Support compiling rustc without LLVM (try 2) by bjorn3 · Pull Request #42932 · rust-lang/rust

bjorn3 · 2017-06-27T16:21:46Z

Now doesn't change rustc_driver.

Supersedes #42752

rust-highfive · 2017-06-27T16:21:50Z

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

eddyb · 2017-06-27T17:51:35Z

cc @rust-lang/compiler
I am personally not in favor of the approach taken here and would rather make rustc_trans LLVM-only and extract all the other functionality out of it. There are just too many #[cfg(llvm)].

Keruspe · 2017-06-30T08:02:12Z

The comment seems to imply that the build with llvm is broken, not the build without it. Should be the other way round, right?

bors · 2017-07-04T18:29:51Z

☔ The latest upstream changes (presumably #43051) made this pull request unmergeable. Please resolve the merge conflicts.

oli-obk · 2017-07-12T11:02:13Z

I am personally not in favor of the approach taken here and would rather make rustc_trans LLVM-only and extract all the other functionality out of it. There are just too many #[cfg(llvm)].

So llvm would just be a cargo feature of rustc_driver either including rustc_trans_llvm or not, depending on whether the feature is active or not?

eddyb · 2017-07-12T11:57:03Z

@oli-obk More or less, yes.

carols10cents · 2017-07-17T12:21:50Z

friendly ping to keep this on your radar, @bjorn3!

nikomatsakis · 2017-07-17T13:45:45Z

Without looking in detail, I definitely agree with @eddyb that we should strive to factor things better and avoid quite so many #[cfg] gates. @bjorn3, do you need more detailed guidance on how to achieve that?

bjorn3 · 2017-07-20T13:16:55Z

@eddyb is this a bit like you mean. If so I will work it out further in a few days.

bors · 2017-07-22T20:59:13Z

☔ The latest upstream changes (presumably #43059) made this pull request unmergeable. Please resolve the merge conflicts.

alexcrichton · 2017-07-27T23:21:44Z

Looks like this may be ready for another round of review @eddyb!

@bjorn3 it also looks like there may be rebase conflicts here now.

eddyb · 2017-07-28T05:24:04Z

Missing newline at the end of file.

eddyb

Now this looks much more streamlined!

eddyb · 2017-07-28T05:28:26Z

@bjorn3 Ahh how I wish I had a way to un-dismiss github notifications... I keep losing track of some I've gone through on my phone 😞. @alexcrichton Thanks for the ping!

bjorn3 · 2017-07-31T08:19:35Z

Rebased

eddyb · 2017-07-31T08:21:19Z

@bjorn3 Can you squash the commits to remove the back-and-forth (X followed by revert X)?

eddyb · 2017-07-31T08:23:35Z

Was this copied wholesale from rustc_trans? Does that mean ar::Archive is independent of LLVM?
That could make rustc_metadata target-agnostic in the non-dynamic-library case!
cc @alexcrichton @rkruppe

The ar crate is a rust only .a reader and writer.

Then it should be used directly by rustc_metadata and get_rlib_metadata shouldn't exist as a pluggable API (but rather the code live in rustc_metadata) which means loader: Box<MetadataLoader> in rustc_metadata becomes just:
get_dylib_metadata:fn(&Target,&Path) -> Result<ErasedBoxRef<[u8]>,String>
When not available, one can just pass in |_, _| Err(String::from("unsupported")).

bjorn3 · 2017-07-31T12:16:07Z

Squashed

bjorn3 · 2017-08-05T15:35:10Z

Doesn't fail locally (mac os)

alexcrichton · 2017-08-10T19:00:12Z

Hm so looking over this it looks like this is changing how we're reading rlib metadata? The current construction is done specifically for speed as LLVM's ar parsing/reading is much speedier than many other solutions. Are we sure the ar crate is sufficiently fast enough? It initially looks that it is not, as one big difference is using memory maps vs actually reading the file into memory, which can make a huge difference here.

Also from a maintainability perspective, this seems like the kind of patch that'd bitrot almost instantly? Is there something we can do to avoid the #[cfg] jungle that will be very difficult to maintain over time?

eddyb · 2017-08-10T19:21:20Z

@alexcrichton OH! Uhm, when I asked about it it sounded like we already moved away from using the LLVM support - but if that's not true, then the "Move rlib metadata reading code to rustc_metadata and replace MetadataLoader with a function pointer" commit was a mistake...

bjorn3 · 2017-08-11T10:57:39Z

Travis passes.

eddyb · 2017-08-11T11:17:37Z

@bjorn3 Please squash the original commits with the reverts.

bjorn3 · 2017-08-11T13:26:56Z

Squashed and travis passes

eddyb · 2017-08-11T13:36:46Z

@bors r+

bors · 2017-08-11T13:36:47Z

📌 Commit e539996 has been approved by eddyb

bors · 2017-08-11T15:45:16Z

⌛ Testing commit e539996 with merge edd82ee...

