Patch by Pino Toscano. Reported in http://bugs.debian.org/835665
Details
Diff Detail
Event Timeline
cmake/LLDBDependencies.cmake | ||
---|---|---|
168 | This really should turn into a check so that we don't need this to be extended for every single OS that gets added. |
cmake/LLDBDependencies.cmake | ||
---|---|---|
168 | There is now support for CMake >= 3.0, which offers a builtin check for it. |
- in the future can you please upload with context (e.g. git diff -U9999)
- I think the OS names in utilsOsType.py are kept sorted and the actual enum values are not part of an ABI
cmake/LLDBDependencies.cmake | ||
---|---|---|
168 | I agree with previous comments - if we can convert this to a CMake built-in check, that would be ideal. |
@sylvestre.ledru, I think this will be ready to go if you can convert that check to a CMake built-in check. Maybe you can pass that along to Pino Toscano?
cmake/modules/LLDBConfig.cmake | ||
---|---|---|
414 ↗ | (On Diff #72758) | Hi Sylvestre! It's hard to tell without more context, but it looks like this location has most/all configurations going through it. For OSes that don't actually have a backtrace package, I think this emits a cmake error, doesn't it? This might need to be protected by the systems that need the backtrace package. (Probably Unix-like systems only?) |
cmake/modules/LLDBConfig.cmake | ||
---|---|---|
414 ↗ | (On Diff #72758) | It's sufficient to drop the REQUIRED keyword and it will be fine. ${Backtrace_LIBRARY} will be evaluated to NIL in case of lack of this dependency. I'm unsure whether find_pacakge() might be after usage of ${Backtrace_LIBRARY}. I would reorder it at least for clarity for a reader. |
This really should turn into a check so that we don't need this to be extended for every single OS that gets added.