On Mon, Feb 10, 2014 at 2:34 PM, Reid Kleckner <[email protected]> wrote: > On Mon, Feb 10, 2014 at 11:28 AM, Aaron Ballman <[email protected]> > wrote: >> >> (Removing llvm-admin since this doesn't concern them.) >> >> On Mon, Feb 10, 2014 at 2:21 PM, Reid Kleckner <[email protected]> wrote: >> > I'm conflicted about this. The less time I spend editing .td files, the >> > happier I am. Vim has reasonable support for highlighting and editting >> > rst >> > out of the box, and I'd rather edit full text in that directly. >> > >> > We're pretty good at keeping LLVM's LangRef.rst up to date through code >> > review. Could we use the same approach with something in >> > docs/AttributesReference.rst, or do people think it would quickly become >> > stale? >> >> We could, but I believe this approach is an improvement. For starters, >> it's a way that we can *require* attributes to have some documentation >> which is a drastic improvement over the LangRef.rst approach (at least >> as far as users are concerned). Also, there are things we can >> automatically generate for attributes which would require duplicated >> effort if it was done manually (such as what syntaxes are supported, >> what subjects the attribute can appertain to, etc). The declarative >> nature of the attribute tablegen lends itself nicely to this sort of >> automation. >> >> Honestly, I think the benefits are significant enough to warrant this >> approach. > > > I'm OK with repeating the complete set of attribute names once in the > documentation, but doing a nice pretty printing of all the spellings seems > like a nice win that I hadn't thought of.
The other big win, as far as I am concerned, is that you cannot add an attribute without documenting it in some form or fashion (I suspect that new attributes getting an Undocumented designator will not pass review very often). ~Aaron _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
