| Age | Commit message (Collapse) | Author |
|---|
|
|
|
This ensures that the Python source files are only generated and copied
into the output directory if inputs have changed, rather than being done
unconditionally.
|
|
This ensures that CMake detects when files that match the glob are added
or removed.
|
|
|
|
|
|
Fixes https://github.com/Vector35/binaryninja-api/issues/7194
|
|
Forgot to update these after the guid regression fix, there is no cause for alarm :)
|
|
Previously we only added the warp function activities to the meta workflow, however objective-c plugin currently registers its own
workflow that we must add the activities to as well.
Fixes https://github.com/Vector35/binaryninja-api/issues/7172
|
|
|
|
Fixes https://github.com/Vector35/binaryninja-api/issues/7116
|
|
matched functions
One day I will get the QT incantations correct
|
|
guids generated
Just a sanity check so we can alert the user, they can just reanalyze their functions and the guid will be constructed
|
|
We emit jumps at the end of some blocks that the architecture is not aware of, they are not actually apart of the functions instructions
rather they are used in our analysis... somehow, I have not figured out what they are supposed to be used for but regardless, they should be
ignored by WARP.
This was just a case where the filtering was including jumps in non start instructions, where we actually wanted the opposite.
|
|
This removes more barriers for those working with firmware to use WARP, previously if you had defined any sections (such as a vector interrupt table) then backed segments would
end up being marked non-relocatable. This, combined with the use of ELF intermediate object files created inconsistent GUID's, using the segments backing the real functions fixes this.
|
|
|
|
It is apparently "too bright" and I need to get "my eyes checked"
|
|
|
|
wants the symbol
We were previously constructing the function type, comments and variables to just throw away, wasting precious cpu cycles!
|
|
Will also adjust the worker thread count so that there is enough for view destruction and what not.
The default will be filled as well for the user.
|
|
Just added noise to the logs, probably should do this for rtti as well
|
|
|
|
|
|
- Fix project file filter not blacklisting warp files
- Fix project name not being used for default file save name
- Prevent adding the same file to the available sources
|
|
This previously was not allowed as the information would sometimes include a zero based segment, however this is still preferable in the case of firmware, where sections will not be available by default. This improves the usability of WARP with firmware by removing the requirement of filling section information prior to function analysis.
|
|
This is going to be disabled by default on this upcoming stable, however users may enable it once we deploy the public server. The data from the server is done through `Container::fetch_functions` independent of the nonblocking function lookup functions.
The sidebar has been updated to drive fetching so that when users navigate to a new function the fetcher will kick off. This fetcher operates on a separate thread, in the event of a user navigating to many functions before the current fetch has completed they all will be batched together in a single fetch.
Networked container currently is limited to just function prototypes, other type information separate from the function object will be omitted.
|
|
|
|
|
|
|
|
Was previously not exposed, we want it for unit tests.
|
|
With the new instruction retrieval we are getting the instructions at the end of the lifted function which are not really apart of the function, but they share the address of the last ret/jump.
|
|
|
|
|
|
appended for a given instruction
|
|
|
|
Another set of bugs caused by undo and main thread actions, when will it end
|
|
Was not able to find the core plugin in the plugins directory
|
|
Removing the nesting because we now scope view plugin commands out of the context menu
|
|
|
|
- Added FFI
- Added a sidebar to the UI
- Added project, directory and archive processing
- Added generic `Container` interface for extensible stores of WARP data
- Fixed type references being constructed and pulled incorrectly
- Added HTML, Markdown and JSON report generation
- Made the WARP information added as an analysis activity
- Flattened the signatures directory, the target information is stored in the file now
- Matched function information is stored as function metadata in the database to reliably persist, alongside the function GUID
- Split the matching out from the application, allowing you to match on a given function without applying it
- Added more/better tests
- Added support for binaries with multiple architectures, the functions are now also queried based off the Target, see WARP spec for more details
- Greatly improved support for RISC architectures, see WARP spec for more details
- Greatly improved UX when loading files after the fact, will now sanely rerun the matcher
- Omitted the function type if not a user type, this greatly reduces file size
- Improved support for functions that reference a page aligned base pointer, see WARP spec for more details
- Removed some extra cache structures that were causing erroneous behavior
- Fixed edge-case in LLIL traversal missing some constant pointers, this was a bug in the Rust bindings
- Added support for function comments
- Made long running tasks, such as generating, matching and loading signatures, cancellable where possible
- Made function constraints more versatile, allowing for easy extensions in the future, see WARP spec for details
- Added options to signature generation, such as what data to store, and whether to compress the data or not
- Made all long running tasks prompt the user for required information before the task starts, allowing users to "set it and forget it" and not have to baby sit the finalization of the task
- Myriad of other changes to the actual WARP format that impact performance, file size and general feature set, see https://github.com/Vector35/warp for more details
|
|
Highlights all variant instructions and blacklisted instructions
Variant instructions are highlighted red, blacklisted will be orange.
|
|
Need to do this for exposing to FFI and things like render layers
|
|
Also update minimum CMake version.
|
|
|
|
We don't do enough with the lifted il != non lifted il to justify the bound.
This makes modifying IL much less work as the historical lifted il bound is gone.
|
|
This is being done to reduce complexity in function signatures, specifically many of the strings we are passing ultimately should be new types themselves instead of "just strings", things such as type ids.
Another place which was confusing was dealing with filesystem related APIs, this commit turns most of those params into a stricter `Path` type.
This is bringing the rust api more inline with both python and C++, where the wrapper eagerly converts the string into the languages standard string type.
Special consideration must be made for symbols or other possible non utf-8 objects.
This commit will be followed up with one that adds the `IntoCStr` bound on API's we want to keep as invalid utf-8 so we can for example, get section by name on a section with invalid utf-8.
|
|
- Removed `to_string` shortcut from `BnString`.
- Misc formatting
|
|
conversion can be tolerated
Still need to go and audit all usage, but realistically the most important places to give the user control are with symbols, where the data can come from non utf8 sources
This is still incomplete, I just looked for usage of -> BnString so any other variant was omitted.
|
|
|
|
IDK why this was not done prior, leaking the returned core activity and workflow object.
|
|
|
