[Development] Qt 4.x and Qt 5 frameworks should use @rpath (QTBUG-31814)

Adam Strzelecki ono at java.pl
Wed Aug 13 13:05:08 CEST 2014


> Afaics the suggestion is to never have the absolute path to the Qt frameworks added to the rpath at all. Which would make the rpaths *only* be useful for the bundled dependencies. So, when not using the “bundle stuff” option, it would require setting either DYLD_.... in the run environment, or explicitly adding the absolute Qt path to your binaries’ rpath, to be able to run the 
> result. I still do not see why the absolute path shouldn’t be added when _not_ bundling Qt though.

From very beginning a was about to add absolute rpath to Qt libraries in built executable, with one exception - when bundling is done on build, then we can add relative path. I was not about to touch existing workflow. Now we call it "Phase I" and changes are ready.

If bundling is done as separate manually triggered step, we need to replace absolute rpath to relative in the executables, also to not keep any leftovers.

From very beginning the reason I started this discussion was to never touch Qt libraries binaries once they are built (as now its done on install and macdeployqt). That's it. All other workflow changes was somehow ideas of the others not mine. Yet because I was asked how do I see "Phase II" (aka workflow changes) I stated my opinion.

Since both of us don't like idea of using DYLD_ and tweaking Qt Creator for that, absolute rpath is IMHO necessary. Therefore I may ask you for a favor to look at following changes:

	https://codereview.qt-project.org/#/c/92072/9//ALL,unified

Because I am constantly loosing Jake in this and frankly I am getting confused what to do.

--Adam


More information about the Development mailing list