OSX users out there? Serious bug (I think)

Denis Shelomovskij verylonglogin.reg at gmail.com
Wed Apr 3 09:17:45 PDT 2013


02.04.2013 18:37, Nick Sabalausky пишет:
> On Tue, 02 Apr 2013 11:20:26 +0400
> Denis Shelomovskij <verylonglogin.reg at gmail.com> wrote:
>>
>> P.S.
>> Also, Nick, if you have time, could you show me, how do you suggest
>> "to use D reliably" in non toy Windows projects as you "never said
>> anything" against it?
>>
>
> Don't do incremental compilation. Stick to the releases instead of
> Git master, and if you do use Git master, make sure to 1. make clean
> everything (dmd, phobos, druntime) before recompiling and 2. make sure
> all three, dmd, phobos, druntime always match - don't update one
> without updating the others.
>
> Aside from that stuff though, I haven't personally run into the problems
> you describe, so I really don't know.
>

You are talking about different thing here. I mean known old bugs which 
may break every program's linking because of every small code change as 
I described and it isn't connected with any dmd/phobos/druntime version.

A also noted that it was triggered on both my projects to highlight the 
failures are not seldom (and Andrei though I was playing the victim 
because of that) and I don't understand how they can be ignored for such 
a long time as I don't see any other blockers in the language. The only 
other bad thing is a high probability of wrong-code when using contracts 
or closures but it is not a blocker and even not a PITA, one just have 
to know about it and put (lots of) workarounds which doesn't spend a lot 
of time (yes, uglier code, but D is still wins over e.g. C++ a lot here).

-- 
Денис В. Шеломовский
Denis V. Shelomovskij


More information about the Digitalmars-d mailing list