#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;
}


}
><td><span title='2020-03-25 16:02:53 +0100'>2020-03-25</span></td><td><a href='/guix/commit/tests/ui.scm?id=c136da3fbd92a5db1f889d37f0a7c938ccd63648'>tests: Adjust to 'show-manifest-transaction' changes.</a><span class='msg-avail'>...<span class='msg-tooltip'>This is a followup to 3e5ab0a7a9399bb098b9ced46bf3cbf4085c6bab.

* tests/ui.scm ("show-manifest-transaction"): Update regexp.
* tests/guix-package.sh: Adjust Emacs regexp in --with-source test.
</span></span></td><td>Ludovic Courtès</td></tr>
<tr><td><span title='2020-02-23 12:47:13 +0100'>2020-02-23</span></td><td><a href='/guix/commit/tests/ui.scm?id=fa99c4bbc7acdb8def9ce14a05aacb73f99fe3b1'>ui: (size-&gt;number "1.M") is correctly parsed.</a><span class='msg-avail'>...<span class='msg-tooltip'>Reported by Pierre Neidhardt &lt;mail@ambrevar.xyz&gt;.

* guix/ui.scm (size-&gt;number)[unit-pos]: Add #\. to CHAR-SET:DIGIT.
* tests/ui.scm ("size-&gt;number, 1.M"): New test.
</span></span></td><td>Ludovic Courtès</td></tr>
<tr><td><span title='2019-09-19 23:24:04 +0200'>2019-09-19</span></td><td><a href='/guix/commit/tests/ui.scm?id=d2cdef65605b9e14bfa02c3bf1612ab6b62f4a89'>ui: 'relevance' connects regexps with a logical and.</a><span class='msg-avail'>...<span class='msg-tooltip'>Fixes &lt;https://bugs.gnu.org/36763&gt;.
Previously, the logical and connecting the regexps did not output the expected
results (introduced in 8874faaaac665100a095ef25e39c9a389f5a397f).

* guix/ui.scm (relevance)
[score]: Change its arguments.
[regexp-&gt;score]: New procedure.
* tests/ui.scm ("package-relevance"): Add test.

Signed-off-by: Ludovic Courtès &lt;ludo@gnu.org&gt;
</span></span></td><td>zimoun</td></tr>
<tr><td><span title='2019-06-27 11:14:40 +0200'>2019-06-27</span></td><td><a href='/guix/commit/tests/ui.scm?id=8874faaaac665100a095ef25e39c9a389f5a397f'>ui: 'relevance' considers regexps connected with a logical and.</a><span class='msg-avail'>...<span class='msg-tooltip'>* guix/ui.scm (relevance)[score]: Change to return 0 when one of REGEXPS
doesn't match.
* tests/ui.scm ("package-relevance"): New test.
</span></span></td><td>Ludovic Courtès</td></tr>