Glenn Smith
Glenn Smith
Couple updates on this from a few weeks ago: 1. You can now specify custom type parser arguments with the new Platform::AdjustTypeParserInput callback, which you can use to give clang...
Tried with a PDB a little while ago and they seem to respect the base address and work as expected
As well, we may want to look into a "collapse sub-structures" option, since most STL types are unholy hydras of nested zero-size structures and unions.
Can confirm. Looks like there is deadlock creating the Undo Actions for the Clear action. Details: The UI is in a worker thread, waiting on the main thread to commit...
This is fixed in >= 4.1.5394 by having the analysis completion event spawn a main thread task to clear the type, instead of doing it directly in the event.
This is just a consequence of the `analysis.types.paddingThreshold` setting being rather low by default. Any types >= that setting are rendered in a single line for performance reasons (previous users...
So there are two different issues here: 1. Parsing types takes a while (especially if it's invalid) 2. The UI waits on the background thread I can add a setting...
Build 4.1.5071-dev will have two changes to address this: 1. You can disable the UI background type checking via the new Setting `ui.types.checkParse` 2. Parts of the parsing setup process...
These are now implemented as of >= 4.1.5339-dev. You can create them using `void* __based(start)` and various other forms as documented in the new type annotations docs.
Sure. See https://dev-docs.binary.ninja/guide/types/attributes.html#based-pointers