ryanofsky
commented at 11:11 pm on March 21, 2024:
contributor
Return util::Result objects from all functions that can trigger fatal errors.
There are many validation functions that handle failures by calling AbortNode and triggering shutdowns, without returning error information to their callers. This makes error handling in libbitcoinkernel application code difficult, because the only way to handle these errors is to register for notification callbacks. Improve this by making all functions that trigger fatal errors return util::Result objects with the error information.
This PR is a pure refactoring that returns extra result information from functions without changing their behavior. It’s a possible alternative to and subset of #29642, which adds similar return information but also makes behavior changes and exposes a FatalError type.
This is based on #25665. The non-base commits are:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
ryanofsky marked this as a draft
on Mar 21, 2024
DrahtBot added the label
CI failed
on Mar 22, 2024
DrahtBot
commented at 0:41 am on March 22, 2024:
contributor
🚧 At least one of the CI tasks failed. Make sure to run all tests locally, according to the
documentation.
Possibly this is due to a silent merge conflict (the changes in this pull request being
incompatible with the current code in the target branch). If so, make sure to rebase on the latest
commit of the target branch.
Leave a comment here, if you need help tracking down a confusing failure.
0 test 2024-06-18T23:26:56.881000Z TestFramework (ERROR): Assertion failed
1 Traceback (most recent call last):
2 File "/ci_container_base/ci/scratch/build/bitcoin-x86_64-pc-linux-gnu/test/functional/test_framework/test_framework.py", line 132, in main
3 self.run_test()
4 File "/ci_container_base/ci/scratch/build/bitcoin-x86_64-pc-linux-gnu/test/functional/feature_pruning.py", line 447, in run_test
5 self.reorg_back()
6 File "/ci_container_base/ci/scratch/build/bitcoin-x86_64-pc-linux-gnu/test/functional/feature_pruning.py", line 231, in reorg_back
7 assert not self.nodes[2].verifychain(checklevel=4, nblocks=0)
8 AssertionError
ryanofsky force-pushed
on Jun 20, 2024
DrahtBot added the label
Needs rebase
on Jun 25, 2024
refactor: Add util::Result failure values
Add util::Result support for returning more error information. For better error
handling, adds the ability to return a value on failure, not just a value on
success. This is a key missing feature that made the result class not useful
for functions like LoadChainstate() which produce different errors that need to
be handled differently.
This change also makes some more minor improvements:
- Smaller type size. util::Result<int> is 16 bytes, and util::Result<void> is 8
bytes. Previously util::Result<int> and util::Result<void> were 72 bytes.
- More documentation and tests.
e0b06ee8ec
ryanofsky force-pushed
on Jul 10, 2024
ryanofsky
commented at 5:45 pm on July 10, 2024:
contributor
Make sure to run all tests locally, according to the documentation.
The failure may happen due to a number of reasons, for example:
Possibly due to a silent merge conflict (the changes in this pull request being
incompatible with the current code in the target branch). If so, make sure to rebase on the latest
commit of the target branch.
A sanitizer issue, which can only be found by compiling with the sanitizer and running the
affected test.
An intermittent issue.
Leave a comment here, if you need help tracking down a confusing failure.
DrahtBot added the label
CI failed
on Jul 18, 2024
Previous commit causes the warning to be triggered, but the suboptimal return
from const statement was added earlier in 517e204bacd9d.
Warning was causing CI failure https://cirrus-ci.com/task/6159914970644480
eb006748db
refactor: Add util::Result::Update() method
Add util::Result Update method to make it possible to change the value of an
existing Result object. This will be useful for functions that can return
multiple error and warning messages, and want to be able to change the result
value while keeping existing messages.
1e6b58ed04
refactor: Use util::Result class in LoadChainstate and VerifyLoadedChainstated43dcc10b2
refactor: Add util::Result multiple error and warning messages
Add util::Result support for returning warning messages and multiple errors,
not just a single error string. This provides a way for functions to report
errors and warnings in a standard way, and simplifies interfaces.
The functionality is unit tested here, and put to use in followup PR
https://github.com/bitcoin/bitcoin/pull/25722
ebb45ad32d
test: add static test for util::Result memory usage
Suggested by Martin Leitner-Ankerl <martin.ankerl@gmail.com>
https://github.com/bitcoin/bitcoin/pull/25722#discussion_r1174298529
Co-authored-by: Martin Leitner-Ankerl <martin.ankerl@gmail.com>
15b673d122
ryanofsky force-pushed
on Jul 19, 2024
DrahtBot removed the label
CI failed
on Jul 19, 2024
DrahtBot added the label
Needs rebase
on Jul 26, 2024
Add new InfoType field to util::Result which unlike SuccessType and FailureType
is useful for returning extra information in a result regardless of whether the
function succeeded for failed.
This will be used to return FlushStatus values from libbitcoinkernel functions
indicating if flushed data in addition to whether they succeeded or failed.
The InfoType field is stored directly in the util::Result object, unlike
FailureType and MessagesType which are stored in dynamically allocated memory
and require an extra memory allocation if they are accessed.
83bce7d853
refactor, blockstorage: Return FlushResult from flush methods
Return FlushResult instead of bool from BlockStorage FlushUndoFile,
FlushBlockFile, FlushChainstateBlockFile methods and update all callers of
these methods to use the FlushResult type internally and provide context
information for the flush failure. Three callers:
BlockManager::FindNextBlockPos, BlockManager::WriteUndoDataForBlock, and
Chainstate::FlushStateToDisk will be updated in upcoming commits to bubble
results up to their callers.
d0717d122a
refactor, blockstorage: Return fatal errors from block writes
Return fatal errors from BlockManager methods that write block data. Also
update callers to use the new result types. The callers will be changed to
bubble up the results to their callers in subsequent commits.
