Don't compress metadata #8745
Closed
Don't compress metadata #8745
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
|
So this should cut off ~3 minutes from cycle time? Also:
Not true. The Ubuntu PPA and the Arch repo both have nightlies, bloating that by 7x would annoy at least me, they're already ~17MB. |
|
@cmr: I wouldn't be surprised if the Arch package got smaller with this compression off, because |
|
@cmr Yes, 3 minutes sounds about right. As @thestinger noted when we distribute binaries they tend to be compressed independently. Additionally it's been suggested that in cases where we need to save that space we probably want to strip the metadata from the binaries completely. |
bors
added a commit
that referenced
this pull request
Aug 25, 2013
This does two things: 1) stops compressing metadata, 2) stops copying the metadata section, instead holding a reference to the buffer returned by the LLVM section iterator. Not compressing metadata requires something like 7x the storage space, but makes running tests about 9% faster. This has been a time improvement on all platforms I've tested, including windows. I considered leaving compression as an option but it doesn't seem to be worth the complexity since we don't currently have any use cases where we need to save that space. In order to avoid copying the metadata section I had to hack up extra::ebml a bit to support unsafe buffers. We should probably move it into librustc so that it can evolve to support the compiler without worrying about having a crummy interface. r? @graydon
flip1995
pushed a commit
to flip1995/rust
that referenced
this pull request
Jun 4, 2022
…tiple, r=Manishearth Fix `manual_range_contains` false negative with chains of `&&` and `||` Fixes rust-lang#8745 Since the precedence for `&&` is the same as itself the HIR for a chain of `&&` ends up with a right skewed tree like: ``` && / \ && c2 / \ ... c1 ``` So only the leftmost `&&` was actually "fully" checked, the top level was just `c2` and `&&` so the `manual_range_contains` lint won't apply. This change makes it also check `c2` with `c1`. There's a bit of a hacky solution in the [second commit](rust-lang/rust-clippy@257f097) to check if the number of open/closing parens in the snippet match. This is to prevent a case like `((x % 2 == 0) || (x < 0)) || (x >= 10)` from offering a suggestion like `((x % 2 == 0) || !(0..10).contains(&x)` which now won't compile. Any suggestions for that paren hack welcome, kinda new to working on this so not too sure about possible solutions :) it's weird because I don't know how else to check for parens in HIR considering they're removed when lowering AST. changelog: Fix [`manual_range_contains`] false negative with chains of `&&` and `||`
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.
This does two things: 1) stops compressing metadata, 2) stops copying the metadata section, instead holding a reference to the buffer returned by the LLVM section iterator.
Not compressing metadata requires something like 7x the storage space, but makes running tests about 9% faster. This has been a time improvement on all platforms I've tested, including windows. I considered leaving compression as an option but it doesn't seem to be worth the complexity since we don't currently have any use cases where we need to save that space.
In order to avoid copying the metadata section I had to hack up extra::ebml a bit to support unsafe buffers. We should probably move it into librustc so that it can evolve to support the compiler without worrying about having a crummy interface.
r? @graydon