... I'm not convinced: 1. compile-time assertions will be primarily for the advanced user and library programmer. The library's clients will get the benefit of...
1081
Andy Glew
glew@...
Dec 1, 1999 11:57 am
... So, there's a whole slew of possible reference counted and garbage collected pointers, with implementations ranging from + various reference count and GC +...
1082
Bill Wade
bill.wade@...
Dec 1, 1999 3:07 pm
... A trick to consider (I've never actually used this) is to require a mutex& as an unused argument in all constructors. So mutex(); // Compiler error ...
1083
scleary@...
Dec 1, 1999 3:17 pm
... member ... of ... change ... I agree with allowing empty_member to be able to hold function pointers (or any other non-class type). I also agree with...
1084
Greg Colvin
gcolvin@...
Dec 1, 1999 4:27 pm
From: Andy Glew <glew@...> ... I've been backing into this one for years. One of my committee proposals had counted_ptr parameterized on an allocator,...
1085
Andy Glew
glew@...
Dec 1, 1999 6:26 pm
... Can we enumerate the different semantics? I may have missed part of cyclic_ptr<T>. Repeating part of my last email, I see two main classes of semantics: ...
1086
Darin Adler
darin@...
Dec 1, 1999 7:16 pm
... It seems pretty easy. Here's an example: #include <iostream> template<typename T> struct is_pointer_trait { enum { value = false }; }; template<typename T>...
1087
scleary@...
Dec 1, 1999 7:59 pm
... OK, I'm convinced; let's use Howard's traits approach. -Steve...
1088
Greg Colvin
gcolvin@...
Dec 2, 1999 5:08 am
From: Andy Glew <glew@...> ... I guess it depends on how minor GC is, but yes, it's just like shared pointer otherwise. ... The point is that once you...
1089
jsiek@...
Dec 2, 1999 8:49 am
This is not a proposal for a library. It is a proposal for an interface (or set of "concepts" in the SGI STL meaning of the term). This is the interface for...
1090
John Maddock
John_Maddock@...
Dec 2, 1999 12:53 pm
Howard - ... getting decremented?). If this is the loop body and you've edited out the loop, then the by-value looks like a very big win to me for this ...
1091
Gavin Collings
gcollings@...
Dec 2, 1999 12:56 pm
... I think Andy's point is valid. I am part way through implementing a smart pointer (interface defined by boost::shared_ptr) parameterized by sharing policy...
1092
scleary@...
Dec 2, 1999 1:16 pm
... I second this motion. I've looked around for a good generic graph library. The GGCL is much more advanced than the GTL (Graph Template Library), and if we...
1093
Andy Sawyer
boost@...
Dec 2, 1999 1:42 pm
... As has been pointed out on more than one occasion, the most common use of allocators [in the standard library, at least] is as hooks into the type system...
1094
Andy Sawyer
boost@...
Dec 2, 1999 4:55 pm
On 02 December 1999 12:55, John Maddock wrote: AMS> Of course, it's possible to generate (or at least write by hand) _much_ ... JM> Exactly, I really wish that...
1095
Joseph Gottman
joegottman@...
Dec 3, 1999 3:09 am
Dear Sirs, Are any of the heaps in your heap library stable, in the sense that if two equivalent objects are put on the heap it is guaranteed that the first...
1096
Valentin Bonnard
Bonnard.V@...
Dec 3, 1999 3:10 am
... Hum... I don't see how it can be parsed... -- Valentin Bonnard...
1097
Dave Abrahams
abrahams@...
Dec 3, 1999 9:21 am
... Wow, I've just been struggling with this very problem at work (all night). I'd like to work on this sometime. Heck, I'd like to read your whole email ...
1098
Moore, Paul
Paul.Moore@...
Dec 3, 1999 2:31 pm
Hi, I was amazed to find that there isn't a basic rational numbers class available anywhere on the net. It strikes me as pretty fundamental, like complex. ...
1099
Dietmar Kuehl
dietmar_kuehl@...
Dec 3, 1999 4:23 pm
Hi, ... You have to abstract quite a lot of stuff out of the algorithms. It is necessary to keep them general. Actually, I'm not sure whether GGCL already...
1100
Dietmar Kuehl
dietmar_kuehl@...
Dec 3, 1999 4:23 pm
Hi, ... No, none of the heaps in the heaps library is stable. The reason for this are the performance characteristics: You cannot maintain the performance of...
1101
Andy Glew
glew@...
Dec 3, 1999 4:52 pm
???????? Creating a stable rooted heap with O(1) inspection of maximum is trivial: Extend the keys with an insertion count, and use that in comparisons. A...
1102
Bill Wade
bill.wade@...
Dec 3, 1999 5:32 pm
... Well if you have N elements and M much less than N values, the asymptotic behavior for set<mutable queue<T> > is better. It also doesn't suffer the...
1103
Nathan Myers
ncm@...
Dec 3, 1999 7:12 pm
... This has a problem: multiset is not guaranteed to have the desired property. Typical implementations do, but your code will not be portable. ... This...
1104
jsiek@...
Dec 3, 1999 11:42 pm
Just wondering if anyone has thought of doing a traits class for getting the appropriate value for zero and one given some number type. Cheers, Jeremy...
1105
Howard Hinnant
hinnant@...
Dec 3, 1999 11:51 pm
Not sure if this is exactly what you're looking for, but I'm currently using this in my work: template <typename Size, Size sz> struct number { operator Size()...
1106
jsiek@...
Dec 4, 1999 12:05 am
That solves a different problem than the one I was think of :) I'll elaborate. Suppose you're writing a generic algorithm, and need to call one of the STL...
1107
Darin Adler
darin@...
Dec 4, 1999 6:23 pm
After Greg Colvin posted his implementation of weak pointers and garbage collection, I started thinking about why I use smart pointers in my programs, and what...
1108
Andy Glew
glew@...
Dec 4, 1999 8:43 pm
... I like your semantic breakdown: strong/weak, non-null/optional null. It is better than my keepalive/live breakdown, because it is more consistent with how...
1109
Greg Colvin
gcolvin@...
Dec 4, 1999 9:34 pm
From: Darin Adler <darin@...> ... So it has no default constructor? ... This is an interesting and useful interface, but it isn't what I think is...
