Insight into loading CMake files and building of symbols?

I just learned that another team at my job has a project of a similar size to me and their reload CMake using all four configuration types takes 12 seconds to run, while my project takes nearly two minutes when running all four configuration types.

I recently learned the trick to change the configuration types (Thanks Anna Filipova!) so that helps.

But even so, it seems if there is a fundamental problem with my project. We are sharing most of the CMake include files and the project only differs pretty much at the top level CMakeLists file. We also share most of the libraries.

Is there a way to examine the complexity of the built indexes or cmake behavior?

2 comments
Comment actions Permalink

Just wanted to give this a bump - this is such a huge problem with CLion right now.  I love using the tool but I dread making any changes to my CMakeLists.txt because it takes 3-4 minutes to reload (even after the most trivial of changes).  Changing to one configuration type doesn't really seem to help.

Can you give us any hope that this will be fixed in a meaningful way?  It seems like a fundamental flaw in the data model so I'm concerned that its always going to be a problem.

1
Comment actions Permalink

Wanted to add an update that this problem is tremendously improved in 2016.3.  I'm down to a minute when switching between branches that have significant changes in the CMake file and more importantly the time for adding a new file is down to only a few seconds.  Thanks to the CLion team for addressing this issue!

0

Please sign in to leave a comment.