Alignment typed storage #1023
Merged
Alignment typed storage #1023
Conversation
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
jwellbelove
changed the base branch from
master
to
pull-request/#1023-Alignment-typed-storage
|
If you can implement these changes, then I can include this in the next release. |
|
It is implemented in the PR, in etl::typed_storage, isn't it? (As demonstrated in the unit tests.) |
|
It says my requested review changes are 'Pending'. |
|
I have the feeling I'm missing sth. here. :-) Can you please point me to your change requests? I currently can't see any in or https://github.com/ETLCPP/etl/pull/1023/files ? |
jwellbelove
reviewed
Mar 7, 2025
include/etl/alignment.h
Outdated
| public: | ||
|
|
||
| typed_storage_error(string_type file_name_, numeric_type line_number_) | ||
| : alignment_exception(ETL_ERROR_TEXT("typed_storage:error", ETL_ALIGNMENT_FILE_ID"A"), file_name_, line_number_) |
There was a problem hiding this comment.
This should be ETL_ALIGNMENT_FILE_ID"B" as it's the second alignment_exception
There was a problem hiding this comment.
Adjusted accordingly.
include/etl/alignment.h
Outdated
| { | ||
| ETL_ASSERT(!has_value(), ETL_ERROR(etl::typed_storage_error)); | ||
| valid = true; | ||
| return *new (data.template get_address<char>()) value_type(etl::forward<Args>(args)...); |
There was a problem hiding this comment.
I would use ::new here to explicitly reference the global new.
There was a problem hiding this comment.
Adjusted accordingly.
include/etl/alignment.h
Outdated
| /// \returns the instance of T which has been constructed in the internal byte array. | ||
| //*************************************************************************** | ||
| template<typename... Args> | ||
| reference emplace(Args&&... args) |
There was a problem hiding this comment.
For the other classes in the ETL that have a member function that constructs an object, I have used create(Args&&... args).
I can see the reason for using emplace, but it would be inconsistent naming.
Also, create is the common antonym of destroy.
There was a problem hiding this comment.
Are you sure? emplace() is already being used in more ETL classes than create() is:
$ grep -r 'create(Args' include/etl/ -m1 |wc -l
5
$ grep -r 'emplace(Args' include/etl/ -m1 |wc -l
13
$
rolandreichweinbmw
force-pushed
the
alignment-typed-storage
branch
from
8274104 to
d22ca6c
Compare
|
In the last commit in the PR, I replaced emplace() with create(). So depending how you decide regarding this name question above, you can take this one or leave this last commit out. |
|
The feature is already available in the feature extended version of ETL at https://github.com/rolandreichweinbmw/etlplus |
jwellbelove
merged commit 7d91e1f
into
ETLCPP:pull-request/#1023-Alignment-typed-storage
63 checks passed
github-project-automation
bot
moved this from In progress
to Done
in Embedded Template Library
jwellbelove
pushed a commit
that referenced
this pull request
Apr 12, 2025
…github.com/ETLCPP/etl into pull-request/#1023-Alignment-typed-storage
rolandreichweinbmw
added a commit
to rolandreichweinbmw/etl
that referenced
this pull request
Sep 8, 2025
In a recent change to alignment.h, the etl::typed_storage was changed in a way that in case of an already constructed object, the object is created via assignment. However, this contradicts the original use case that led to etl::typed_storage in the first place: ETLCPP#1023 The goal is to omit destructors (and at the same time support classes with deleted assignment operators), so they can be optimized out at link time. This change reverts commit ac7b268 to restore the original functionality and changes the test to reflect the original use case.
rolandreichweinbmw
added a commit
to rolandreichweinbmw/etl
that referenced
this pull request
Sep 8, 2025
In a recent change to alignment.h, the etl::typed_storage was changed in a way that in case of an already constructed object, the object is created via assignment. However, this contradicts the original use case that led to etl::typed_storage in the first place: ETLCPP#1023 The goal is to omit destructors (and at the same time support classes with deleted assignment operators), so they can be optimized out at link time. This change reverts commit ac7b268 to restore the original functionality and changes the test to reflect the original use case.
