#include "references.hh"
#include "hash.hh"
#include "util.hh"
#include "archive.hh"

#include <map>
#include <cstdlib>


namespace nix {


static unsigned int refLength = 32; /* characters */


static void search(const unsigned char * s, unsigned int len, 
    StringSet & hashes, StringSet & seen)
{
    static bool initialised = false;
    static bool isBase32[256];
    if (!initialised) {
        for (unsigned int i = 0; i < 256; ++i) isBase32[i] = false;
        for (unsigned int i = 0; i < base32Chars.size(); ++i)
            isBase32[(unsigned char) base32Chars[i]] = true;
        initialised = true;
    }
    
    for (unsigned int i = 0; i + refLength <= len; ) {
        int j;
        bool match = true;
        for (j = refLength - 1; j >= 0; --j)
            if (!isBase32[(unsigned char) s[i + j]]) {
                i += j + 1;
                match = false;
                break;
            }
        if (!match) continue;
        string ref((const char *) s + i, refLength);
        if (hashes.find(ref) != hashes.end()) {
            debug(format("found reference to `%1%' at offset `%2%'")
                  % ref % i);
            seen.insert(ref);
            hashes.erase(ref);
        }
        ++i;
    }
}


struct RefScanSink : Sink
{
    HashSink hashSink;
    StringSet hashes;
    StringSet seen;

    string tail;

    RefScanSink() : hashSink(htSHA256) { }
    
    void operator () (const unsigned char * data, size_t len);
};


void RefScanSink::operator () (const unsigned char * data, size_t len)
{
    hashSink(data, len);

    /* It's possible that a reference spans the previous and current
       fragment, so search in the concatenation of the tail of the
       previous fragment and the start of the current fragment. */
    string s = tail + string((const char *) data, len > refLength ? refLength : len);
    search((const unsigned char *) s.data(), s.size(), hashes, seen);

    search(data, len, hashes, seen);

    unsigned int tailLen = len <= refLength ? len : refLength;
    tail =
        string(tail, tail.size() < refLength - tailLen ? 0 : tail.size() - (refLength - tailLen)) +
        string((const char *) data + len - tailLen, tailLen);
}


PathSet scanForReferences(const string & path,
    const PathSet & refs, HashResult & hash)
{
    RefScanSink sink;
    std::map<string, Path> backMap;

    /* For efficiency (and a higher hit rate), just search for the
       hash part of the file name.  (This assumes that all references
       have the form `HASH-bla'). */
    foreach (PathSet::const_iterator, i, refs) {
        string baseName = baseNameOf(*i);
        string::size_type pos = baseName.find('-');
        if (pos == string::npos)
            throw Error(format("bad reference `%1%'") % *i);
        string s = string(baseName, 0, pos);
        assert(s.size() == refLength);
        assert(backMap.find(s) == backMap.end());
        // parseHash(htSHA256, s);
        sink.hashes.insert(s);
        backMap[s] = *i;
    }

    /* Look for the hashes in the NAR dump of the path. */
    dumpPath(path, sink);

    /* Map the hashes found back to their store paths. */
    PathSet found;
    foreach (StringSet::iterator, i, sink.seen) {
        std::map<string, Path>::iterator j;
        if ((j = backMap.find(*i)) == backMap.end()) abort();
        found.insert(j->second);
    }

    hash = sink.hashSink.finish();
        
    return found;
}


}
'msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-15 07:43:09 -0400'>2019-08-15</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=9c70c460a05b2bc60f3f3602f0a2dba0f79ce86c'>machine: Implement 'roll-back-machine'.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-15 07:43:03 -0400'>2019-08-15</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=5ea7537b9a650cfa525401c19879080a9cf42e13'>machine: Allow non-root users to deploy.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-14 15:38:09 -0400'>2019-08-14</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=2c8e04f13670c8c7ad8c7195c305960dd1905363'>remote: Build derivations appropriate for the remote's</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-07 18:29:07 -0400'>2019-08-07</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=d84e9b75b2aba98583531f2876d9298871389585'>machine: Add 'build-locally?' field for managed hosts.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-06 16:40:25 -0400'>2019-08-06</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=fd3119db4fed02a1f3491460be776a3b5d2b2cd3'>machine: Implement safety checks.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-08-06 15:48:51 -0400'>2019-08-06</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=d97ce20400b5d2967bfbc0e9fbfb99bbb2daed4f'>machine: Rename 'system' field.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-07-26 19:19:49 +0200'>2019-07-26</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=5c793753b31b1dcd9a554bce953124f7ae88ca9a'>guix system: Add 'reconfigure' module.</a><span class='msg-avail'>...</span></td><td>Jakob L. Kreuze</td></tr>
<tr><td><span title='2019-07-18 12:57:00 +0200'>2019-07-18</span></td><td><a href='/guix/commit/gnu/machine/ssh.scm?id=53f21b3fe9ed57f0feef003185b561efc9b3e30c'>machine: Fix typo.</a><span class='msg-avail'>...</span></td><td>Tobias Geerinckx-Rice</td></tr>