Re: [chuck-users] implementing a set (was object / class hacking)
Hans: To simultaneously stand my ground and display my ignorance: I've written several flavors of hash tables in my life, but I'm not sure how the Stroustrup code sample helps in this case. If you are suggesting that I hack the Chuck source code, you should know that I do entertain such ideas from time to time, but I usually come back to my senses within 25::samp -- it's just not on my list of priorities. A "written entirely in ChucK" solution -- albeit non-optimal in execution speed -- is good enough for my needs. I'm more into optimizing my days than my milliseconds. Thanks, though. - Rob On 30 Sep 2009, at 14:01, Hans Aberg wrote:
On 30 Sep 2009, at 22:12, Robert Poor wrote:
To clarify: I'd like to implement a set for storing a collection of homogenous objects. Specifically, I need to be able to test for the presence of an object in the set, add an object to the set (if not already present), delete a object from the set, and iterate over all the objects in the set. This last requirements, sadly, eliminates Chuck's "associative arrays".
It's worth pointing out that:
arbitraryobject.toString()
will generate a string of the form "ClassName:hexLocation". Assuming that Chuck NEVER (ever) relocates a live object in memory (Chuck has a reference counting GC, not a copying GC, true?), that string could be useful as a unique ID. But Chuck lacks the string operations to reduce a string into a hash key.
Bjarne Stroustrup, "The C++ Programming Language", 3.1.3, has a simple hash function: name* look(const char* p, int ins=0); name* insert(const char* s) { look(s, 1); }
struct name { char* string; name* next; double value; }
const TBLSZ = 23; name* table[TBLSZ];
name* look(const char* p, int ins=0) { int ii = 0; const char* pp = p; while (*pp) ii = ii<<1 ^ *pp++; // Hash function using eor: ii<<1; ii ^= *pp++;
if (ii < 0) ii = -ii; // Reduce to array range. ii %= TBLSZ;
for (name* n = table[ii]; n; n = n->next) // Search if (strcmp(p, n->string) == 0) return n;
if (ins == 0) error("name not found");
name* nn = new name; // Insert nn->string = new char[strlen(p) + 1]; strcpy(nn->string, p); nn->value = 1; nn->next = table[ii]; table[ii] = nn; return nn; }
Hans
On 30 Sep 2009, at 23:28, Robert Poor wrote:
I've written several flavors of hash tables in my life, but I'm not sure how the Stroustrup code sample helps in this case. If you are suggesting that I hack the Chuck source code, you should know that I do entertain such ideas from time to time, but I usually come back to my senses within 25::samp -- it's just not on my list of priorities.
A "written entirely in ChucK" solution -- albeit non-optimal in execution speed -- is good enough for my needs. I'm more into optimizing my days than my milliseconds.
You might combine it with the earlier idea: class ObjectKey { Object object_; int key_; } If using strings as key, you would need to extract the characters. So perhaps use a global int incremented for each instantiation. Then hash with %. Hans
participants (2)
-
Hans Aberg
-
Robert Poor