Optimize text_style using bit packing
#4363
Open
+173
−91
Conversation
This file contains 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
In preparation for optimizing `text_style`.
LocalSpook
commented
Feb 25, 2025
| -> text_style { | ||
| return lhs |= rhs; | ||
| } | ||
|
|
||
| FMT_CONSTEXPR auto operator==(text_style rhs) const noexcept -> bool { |
There was a problem hiding this comment.
operator== and operator!= should really be friends, but I ran into this GCC bug when doing that.
LocalSpook
commented
Feb 25, 2025
| // bit is impossible in that case, because 00 (unset color) means the | ||
| // 24 bits that precede the discriminator are all zero. | ||
| // | ||
| // This test can be applied to both colors simultaneously. |
There was a problem hiding this comment.
This is a very bit hacky algorithm; I hope the detailed explanation compensates for that.
vitaut
reviewed
Feb 26, 2025
| return (value_ & (1 << 25)) != 0; | ||
| } | ||
|
|
||
| FMT_CONSTEXPR auto get_value() const noexcept -> uint32_t { |
There was a problem hiding this comment.
I suggest dropping get_ here.
Comment on lines
+270
to
+279
| // We do that check by adding the styles. Consider what adding does to each | ||
| // possible pair of discriminators: | ||
| // 00 + 00 = 000 | ||
| // 01 + 00 = 001 | ||
| // 11 + 00 = 011 | ||
| // 01 + 01 = 010 | ||
| // 11 + 01 = 100 (!!) | ||
| // 11 + 11 = 110 (!!) | ||
| // In the last two cases, the ones we want to catch, the third bit——the | ||
| // overflow bit——is set. Bingo. |
|
One potential concern is the ABI but it seems to be OK because all the functions involved are inline or templated. |
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Outdated suggestions cannot be applied.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.
Currently,
text_styletakes up 20 bytes, but we can pack it into 8.First, a side benefit of that is a smaller and faster implementation of
text_style::operator|(before-and-after), but that function I imagine is usually evaluated at compile-time anyway.The main benefit is reduced binary size per function call using colors. This comes from two factors:
text_stylethrough registers (before-and-after); andstyled_argbecomes smaller, so less data needs to be copied around (before-and-after).This absolutely breaks ABI. (Should the
v11namespace be bumped tov12?) I'm actually not sure if it breaks API;it changes the API of
detail::color_type—an internal type, you would think—but one that's part of the public interface oftext_style.The first commit adds more color tests, because I don't feel the existing suite is comprehensive enough to prove that this change doesn't affect semantics.