On 16 Jan 2016, at 20:32, Stephen Leake <[email protected]> wrote: > > Stephen Leake <[email protected]> writes: > >> Simon Wright <[email protected]> writes: >> >>> My recent changes are in monotone at org.emacs.ada-mode.sjw at >>> 279a675319535670119e771355becbf6170cb22d - this includes the gpr-mode >>> changes I wrote about earlier. >> >> Right, I'll work on merging those. > > This is now done, and all tests are passing. > > You mentioned makefile changes; I didn't see any in the merge.
In build/common.make at :52, I use '-' to ignore failure (which happens if the files are different, and halts the test). In build/wisi/Makefile at :65, I check for Darwin and treat it the same as Linux. I'd have preferred to check for Windows first and assume Unix otherwise, but not sure of a "universal" test for Windows. Thes change are both in the latest org.emacs.ada-mode, so I guess they crept under the radar! > You mentioned problems running the tests; what OS are you on, and what > problems are you having? The above change to build/wisi/Makefile fixed them. > Did you get anywhere with imenu? It looks to me that the grammar-based approach will be very slow; it'd need to traverse the buffer several times. Perhaps I only need it for subprograms; task & protected objects/types aren't so difficult. The problem arises (in the current regex-based approach) because of the need to distinguish a subprogram spec from a subprogram body; the 'is' tells you, but you wouldn't want to charge past the end of a spec to find the 'is' from a subsequent body. _______________________________________________ Emacs-ada-mode mailing list [email protected] http://host114.hostmonster.com/mailman/listinfo/emacs-ada-mode_stephe-leake.org
