Protocol version 2 #3873
Protocol version 2 #3873
Conversation
|
The use of |
|
(quite a big patch :) @arvidn or @ssiloti, |
|
The |
Codecov Report
@@ Coverage Diff @@
## master #3873 +/- ##
==========================================
+ Coverage 70.03% 70.31% +0.27%
==========================================
Files 438 445 +7
Lines 60390 63082 +2692
Branches 8549 8914 +365
==========================================
+ Hits 42297 44354 +2057
- Misses 13759 14239 +480
- Partials 4334 4489 +155
Continue to review full report at Codecov.
|
|
https://travis-ci.org/arvidn/libtorrent/jobs/554481893#L5561 This appears to be a bug in clang, https://bugs.llvm.org/show_bug.cgi?id=21629 |
There was a problem hiding this comment.
You also need this change:
diff --git a/docs/gen_settings_doc.py b/docs/gen_settings_doc.py
index 5f5090af4..fa29a797a 100755
--- a/docs/gen_settings_doc.py
+++ b/docs/gen_settings_doc.py
@@ -76,6 +76,8 @@ for line in f:
mode = 'int'
if '#if TORRENT_ABI_VERSION == 1' in line:
mode += 'skip'
+ if '#if TORRENT_ABI_VERSION <= 2' in line:
+ mode += 'skip'
if '#endif' in line:
mode = mode[0:-4]
|
@ssiloti I ended up putting my changes in a patch: ssiloti#50 |
src/torrent.cpp
Outdated
| @@ -1492,7 +1510,7 @@ bool is_downloading_state(int const st) | |||
| m_save_path, | |||
| static_cast<storage_mode_t>(m_storage_mode), | |||
| m_file_priority, | |||
| m_info_hash | |||
| m_info_hash.has_v1() ? m_info_hash.v1 : m_info_hash.get(protocol_version::V2) | |||
There was a problem hiding this comment.
this basically looks like "get_worst()". would it not work to just use get_best() instead?
There was a problem hiding this comment.
I was probably thinking of backwards compatibility when I wrote this. In hindsight that's not an issue here since hybrid torrents will have a different v1 info-hash from an equivalent v1 only torrent.
| if (req.base == 0) | ||
| { | ||
| std::fill_n(m_hash_verified[req.file].begin() + req.index, req.count, true); | ||
| // TODO: add passing pieces to ret.hash_passed |
There was a problem hiding this comment.
doing this is not critical for the hash_picker's function I take it?
| m_merkle_trees[req.file][merkle_to_flat_index(base_layer_index, req.index + i)] = hashes[i]; | ||
| TORRENT_ASSERT(num_leafs == m_files.piece_length() / default_block_size); | ||
| //verify_block_hashes(m_files.file_offset(req.file) / m_files.piece_length() + req.index); | ||
| // TODO: add to failed hashes |
There was a problem hiding this comment.
this also sounds like an important thing to do
There was a problem hiding this comment.
this is the case where the block-hashes we got from a peer don't match what already know about the merkle tree. is that right?
There was a problem hiding this comment.
This is for the case were the client downloaded piece data before it knew the corresponding piece hashes. Now it has the hashes and discovered that some of the piece data was bad.
src/bt_peer_connection.cpp
Outdated
| while (blocks_per_piece >>= 1) ++layer_index; | ||
| detail::write_uint32(layer_index, ptr); | ||
| } | ||
| #endif |
There was a problem hiding this comment.
we should decide what to do with this block I think. Is this here as a proposed future improvement? (if so, there should probably be a comment to that effect)
There was a problem hiding this comment.
This is obsolete, I'll remove it.
| sha256_hash const& ih = info.info_hash().v2; | ||
| ret += "xt=urn:btmh:1220"; | ||
| ret += aux::to_hex(ih); | ||
| } |
There was a problem hiding this comment.
there should be a test for make_magnet_uri() that creates a v2-only as well as a hybrid magnet URI
… files for v2 torrents
|
https://ci.appveyor.com/project/arvidn/libtorrent/builds/26577613/job/7w5x5kaschyx3nu8#L530 The |
|
did |
fix test_torrent_info.cpp
add missing file, v2_symlinks.torrent
No description provided.