The attached patch adds two variables to the kernel-2 eclass to extend functionality: K_EXTRAEINFO_F and K_EXTRAEWARN_F. The _F stands for format which allows users to embed \n and \t in the K_EXTRAE* text strings. Later, during the postinst_sources function, echo -e is used in conjuction with fmt -w75 -s in order to print nicely formatted output. I did this because I wanted a line break in the middle of a string I had, but fmt kept wrapping all text together. This enhancement allows the user to optionally choose to use the _F and format his/her K_EXTRA* string. The patch does not modify any existing code so current functionality is maintained.
Created attachment 88690 [details, diff] kernel-2.eclass.diff patch to add capability to print formatted K_EXTRAEINFO/WARN output.
A new enhancement has been added which allows the user to set a flag UNIPATCH_GENPATCHES_LAST which will permit the application of custom patchsets (a la ck, beyond) PRIOR to the application of genpatches. This will remove many patch errors that result when genpatches are applied first.
Created attachment 88829 [details, diff] kernel-2.eclass.diff adds UNIPATCH_GENPATCHES_LAST flag and application in unipatch function.
That limits what ebuilds can do somewhat. I think it would be better to leave existing behaviour as-is, but add another list variable which contains patches to be applied after genpatches. That way an ebuild can specify patches to be applied both before and after.
Comment on attachment 88829 [details, diff] kernel-2.eclass.diff patch order patch moved to 136271 per dsd.
John, per your request, this is the enhancement bug. If you're on the kernel@g.o email, then just remove the duplicate cc. In regards to: "however on immediate thought I can't see why formatting this can't be applied to the main variables." All text just wraps together and split lines are joined. fmt does this. Using echo -e and making some small changes to fmt fixes this (unless I missed something obvious!)
upstream inaction.
