#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;
}


}
ive input instead of 'self-native-input?'.</a></td><td>Ludovic Courtès</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
* gnu/packages/file.scm (file)[self-native-input?]: Remove.
[native-inputs]: New field.
* gnu/packages/guile.scm (guile-1.8)[self-native-input?]: Remove.
[native-inputs]: New field.
(guile-2.0)[self-native-input?]: Remove.
[native-inputs]: Add THIS-PACKAGE when (%current-target-system) is true.
* gnu/packages/ncurses.scm (ncurses)[self-native-input?]: Remove.
[native-inputs]: Add THIS-PACKAGE when (%current-target-system) is true.
* gnu/packages/python-xyz.scm (python-file)[self-native-input?]: Remove.
[native-inputs]: New field.


</td></tr>
<tr class='logheader'><td><span title='2019-02-07 22:10:16 +0100'>2019-02-07</span></td><td class='logsubject'><a href='/guix/commit/gnu/packages/file.scm?id=cc43d807c08af02d31770a011881990d46428737'>gnu: file: Incorporate grafted changes.</a></td><td>Marius Bakke</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
* gnu/packages/file.scm (file)[replacement]: Remove.
[source](patches): New field.
(file/fixed): Remove variable.


</td></tr>
<tr class='logheader'><td><span title='2018-12-24 11:12:52 +0200'>2018-12-24</span></td><td class='logsubject'><a href='/guix/commit/gnu/packages/file.scm?id=e47c69f22638dc92ee962942b8e24284c8dee832'>gnu: file: Fix CVE-2018-10360.</a></td><td>Efraim Flashner</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
* gnu/packages/patches/file-CVE-2018-10360.patch: New file.
* gnu/packages/file.scm (file)[replacement]: New field.
(file/fixed): New variable.
* gnu/packages/commencement.scm (file-boot0): Use 'package/inherit' to
receive security fixes.
* gnu/local.mk (dist_patch_DATA): Register it.


</td></tr>
<tr class='logheader'><td><span title='2018-05-01 03:35:06 +0200'>2018-05-01</span></td><td class='logsubject'><a href='/guix/commit/gnu/packages/file.scm?id=32c18fcb3c3cb10303aa4ef07571867e8ce3fe8d'>gnu: file: Update to 5.33.</a></td><td>Tobias Geerinckx-Rice</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
* gnu/packages/file.scm (file): Update to 5.33.


</td></tr>
<tr class='logheader'><td><span title='2018-01-04 21:21:04 -0500'>2018-01-04</span></td><td class='logsubject'><a href='/guix/commit/gnu/packages/file.scm?id=247d23680a222bc5440dec7c3fe0d36ad463a121'>Merge branch 'master' into core-updates</a></td><td>Mark H Weaver</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
</td></tr>
<tr class='logheader'><td><span title='2018-01-04 12:44:03 -0500'>2018-01-04</span></td><td class='logsubject'><a href='/guix/commit/gnu/packages/file.scm?id=def8a62f627d00503e1fb84ac335cfa3d95a0438'>gnu: file: Mention libmagic.</a></td><td>Leo Famulari</td></tr>
<tr class='nohover-highlight'><td/><td colspan='3' class='logmsg'>
* gnu/packages/file.scm (file)[description]: Mention libmagic.


</td></tr>