_id(
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ID_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand Down Expand Up @@ -333,23 +333,23 @@ extern "C" auto llama_rs_parsed_chat_tool_call_name(
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_NAME_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand Down Expand Up @@ -382,23 +382,23 @@ extern "C" auto llama_rs_parsed_chat_tool_call_arguments(
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_TOOL_CALL_ARGUMENTS_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand Down Expand Up @@ -426,23 +426,23 @@ extern "C" auto llama_rs_parsed_chat_content(
}
return LLAMA_RS_PARSED_CHAT_CONTENT_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_CONTENT_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_CONTENT_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_CONTENT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_CONTENT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_CONTENT_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_CONTENT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_CONTENT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_CONTENT_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}

Expand Down Expand Up @@ -470,22 +470,22 @@ extern "C" auto llama_rs_parsed_chat_reasoning_content(
}
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_OK;
} catch (const std::bad_alloc &) {
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_VENDORED_OUT_OF_MEMORY;
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_LLAMA_CPP_OUT_OF_MEMORY;
} catch (const std::exception & err) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string(err.what());
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_LLAMA_CPP_THREW_CXX_EXCEPTION;
} catch (...) {
if (out_error != nullptr) {
*out_error = llama_rs_dup_string("unknown c++ exception");
if (*out_error == nullptr) {
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_ERROR_STRING_ALLOCATION_FAILED;
}
}
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_VENDORED_THREW_CXX_EXCEPTION;
return LLAMA_RS_PARSED_CHAT_REASONING_CONTENT_LLAMA_CPP_THREW_CXX_EXCEPTION;
}
}
Loading
Loading