i128 and u128 support by est31 · Pull Request #38482 · rust-lang/rust

est31 · 2016-12-20T01:22:45Z

Brings i128 and u128 support to nightly rust, behind a feature flag. The goal of this PR is to do the bulk of the work for 128 bit integer support. Smaller but just as tricky features needed for stabilisation like 128 bit enum discriminants are left for future PRs.

Rebased version of #37900, which in turn was a rebase + improvement of #35954 . Sadly I couldn't reopen #37900 due to github. There goes my premium position in the homu queue...

[plugin-breaking-change]

cc #35118 (tracking issue)

est31 · 2016-12-20T01:22:54Z

r? @eddyb

rust-highfive · 2016-12-20T01:22:58Z

r? @alexcrichton

(rust_highfive has picked a reviewer for you, use r? to override)

eddyb · 2016-12-20T01:25:34Z

@bors r+

bors · 2016-12-20T01:25:35Z

📌 Commit 84fd6c3 has been approved by eddyb

est31 · 2016-12-20T01:35:14Z

I've removed the workaround for #37686 as the bootstrap compiler got updated.

r? @eddyb

eddyb · 2016-12-20T01:38:21Z

@bors r+

bors · 2016-12-20T01:38:22Z

📌 Commit 1abfc5c has been approved by eddyb

eddyb · 2016-12-20T05:18:45Z

@bors p=100 (to avoid bitrotting)

retep998 · 2016-12-20T05:33:07Z

Sadly I couldn't reopen #37900 due to github.

If you push to a PR branch while the PR is closed, github won't let you reopen it. You have to reopen it first, then push the new version.

est31 · 2016-12-20T05:33:53Z

@retep998 yes, found that out after I've already pushed it. :/

samueltardieu

I read the diff out of curiosity and was surprised by this.

samueltardieu · 2016-12-20T10:17:57Z

+ -0x0000_0000_0000_8000...0x0000_0000_0000_7fff => I16,
+ -0x0000_0000_8000_0000...0x0000_0000_7fff_ffff => I32,
+ -0x8000_0000_0000_0000...0x7fff_ffff_ffff_ffff => I64,
+ _ => I128


Match argument is an i64 there, I128 will never be returned.

See #37900 (comment)

Would it be better to actually add the "proper" case, but have the *128 case return a *64 instead, with a comment linking to the relevant issue?

samueltardieu · 2016-12-20T10:18:27Z

+ 0...0x0000_0000_0000_ffff => I16,
+ 0...0x0000_0000_ffff_ffff => I32,
+ 0...0xffff_ffff_ffff_ffff => I64,
+ _ => I128,


Match argument is an u64 there, I128 will never be returned.

bors · 2016-12-20T17:12:52Z

⌛ Testing commit 1abfc5c with merge 764d2f6...

bors · 2016-12-20T17:41:56Z

💔 Test failed - auto-mac-32-opt

alexcrichton · 2016-12-20T17:44:53Z

@bors: retry

…

On Tue, Dec 20, 2016 at 9:41 AM, bors ***@***.***> wrote: 💔 Test failed - auto-mac-32-opt <https://buildbot.rust-lang.org/builders/auto-mac-32-opt/builds/11397> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#38482 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAD95FjO7LrbLl3Ari1p96O-5GH0qgMpks5rKBNmgaJpZM4LRXRV> .

alexcrichton · 2016-12-20T18:37:49Z

@bors: r-

Looking at some travis logs it looks like this doesn't compile on 32-bit maybe? I think the error is:

error[E0053]: method `to_ret` has an incompatible type for trait
--> /checkout/src/libcompiler_builtins/lib.rs:505:28
|
426 | fn to_ret(self) -> Self::Ret;
| --------- type in trait
...
505 | fn to_ret(self) -> u128ret {
| ^^^^^^^ expected i128, found u128
error: aborting due to previous error

bors · 2016-12-21T02:52:15Z

☔ The latest upstream changes (presumably #38499) made this pull request unmergeable. Please resolve the merge conflicts.

est31 · 2016-12-21T03:46:01Z

Rebased. r? @eddyb

eddyb · 2016-12-21T04:15:04Z

@bors r+

bors · 2016-12-21T04:15:05Z

📌 Commit 5b4fe10 has been approved by eddyb

bors · 2016-12-21T06:46:33Z

⌛ Testing commit 5b4fe10 with merge 1602354...

bors · 2016-12-21T07:28:03Z

💔 Test failed - auto-win-msvc-32-opt

est31 · 2016-12-21T09:16:15Z

Ok, this seems like another windows ABI issue with the intrinsics, this time on 32 bit. I'm trying to reproduce the issue right now.