Support compiling rustc without LLVM (try 2) Now doesn't change rustc_driver. Supersedes #42752

bors · 2017-08-11T18:37:24Z

☀️ Test successful - status-appveyor, status-travis
Approved by: eddyb
Pushing edd82ee to master...

alexcrichton · 2017-08-11T18:48:26Z

@eddyb and @bjorn3 did y'all have no comments about the maintainability aspect above? Specifically:

Also from a maintainability perspective, this seems like the kind of patch that'd bitrot almost instantly? Is there something we can do to avoid the #[cfg] jungle that will be very difficult to maintain over time?

eddyb · 2017-08-11T18:55:46Z

@alexcrichton So I took a look at the diff to count them, remembering only 5-6... and now there's a lot more. And they're mostly in b7314c7 "Actually make rustc_driver compile without llvm".

eddyb · 2017-08-11T19:09:03Z

So that is definitely my fault, and if @alexcrichton wants this reverted then I'm on board.
@bjorn3 What are you willing to do here? I know I've asked for a lot during this PR.

bjorn3 · 2017-08-12T08:50:17Z

Im writing a pr to remove most of the cfg's. It will likely be ready tomorrow.

alexcrichton · 2017-08-12T15:10:05Z

Thanks @bjorn3! Feel free to cc or r? me on that

@alexcrichton

Cleanup for "Support compiling rustc without LLVM (try 2)" This includes a small patch to allow running tests without llvm. Also check if you are not trying to compile a dylib. cc #42932 r? @alexcrichton

rust-highfive assigned eddyb Jun 27, 2017

bjorn3 mentioned this pull request Jun 27, 2017
Support compiling rustc without LLVM #42752
Closed
2 tasks

arielb1 added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Jun 27, 2017

Keruspe reviewed Jun 30, 2017
View reviewed changes

bjorn3 force-pushed the no_llvm_try2 branch from 7b9617e to 36a3bf3Compare July 20, 2017 13:15

alexcrichton added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 27, 2017

eddyb reviewed Jul 28, 2017
View reviewed changes

eddyb approved these changes Jul 28, 2017
View reviewed changes

bjorn3 force-pushed the no_llvm_try2 branch 2 times, most recently from 4c5839c to 7f191bcCompare July 31, 2017 08:18

eddyb reviewed Jul 31, 2017
View reviewed changes

bjorn3 force-pushed the no_llvm_try2 branch from 7f191bc to af8838dCompare July 31, 2017 12:15

bjorn3 force-pushed the no_llvm_try2 branch 3 times, most recently from 7b64996 to 47cd964Compare August 1, 2017 13:26

bjorn3 force-pushed the no_llvm_try2 branch from 5c1bd88 to e20cdbaCompare August 11, 2017 08:36

bjorn3 added 2 commits August 11, 2017 10:38

Build support for no llvm
e152a16

Make librustc_driver work without librustc_trans
b43c02b

bjorn3 force-pushed the no_llvm_try2 branch from e20cdba to 7e6afacCompare August 11, 2017 09:08

bjorn3 added 3 commits August 11, 2017 14:00

Actually make rustc_driver compile without llvm
b7314c7

It now completely compiles without LLVM!!!
b8d5c74

Remove some unused dependencies from rustc_trans_utils
e539996

bjorn3 force-pushed the no_llvm_try2 branch from 7e6afac to e539996Compare August 11, 2017 12:00

bors added a commit that referenced this pull request Aug 11, 2017
Auto merge of#42932- bjorn3:no_llvm_try2, r=eddyb
edd82ee
Support compiling rustc without LLVM (try 2) Now doesn't change rustc_driver. Supersedes #42752

bors merged commit e539996 into rust-lang:masterAug 11, 2017

This was referenced Aug 11, 2017
Support dynamically-linked and/or native musl targets #40113
Merged
WIP implement a machine readable print-type-info API #43761
Closed

bjorn3 mentioned this pull request Aug 13, 2017
Cleanup for "Support compiling rustc without LLVM (try 2)" #43842
Merged

bjorn3 deleted the no_llvm_try2 branch August 14, 2017 09:44

Uh oh!

Uh oh!

Conversation

bjorn3 commented Jun 27, 2017

Uh oh!

rust-highfive commented Jun 27, 2017

Uh oh!

eddyb commented Jun 27, 2017

Uh oh!

KeruspeJun 30, 2017

Choose a reason for hiding this comment

Uh oh!

bjorn3Jun 30, 2017

Choose a reason for hiding this comment

Uh oh!

bors commented Jul 4, 2017

Uh oh!

oli-obk commented Jul 12, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eddyb commented Jul 12, 2017

Uh oh!

carols10cents commented Jul 17, 2017

Uh oh!

nikomatsakis commented Jul 17, 2017

Uh oh!

bjorn3 commented Jul 20, 2017

Uh oh!

bors commented Jul 22, 2017

Uh oh!

alexcrichton commented Jul 27, 2017

Uh oh!

eddybJul 28, 2017

Choose a reason for hiding this comment

Uh oh!

bjorn3Jul 31, 2017

Choose a reason for hiding this comment

Uh oh!

eddyb left a comment

Choose a reason for hiding this comment

Uh oh!

eddyb commented Jul 28, 2017

Uh oh!

bjorn3 commented Jul 31, 2017

Uh oh!

eddyb commented Jul 31, 2017

Uh oh!

eddybJul 31, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjorn3Jul 31, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

eddybJul 31, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

bjorn3 commented Jul 31, 2017

Uh oh!

bjorn3 commented Aug 5, 2017

Uh oh!

alexcrichton commented Aug 10, 2017

Uh oh!

eddyb commented Aug 10, 2017

Uh oh!

bjorn3 commented Aug 11, 2017

Uh oh!

eddyb commented Aug 11, 2017

Uh oh!

bjorn3 commented Aug 11, 2017

Uh oh!

eddyb commented Aug 11, 2017

Uh oh!

bors commented Aug 11, 2017

Uh oh!

bors commented Aug 11, 2017

Uh oh!

bors commented Aug 11, 2017

oli-obk commented Jul 12, 2017•
edited
Loading

eddybJul 31, 2017•
edited
Loading

bjorn3Jul 31, 2017•
edited
Loading

eddybJul 31, 2017•
edited
Loading