rolandreichweinbmw
added a commit
to rolandreichweinbmw/etl
that referenced
this pull request
Sep 8, 2025
In a recent change to alignment.h, the etl::typed_storage was changed in a way that in case of an already constructed object, the object is created via assignment. However, this contradicts the original use case that led to etl::typed_storage in the first place: ETLCPP#1023 The goal is to omit destructors (and at the same time support classes with deleted assignment operators), so they can be optimized out at link time. This change reverts commit ac7b268 to restore the original functionality and changes the test to reflect the original use case.
jwellbelove
added a commit
that referenced
this pull request
Sep 10, 2025
* Added coderabbitai configuration * Added builtin mem function tests * Modified etl::typed_storage * Modified etl::typed_storage # Conflicts: # include/etl/alignment.h * Added ETL_NOEXCEPT and ETL_NOEXCEPT_IF_NO_THROW * Added etl::typed_storage_ext and swap for same * Added etl::typed_storage_ext and swap for same # Conflicts: # include/etl/alignment.h * Added release notes * Fixes to GCC -O2 errors * Changed char* parameters to value_type* parameters * Fixed compilation issues for const containers unit tests * Added automatic selection of __builtin_memxxx functions for GCC and clang * Added enhanced coderabbit configuration * Updated version and release notes * Disabled constexpr const container tests for C++11 * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Updated version and release notes * feat: removed unreachable break statements (#1169) * Updated version and release notes * Modified etl::typed_storage # Conflicts: # include/etl/alignment.h * Fix etl::typed_storage by supporting omitted destructors In a recent change to alignment.h, the etl::typed_storage was changed in a way that in case of an already constructed object, the object is created via assignment. However, this contradicts the original use case that led to etl::typed_storage in the first place: #1023 The goal is to omit destructors (and at the same time support classes with deleted assignment operators), so they can be optimized out at link time. This change reverts commit ac7b268 to restore the original functionality and changes the test to reflect the original use case. * Fix missing create() in non-C++11 typed_storage_ext constructor * Typo fix --------- Co-authored-by: John Wellbelove <[email protected]> Co-authored-by: Drew Rife <[email protected]> Co-authored-by: John Wellbelove <[email protected]>
jwellbelove
added a commit
that referenced
this pull request
Sep 11, 2025
* Added coderabbitai configuration * Added builtin mem function tests * Modified etl::typed_storage * Modified etl::typed_storage # Conflicts: # include/etl/alignment.h * Added ETL_NOEXCEPT and ETL_NOEXCEPT_IF_NO_THROW * Added etl::typed_storage_ext and swap for same * Added etl::typed_storage_ext and swap for same # Conflicts: # include/etl/alignment.h * Added release notes * Fixes to GCC -O2 errors * Changed char* parameters to value_type* parameters * Fixed compilation issues for const containers unit tests * Added automatic selection of __builtin_memxxx functions for GCC and clang * Added enhanced coderabbit configuration * Updated version and release notes * Disabled constexpr const container tests for C++11 * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Attempted fixes for MacOS compilation * Updated version and release notes * feat: removed unreachable break statements (#1169) * Updated version and release notes * Modified etl::typed_storage # Conflicts: # include/etl/alignment.h * Fix etl::typed_storage by supporting omitted destructors In a recent change to alignment.h, the etl::typed_storage was changed in a way that in case of an already constructed object, the object is created via assignment. However, this contradicts the original use case that led to etl::typed_storage in the first place: #1023 The goal is to omit destructors (and at the same time support classes with deleted assignment operators), so they can be optimized out at link time. This change reverts commit ac7b268 to restore the original functionality and changes the test to reflect the original use case. * Fix missing create() in non-C++11 typed_storage_ext constructor * Typo fix --------- Co-authored-by: John Wellbelove <[email protected]> Co-authored-by: Drew Rife <[email protected]> Co-authored-by: John Wellbelove <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
I have a recurring use case of globally allocated class instances that I need to explicitly construct in a more defined order than implicit construction of global C++ objects can offer.
Therefore, I have a convenience wrapper etl::typed_storage for aligned_storage_as for instantiating an object explicitly at a defined time.
Previously, I tried with etl::optional, which basically supports the same interface I need, but I can't use this one because it always references ~T. But I want the compiler to optimize away the unneeded dtors because the global objects won't get destroyed anyway until power-off in many embedded scenarios.
(You can still explicitly destroy() if needed.)