Fixing the bug is a matter of finding out what LLVM expects as ABI in this case and replicating it in Rust.

bors · 2016-12-31T17:20:24Z

📌 Commit 29e01af has been approved by eddyb

bors · 2016-12-31T18:54:18Z

⌛ Testing commit 29e01af with merge 38bd207...

i128 and u128 support Brings i128 and u128 support to nightly rust, behind a feature flag. The goal of this PR is to do the bulk of the work for 128 bit integer support. Smaller but just as tricky features needed for stabilisation like 128 bit enum discriminants are left for future PRs. Rebased version of #37900, which in turn was a rebase + improvement of #35954 . Sadly I couldn't reopen #37900 due to github. There goes my premium position in the homu queue... [plugin-breaking-change] cc #35118 (tracking issue)

bors · 2016-12-31T21:00:34Z

☀️ Test successful - status-appveyor, status-travis
Approved by: eddyb
Pushing 38bd207 to master...

jackpot51 · 2016-12-31T23:13:59Z

Awesome!

est31 · 2017-01-01T02:31:26Z

wow, great!!

🎉

it was the

FIRST TRY

I assure you!!

http://devopsreactions.tumblr.com/post/151057980189/getting-your-build-through-the-pipeline

DirkyJerky · 2017-01-01T07:23:08Z

This passed on the first try in 2017! (in utc + (-3)..9 ).
Only acceptable explanation is that 2016 truly was the year of... what animal

We introduced the unadjusted ABI to work around wrong (buggy) ABI expectations by LLVM on Windows [1]. Therefore, it should be solely used on Windows and not on other platforms, like right now is the case. [1]: see this comment for details rust-lang#38482 (comment)

…gisa Don't use "unadjusted" ABI on non windows platforms We introduced the unadjusted ABI to work around wrong (buggy) ABI expectations by LLVM on Windows [1]. Therefore, it should be solely used on Windows and not on other platforms, like right now is the case. [1]: see this comment for details rust-lang#38482 (comment)

As discussed bigint + u8 for decimal representation also avoiding the use of generics for the most part. For now the only imaginable way to have errors in my mind is decimal mismatch. Lots to fill in here and some open questions Do we want this in a separate crate? I imagine it is possible the eth-rpc crate may ultimately depend on this functionality. I just figured out that u128 is natively available in rust on all targets as LLVM has support for it rust simply exposes that. See [this pr](rust-lang/rust#38482) from late 2016. This has been [stablized in 1.26](https://blog.rust-lang.org/2018/05/10/Rust-1.26.html). The representable range here even with a token that is using 18 decimals of precision is 19 9s. If you have 10^19 of something I think it is alright to say that you need to do two transactions. I believe that u128 represents the best combination of speed flexibility and consistency. It is very unlikely to me that any token balance will natively overflow this representation.

rust-highfive assigned alexcrichton Dec 20, 2016

rust-highfive assigned eddyb and unassigned alexcrichton Dec 20, 2016

est31 mentioned this pull request Dec 20, 2016
i128 and u128 support #37900
Closed
2 tasks

est31 force-pushed the i128 branch from 84fd6c3 to 1abfc5cCompare December 20, 2016 01:31

samueltardieu reviewed Dec 20, 2016
View reviewed changes

est31 force-pushed the i128 branch from b73646b to 5b4fe10Compare December 21, 2016 03:45

bors merged commit 29e01af into rust-lang:masterDec 31, 2016

This was referenced Dec 31, 2016
calling convention for MSP430 interrupts #38465
Merged
[WIP] Bump allocator for rustc #38725
Closed

wdv4758h mentioned this pull request Jan 1, 2017
Support i128 and u128 rust-lang/rust.vim#133
Merged

japaric mentioned this pull request Jan 4, 2017
LLVM assertion when using i128 with the NVPTX #38824
Closed

