| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Fixes #942
Fixes #947
|
|
|
|
Presumably they are blacklisted because we can't understand them properly, so be
pessimistic about what we can derive for it.
Fixes #944.
|
|
ir: When something has a definition, return unresolved type references until we parse it.
This fixes #888
|
|
|
|
|
|
|
|
we parse it.
This ensures that we see all the relevant types that are defined when parsing
the definition, avoiding problems like #888.
Fixes #888
|
|
Delete automatic code formatting guidelines
These are currently inaccurate as `cargo fmt` is not enforced for bindgen.
This section can be re-instated if and when code formatting is enforced.
|
|
These are currently inaccurate as `cargo fmt` is not enforced for bindgen.
This section can be re-instated if and when code formatting is enforced.
|
|
Time phases
New module `time_phases` with an RAII timer which prints to stderr, and command-line flag `--time-phases` to enable it.
Fixes #933
|
|
initial timer branch commit
added time_phases builder parameter
create timers for BindgenContex analyses, codegen
time parse
added docstrings
added fitzgens suggestions
add --time-phases to bindgen::Builder::command_line_flags
|
|
Avoid generating fields/methods for opaque objects
WIP fix for issue #929
With libclang 4, passes all tests on my machine except `dump_preprocessed_input`, but that fails on the master branch as well
|
|
|
|
|
|
Bump version to 0.30.0
Version bump, primarily to get #832 and #892 in now that 1.19 is out and untagged unions are usable in stable!
r? @emilio
|
|
|
|
Rewrite `has destructor` analysis as a fixed-point analysis
Fixes #927 . Note that this creates a dependency between the "cannot derive copy" and "has destructor" analysis, i.e. the "has destructor" analysis must run before the "cannot derive copy" analysis, because "cannot derive copy" needs the results of "has destructor".
|
|
monotone framework
|
|
Simplify helpers::blob
Thanks @fitzgen for the detailed instructions.
Fixes #928.
|
|
Thanks @fitzgen for the detailed instructions.
Fixes #928.
|
|
Can derive Eq
Fix: #880 r? @fitzgen
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Rename `TypeKind::Named` to `TypeKind::TypeParam`
Also renames a bunch of other things referring to named types to refer to type parameters.
Fixes #915
|
|
Also renames a bunch of other things referring to named types to refer to type
parameters.
Fixes #915
|
|
Update doc for unions
Addresses #832
I also modified the `--help` text to print the default Rust target.
|
|
Addresses #832
|
|
Recursively include src files, since Cargo pattern-matching changed.
|
|
|
|
Fix misleading comments
r? @fitzgen or @emilio
|
|
|
|
Check if `clang_Type_getNumTemplateArguments` is loaded before use
Older clang don't have it, and while we can't pass our whole test suite with those older clangs, we can still generate simple C bindings, so it makes sense not to panic for C++ only things.
|
|
Older clang don't have it, and while we can't pass our whole test suite with
those older clangs, we can still generate simple C bindings, so it makes sense
not to panic for C++ only things.
|
|
Only compute which types we can derive PartialEq for if we'll use it
If we aren't going to derive `PartialEq`, then it doesn't make sense to even run the analysis.
r? @photoszzt or @emilio
|
|
If we aren't going to derive `PartialEq`, then it doesn't make sense to even run
the analysis.
|
|
Only run analyses when we are going to use their results
Currently, there are various analyses related to deriving various traits that we unconditionally run. However, if we aren't going to derive those traits in codegen, then computing whether or not we can derive the traits is wasteful.
r? @emilio
|
|
Currently, there are various analyses related to deriving various traits that we
unconditionally run. However, if we aren't going to derive those traits in
codegen, then computing whether or not we can derive the traits is wasteful.
|
|
Derive PartialEq Analysis
Fix: #878 r? @fitzgen
|
