Re: += in ada

[Georg Bauhaus <sb463ba@l1-hrz.uni-duisburg.de>]

Russ <18k11tm001@sneakemail.com> wrote:
: Hyman Rosen <hyrosen@mail.com> wrote in message news:<1066400123.238640@master.nyc.kbcfp.com>...
:>  In C++, I use these operators all the time. I
:> just did a quick search through the source code we've added to
:> out trading system and counted around 500 uses of augmented
:> assignment operators (+=, -=, *=, /=), and yes, we actually do
:> have a few cases of /=. We also have a few of those complicated
:> left-hand sides that people talk about; here's one:
:> 
:>      (*fxMap)[exposure.secCurr.id().get_key()] += exposure.secExp;
: 
: Thanks for that little dose of reality. I hope it wakes up a few
: people around here (but I doubt it will).

That makes 500 uses of mostly +=, -=, and *=, thanks for the data.
It doesn't say, though, how many lines of code were counted.
The examples provided (by Hyman and Lutz) do say something
about the uses.
As to reality, I am sure every one in this discussion has seen
+=. But I'd guess that they are still waiting for a detailed
argument about
- whether +:= can be done in Ada without introducing all sorts of
  little problems (as seems to be the case in Python, as someone
  has mentioned)
- whether implementing +:= is a hard problem. A detailed argument
  would, I guess, contain an estimate of the effort involved introducing
  +:= in Ada (not in some other language. SETL does have +:=, and SETL
  seems to have played a role during development of GNAT, and even 
  without that, I'd not assume that Ada compiler writers are out
  of touch with reality because they don't see the big win of
  +:= ?)

Have you answered the question about what should happen if during
assignment something goes wrong?  Do I have to make a backup
copy, then? Have I missed something?



Georg