-
-
Notifications
You must be signed in to change notification settings - Fork 142
crash caused by CPM? #315
Copy link
Copy link
Closed
Labels
CMakecrash(-like)Bug that results in severe state corruption, lock-ups with recovery, or outright terminationBug that results in severe state corruption, lock-ups with recovery, or outright termination
Description
Metadata
Metadata
Assignees
Labels
CMakecrash(-like)Bug that results in severe state corruption, lock-ups with recovery, or outright terminationBug that results in severe state corruption, lock-ups with recovery, or outright termination
As noted in #313, at least on my machine (Ubuntu 21.04) I can reliably reproduce the SEGV in
YAML::LoadFile()call just by the order of theCPMAddPackage()calls inThirdParties.cmake.What is definitely suspicios is the following:
Have a look at the line
-- CPM: libunicode: adding package yaml_cpp ....The other CPM lines look different, i.e. they don't mention two different package names.
It might very well be a bug from CPM.cmake? At least the fact that the order changes the behavior does not sound quite right. /cc @data-man (this is a tracker ticket).