mirror of
https://codeberg.org/guix/guix.git
synced 2026-09-07 06:14:47 +00:00
Reepca Russelstein
64a0824955
Recognizing base32 characters and/or parsing them into byte sequences is done in several places, so it makes sense to provide a common lookup table. * nix/libutil/hash.hh (base32Values, base32ValuesInitialized): new variables. (initializeBase32Values, getBase32Value): new functions. * nix/libutil/hash.cc (base32Values, base32ValuesInitialized): provide definition for variables. (initializeBase32Values, getBase32Value): provide function implementations. (parseHash32): use base32Values. * nix/libstore/references.cc (search): use base32Values. Signed-off-by: Ludovic Courtès <ludo@gnu.org>
114 lines
3.0 KiB
C++
114 lines
3.0 KiB
C++
#include "references.hh"
|
|
#include "hash.hh"
|
|
#include "util.hh"
|
|
#include "archive.hh"
|
|
|
|
#include <map>
|
|
#include <cstdlib>
|
|
#include <cassert>
|
|
#include <format>
|
|
|
|
namespace nix {
|
|
|
|
|
|
static unsigned int refLength = 32; /* characters */
|
|
|
|
|
|
static void search(const unsigned char * s, unsigned int len,
|
|
StringSet & hashes, StringSet & seen)
|
|
{
|
|
for (unsigned int i = 0; i + refLength <= len; ) {
|
|
int j;
|
|
bool match = true;
|
|
for (j = refLength - 1; j >= 0; --j)
|
|
if (base32Values[(unsigned char) s[i + j]] == -1) {
|
|
i += j + 1;
|
|
match = false;
|
|
break;
|
|
}
|
|
if (!match) continue;
|
|
string ref((const char *) s + i, refLength);
|
|
if (hashes.find(ref) != hashes.end()) {
|
|
debug(std::format("found reference to `{}' at offset `{}'", 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'). */
|
|
for (const auto& i : refs) {
|
|
string baseName = baseNameOf(i);
|
|
string::size_type pos = baseName.find('-');
|
|
if (pos == string::npos)
|
|
throw Error(std::format("bad reference `{}'", 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;
|
|
for (auto& 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;
|
|
}
|
|
|
|
|
|
}
|