[io] Write byte counts in streamer field#22430
Open
jblomer wants to merge 6 commits into
Open
Conversation
The version 0 and version 1 streamer fields are exactly the same for small objects (<=1GiB). For larger objects, the version 1 streamer field stores the byte count stack before the data. This is up to a future patch. For the moment, we assert that the streamed buffer is never >1GiB.
jblomer
changed the title
Test Results 22 files 22 suites 3d 10h 28m 44s ⏱️ For more details on these failures, see this check. Results for commit 3b67b13. ♻️ This comment has been updated with latest results. |
Add version 1 streamer field.
Uses RStreamerField as a test bed for (de-)serializing large objects with TBufferFile.
jblomer
force-pushed
the
streamer-field-bytecount-v2
branch
from
fd54941 to
3b67b13
Compare
hahnjo
reviewed
Jun 3, 2026
| if (static_cast<std::size_t>(nbytes) > kMaxSmallBuffer) { | ||
| throw RException(R__FAIL("large objects (>1GiB) not supported by the version 0 streamer field")); | ||
| } else { | ||
| assert(buffer.GetByteCounts().empty()); |
Member
There was a problem hiding this comment.
For back-/forwards-porting this assert will have to be removed...
| Both versions have an identical on-disk representation when the streamed object is smaller than 1GiB. | ||
| Only version 1 supports larger streamed objects. | ||
| For large objects, the version 1 streamer field prepends the large byte counts ("byte count stack") to the byte stream. | ||
| The format for the version 1 byte stream is |
Member
There was a problem hiding this comment.
For objects bigger than 1 GiB, the format [...] is:
| The format for the version 1 byte stream is | ||
|
|
||
| - 64bit unsigned integer: number of elements in the large byte count list | ||
| - List if 64bit unsigned integer pairs with the byte count location and byte count value` |
Member
There was a problem hiding this comment.
"List of" plus stray backtick at the end of the line
| const std::size_t szBufCounts = sizeof(std::uint64_t) * (2 * nCounts + 1); | ||
| if (szBufCounts > nbytes) | ||
| throw RException(R__FAIL("invalid byte count size in streamer field: " + std::to_string(nCounts))); | ||
| bufCounts.resize(szBufCounts); |
Member
There was a problem hiding this comment.
This will zero-initialize the elements. Maybe use Internal::MakeUninitArray as for writing?
Comment on lines
+1353
to
+1356
| std::vector<unsigned char> bufCounts(sizeof(std::uint64_t)); | ||
| fAuxiliaryColumn->ReadV(collectionStart, sizeof(std::uint64_t), bufCounts.data()); | ||
| std::uint64_t nCounts; | ||
| std::size_t pos = Internal::RNTupleSerializer::DeserializeUInt64(bufCounts.data(), nCounts); |
Member
There was a problem hiding this comment.
If using Internal::MakeUninitArray below, this could be a static array on the stack
|
|
||
| namespace { | ||
|
|
||
| class FileRaii { |
Member
There was a problem hiding this comment.
We have this in tree/ntuple/test/ntuple_test.hxx, no?
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds support for streamer field v1, which writes the byte counts of large objects into the stream.
The first two commits should be backported.
Replaces #20608