|Re: [cdt-dev] Persisting discovered scanner configuration
> > Sean, > you are absolutely right. '.cdtproject' is the file where the current user > defined scanner configuration is stored. If it remains where it is it will > impact the whole team if any developer tries to customize it. On the > other hand, if is stored elsewhere then developers will not be able to > share common configuration elements. Maybe we need something like > Eclipse's Exporting/Importing preferences to share common CDT > configuration. > Agreed. The irony is, we started by using the project/resource properties or metadata to save settings and other information. But quickly because of customers/clients demands, we move to make this "sharable" as much as possible. To illustrate, we even have outstanding PRs about the possibility to save/share the breakpoints, sharing the expressions in the Expression view, etc ... There is a real value added to export settings. Recent bugzilla activity shows a suite of PR asking information not to be save(in the project) because it marks the project as "dirty" when using the CVS decorators .... but it does not seem to equate the feedbacks we have so far from our users. I suppose that having the best of both world is the way to go. Include paths, macros, .. are usually save in makefiles or some other persistent storage and that a CVS diff shows things changed was never a serious problem. But because of the "dynamic nature" of the scanner configuration it could be a good thing to have the behaviour configurable.
Back to the top