348e552da5
refactor, validation: Switch to result.Update
Use result.Update in CompleteChainstateInit() and
ChainstateManager::ActivateSnapshot() so it is possible for them to return
warning messages (about flush failures) in upcoming commits.
CompleteChainstateInit() was previously changed to use util::Result in #25665
and ChainstateManager::ActivateSnapshot() was changed to use it in #30267, but
previously these functions only returned Result values themselves, and did not
call other functions that return Result values. Now, some functions they are
calling will also return Result values, so refactor these functions to use
result.Update so they can merge results and return complete error and warning
messages.
57b1f09b30
refactor, blockstorage: Return fatal error from LoadBlockIndex
Return fatal error and interrupt status from LoadBlockIndex functions and
update callers to use new result types.
d916ea6692
refactor, validation: Return fatal errors from FlushStateToDisk
Return fatal errors from the Chainstate::FlushStateToDisk method and several
small, related methods which wrap it: ForceFlushStateToDisk, PruneAndFlush,
ResizeCoinsCaches, and MaybeRebalanceCaches.
Also add nodiscard annotations so callers do not accidentally ignore the result
values. Callers in init and rpc files are updated to explicitly ignore the
flush results, and other callers (AcceptToMemoryPool, ProcessNewPackage,
DisconnectTip, ConnectTip, ActivateBestChainStep, ActivateSnapshot,
MaybeCompleteSnapshotValidation) are updated to store the results in this
commit, and will be updated in upcoming commits to bubble results up to their
callers.
3d0a05e46d
refactor, validation: Return fatal errors from mempool accept functions
Return fatal errors from AcceptToMemoryPool ProcessNewPackage,
ProcessTransaction, MaybeUpdateMempoolForReorg, and LoadMempool functions.
Also add nodiscard annotations so callers handle the result values. Two callers
ActivateBestChainStep and InvalidateBlock will be updated in upcoming commits
to bubble results up to their callers.
359ed188c3
refactor, validation: Return fatal errors from assumeutxo snapshot functions
Return fatal errors from ActivateSnapshot, MaybeCompleteSnapshotValidation, and
ValidatedSnapshotCleanup functions. Also add nodiscard annotations so callers
handle the result values. One caller, ConnectTip, will be updated in an
upcoming commit to bubble results up to its callers.
98ec565e4e
refactor, validation: Return fatal errors from block connect functions
Return fatal errors from ConnectBlock, ConnectTip, DisconnectTip, and
InvalidateBlock. Also add nodiscard annotations so callers handle the result
values. Three callers: ActivateBestChainStep TestBlockValidity, and
CVerifyDB::VerifyDB will be updated in upcoming commits to bubble results up
to their callers.
db00b423b2
refactor, validation: Return fatal errors from activate best chain functions
Return fatal errors from ActivateBestChain, ActivateBestChainStep, and
PreciousBlock functions. Also add nodiscard annotations so callers handle the
result values. Two callers, ProcessNewBlock and LoadExternalBlockFile, will be
updated in an upcoming commits to bubble results up to its callers.
bce8d65d3c
refactor, validation: Return fatal errors from new block functions
Return fatal errors from AcceptBlock, ProcessNewBlock, TestBlockValidity,
LoadGenesisBlock, and LoadExternalBlockFile. Also add nodiscard annotations so
callers handle the result values.
0d6434fde4
refactor, blockstorage: Return fatal error from ImportBlocks
Return fatal error ImportBlocks function and add nodiscard annotation.
c5db278e19
refactor, validation: Return more errors from VerifyDB
Return ConnectBlock errors from VerifyDB.
3ab9a46ac6
in
src/node/blockstorage.cpp:510
in
e22926a383outdated
507 {
508- if (!LoadBlockIndex(snapshot_blockhash)) {
509- return false;
510+ auto result{LoadBlockIndexData(snapshot_blockhash)};
511+ if (!result || IsInterrupted(*result)) {
512+ return result;
Seems fragile to use util::Result<InterruptResult>. Wrapping one result into another is fragile, because call-sites may easily forget to unwrap the inner result, especially if the inner result is otherwise void. My recommendation would be to treat interruption as util::Error, so that callers don’t miss it if they check for errors (but forget to check for interruption), but are free to check it, if they inspect the precise error.
Seems fragile to use util::Result<InterruptResult>. Wrapping one result into another is fragile, because call-sites may easily forget to unwrap the inner result, especially if the inner result is otherwise void. My recommendation would be to treat interruption as util::Error, so that callers don’t miss it if they check for errors (but forget to check for interruption), but are free to check it, if they inspect the precise error.
It sounds like you want Result<InterruptResult, void> to be replaced with Result<void, InterruptResult> which would be ok with me. I don’t think the difference between these things is very significant, but I agree maybe if the caller forgets to check for the interrupted status, incorrectly treating the interrupted status like a failure might be safer than incorrectly treating the interrupted status like a success.
ryanofsky force-pushed
on Aug 7, 2024
ryanofsky
commented at 6:39 pm on August 7, 2024:
contributor
This is a metadata mirror of the GitHub repository
bitcoin/bitcoin.
This site is not affiliated with GitHub.
Content is generated from a GitHub metadata backup.
generated: 2024-09-28 22:12 UTC
This site is hosted by @0xB10C More mirrored repositories can be found on mirror.b10c.me
