On Wednesday, 3 December 2014 at 13:34:42 UTC, Jacob Carlborg
wrote:
On 2014-12-02 23:00, H. S. Teoh via Digitalmars-d wrote:
I'm finding it harder and harder to accept Walter's stance
that symbol
lookups should be kept simple and free from complications and
convoluted
corner cases, etc.. Except that it is *already* full of
convoluted
pitfalls and corner cases you must avoid, as illustrated
above. I wish
we would just acknowledge that the current symbol lookup /
import rules
(or at least the implementation thereof) are inadequate, and
find a
clean solution to this long-standing issue, instead of hoping
that
denying the problem will somehow, magically, make it all go
away.
I fully agree.
+1
So far it seems there is huge amount of anecdotal evidence that D
module system does not work as designed and people keep inventing
hacks and workaround to get over it. And yet Walter still
considers it success.
Well, compared to C includes - no doubt. Compared to actually
usable modle system.. meh. Right now it pretty much requires
relying on coding discipline and custom rules for import
formatting.
