Update variant "all types same/convertible" metaprogramming#3070
Merged
Conversation
Evaluate as a tree instead of a single enormous pack expansion. Fixes microsoft#2770.
To be consistent with STL naming conventions.
Merged
StephanTLavavej
requested changes
Sep 2, 2022
CaseyCarter
commented
Sep 2, 2022
StephanTLavavej
approved these changes
Sep 2, 2022
Member
|
I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed. |
Member
|
Thanks for making |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
And a few other minor changes while we're here. I've separated changes out by commit for ease of review, but this should be squash merged. The most significant change here - others are self-explanatory - replaces the metaprogramming used to enforce the Standard's requirements that "For each valid pack m, e(m) is a valid expression. All such expressions are of the same type
and value category" ([variant.visit]/5).
here are different mechanisms to validate the requirement for
visit<R>and for plainvisit. Previously, we created a single typelist where each type represented one possible configuration of the set of argument variants to a visit call, and then folded over a single giant pack expansion of visitor evaluations ensuring they had the same type and value category (or are all convertible toRforvisit<R>). This was enormously expensive for throughput when the number of possible states is large since the compilers burn a lot of memory maintaining long typelists. The new version uses tree evaluation, where each internal node of the treeands together the result of evaluating its child nodes, which each correspond to a possible state of a single argument variant. Leaves correspond to potential calls and have a boolean value indicating whether that potential call has the same type and value category as the first valid configuration (has a type and value category convertible to the result type forvisit<R>).The resulting speedup for the long-pole
/analyzeconfigs of thevisittests in the "Update LLVM" PR is around 50%.Fixes #2770.
Unblocks #2976.