This was referenced Jan 5, 2017
rust-lang/rust integration metabug rust-lang/compiler-builtins#66
Closed
members of ::std::u128 module must not be stable #38860
Closed

alygin mentioned this pull request Jan 6, 2017
Support u128 and i128 intellij-rust/intellij-rust#895
Closed

FaultyRAM mentioned this pull request Jan 7, 2017
128-bit integers are available in nightly Rust robojeb/pcg_rand#2
Closed

est31 mentioned this pull request Jan 14, 2017
Constants seem to be broken in nightly Rust on powerpc-unknown-linux-gnu #39061
Closed

est31 mentioned this pull request Feb 2, 2017
Don't use "unadjusted" ABI on non windows platforms #39472
Merged

cuviper mentioned this pull request Mar 4, 2017
[beta] rustc_const_math: undefined reference to `rust_eh_personality' #40251
Closed

willstott101 mentioned this pull request Jun 27, 2019
[Rust] Add u128 and i128 primitive types. sublimehq/Packages#2002
Merged

lygstate mentioned this pull request Oct 15, 2020
Suggestions for additional floating-point types rust-lang/rfcs#2629
Open

est31 mentioned this pull request Oct 20, 2020
Fix two small issues in compiler/rustc_lint/src/types.rs #78155
Merged

waynenilsen mentioned this pull request Nov 20, 2020
[COMP-1165] fixed precision decimal implementation (wip) compound-finance/gateway#11
Closed

asterite mentioned this pull request Sep 13, 2021
Int128 compiler-rt methods (Int128 literal support part 1) crystal-lang/crystal#11206
Merged

Uh oh!

Uh oh!

Conversation

est31 commented Dec 20, 2016

Uh oh!

est31 commented Dec 20, 2016

Uh oh!

rust-highfive commented Dec 20, 2016

Uh oh!

eddyb commented Dec 20, 2016

Uh oh!

bors commented Dec 20, 2016

Uh oh!

est31 commented Dec 20, 2016

Uh oh!

eddyb commented Dec 20, 2016

Uh oh!

bors commented Dec 20, 2016

Uh oh!

eddyb commented Dec 20, 2016

Uh oh!

retep998 commented Dec 20, 2016

Uh oh!

est31 commented Dec 20, 2016

Uh oh!

samueltardieu left a comment

Choose a reason for hiding this comment

Uh oh!

samueltardieuDec 20, 2016

Choose a reason for hiding this comment

Uh oh!

sinkuuDec 20, 2016

Choose a reason for hiding this comment

Uh oh!

DirkyJerkyDec 20, 2016

Choose a reason for hiding this comment

Uh oh!

samueltardieuDec 20, 2016

Choose a reason for hiding this comment

Uh oh!

bors commented Dec 20, 2016

Uh oh!

bors commented Dec 20, 2016

Uh oh!

alexcrichton commented Dec 20, 2016 via email

Uh oh!

alexcrichton commented Dec 20, 2016

Uh oh!

bors commented Dec 21, 2016

Uh oh!

est31 commented Dec 21, 2016

Uh oh!

eddyb commented Dec 21, 2016

Uh oh!

bors commented Dec 21, 2016

Uh oh!

bors commented Dec 21, 2016

Uh oh!

bors commented Dec 21, 2016

Uh oh!

est31 commented Dec 21, 2016

Uh oh!

bors commented Dec 31, 2016

Uh oh!

bors commented Dec 31, 2016

Uh oh!

bors commented Dec 31, 2016

Uh oh!

jackpot51 commented Dec 31, 2016

Uh oh!

est31 commented Jan 1, 2017

FIRST TRY

Uh oh!

DirkyJerky commented Jan 1, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone