Hi Mike, On Mon, Jul 22, 2019 at 06:37:35PM -0400, Michael Meissner wrote: > On Mon, Jul 22, 2019 at 03:56:26PM -0500, Segher Boessenkool wrote: > > That still needs an explanation: why is this a good thing, why do you > > want that change? Sometimes that is obvious of course, but here it is > > not. It would be a lot more obvious if there was more context. > > The trouble is to get that much context really relies on about several > additional patches to get to the functions in particular that should be split > out.
In the normal workflow, you send a series of patches when it is ready. And if a series is not ready, you do not send it. You can also give context just in the emails, or send a work-in-progress patch just as FYI (please mark it clearly as such, then). But without context, how can I see if what a patch does is correct and useful? > As I implement stuff, I find myself neeting/wanting to access the stuff in > reg_addr in other files (predicates.md, and the new file rs6000-prefix.c). Then you first need to question if things are split up in a useful way the way things are, and then how it could be done better. > But it would be nice to have that information available to the other .c files > as well as the .md files. If many other modules need the data, then either: a) The module boundaries are not set nicely; or b) This is a central data structure, and should be treated as one. Segher
