bearophile wrote:
This post shows that if you want to write (even simple) numerical code that
uses integral numbers, you need to use everywhere efficient multi-precision
integers, or at least you need integer overflows at runtime. Otherwise you are
programming in the darkness.
Although this
Don:
> Although this type of code is extremely common in programming puzzles, I
> seriously doubt that it's used much anywhere else.
But integer-related bugs infests many other kinds of programs too :-)
Good luck with your bugs too Don,
bearophile
Hi
I like reading the bugs list.
I see lots of patches (for example from Rainer Schuetze) but I don't see all of
them applied (in the latest release).
How does that work?
Regards
Alex
Steven Schveighoffer wrote:
I expect a new dmd will be released in the near future to fix this
issue. My apologies to those who have wasted time trying to debug this...
-Steve
I confirm that with the patch in dmd 2.0.43 my crash problem is gone.
Thanks for the fix and the test case!
j.
Hello Alex,
Hi
I like reading the bugs list.
I see lots of patches (for example from Rainer Schuetze) but I don't
see all of them applied (in the latest release).
How does that work?
While the source for DMD is available (and many parts open source), only
very few people have the "keys to
Hello bearophile,
Don:
Although this type of code is extremely common in programming
puzzles, I seriously doubt that it's used much anywhere else.
But integer-related bugs infests many other kinds of programs too :-)
IIRC FP overflows can trigger hardware based interrupts where as I don't
On 09/04/10 14:08, Alex Strickland wrote:
Hi
I like reading the bugs list.
I see lots of patches (for example from Rainer Schuetze) but I don't see
all of them applied (in the latest release).
How does that work?
Regards
Alex
Quite a few of the 68 patches are planned for inclusion into DMD,
Hello Robert,
Hopefully when he gets some
time again he'll start filtering through the patches and commenting on
them/applying them.
Has anyone tried to reveres engineer a coding standard from Walter's comments/etc.
on patches?
If we can talk him into committing between each patch he apples
BCS:
> If we can talk him into committing between each patch he apples
Nowdays this is mostly true, with few exceptions like the 4004, 4005, 4019,
4020, 4027, 4029 bunch, that was a pack of related changes.
Bye,
bearophile
Michael Rynn Wrote:
>
> In playing around trying to understand and make a better AA, I have
> updated again the dsource/projects/aa/trunk/druntime/aaA.d, this time to
> be a single linked list version.
>
> In the template versions of various implementations, and in Java HashMap,
> I noticed
In playing around trying to understand and make a better AA, I have
updated again the dsource/projects/aa/trunk/druntime/aaA.d, this time to
be a single linked list version.
In the template versions of various implementations, and in Java HashMap,
I noticed that its nice to be able to specif
Michael Rynn, I don't agree. In my opinion it's better to keep the built-in AAs
very flexible and very easy to use. I prefer them to have methods/attributes to
make them more handy to use in real situations (time ago I have listed here
time several missing things, like clear and copy methods, wh
I have tried two of those programs with the latest dmd2 release. The
performance of the euler174 is the same. While the running time of the euler14
goes from (with a fully unloaded PC) 5.49 seconds with v2.042, to 4.13 seconds
with v.2.043, probably thanks to AAs improvements.
Bye,
bearophile
Michael Rynn wrote:
In playing around trying to understand and make a better AA, I have
updated again the dsource/projects/aa/trunk/druntime/aaA.d, this time to
be a single linked list version.
I don't understand, the latest release already changed aaA.d to use a singly
linked list.
On 09/04/2010 19:46, Michael Rynn wrote:
In playing around trying to understand and make a better AA, I have
updated again the dsource/projects/aa/trunk/druntime/aaA.d, this time to
be a single linked list version.
Well,Well Since D2/Phobos is still(quite a while now) lacking a
container lib
15 matches
Mail list logo
