The prime example is the current baselayout ebuild. This ebuild is known to be not package safe, it will not be made package safe, but it is still possible to make a package (emerge -b) even though it will toss several errors if another computer tries to use that package (emerge -k). I would like it if an ebuild cannot be made package safe an "emerge -b" should not be able to make a .tbz2 package to help prevent errors. Reproducible: Always Steps to Reproduce: 1. emerge -b baselayout on portage master 2. emerge -k baselayout on portage slave Actual Results: Since there is a baselayout*.tbz2 created by the emerge -b on an emerge -k, the slave will try to install the binary package which will lead to several errors. Expected Results: On an emerge -b a warning should be tossed that the ebuild is not package safe and no package (.tbz2) should be created. Thus on the slave computers and emerge -k will force an ebuild rather than binary install.
*** This bug has been marked as a duplicate of 25152 ***
How is this considered a duplicate? Both seem to be valid, yet distinct bugs. This one suggests that there should be an option to mark a package as binary unsafe (which is useful; up unil now I didn't know that packages like baselayout were binary unsafe), while bug 25152 states that certain functions are not called when merging a binary package. Please re-open.
Please re-open. Bug was marked duplicate of an unrelated bug.
wrangler
the packages are not 'binary pkg safe' because those functions arent run in other words, there should be no such thing as a 'binary unsafe package' *** This bug has been marked as a duplicate of 25152 ***
