kristina (Kristina Brooks)
Bikeshed Expert

Projects

User Details

User Since
Apr 8 2018, 2:18 PM (24 w, 8 h)

Maintainer of an LLVM backend fork for Broadcom VideoCore4, never went upstream as Julian's GCC port turned out to work a lot better and did not require a separate assembler/linker. Expert on Darwin kernel, Mach-O file format, dynamic linking infrastructure on Darwin (dyld, shared caches, codesigning).

Personal GitHub Account: https://github.com/christinaa
Personal Blog: http://crna.cc/

Not representing any organization.

Recent Activity

Yesterday

kristina added a comment to D45639: [Driver] Support default libc++ library location on Darwin.

I think on Darwin it would not make sense to have libc++fs.a ship in libc++.dylib especially considering that it ends up in the dyld cache and that has a lot of other implications. It would make sense to ship it as a separate library, perhaps as part of the SDK, at least for now. As far as making it a system dylib, it's a possibility as long as no core os components depend on it and it's there solely for consumers.

Sat, Sep 22, 5:09 PM
kristina added a comment to D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..

Want to see what @rnk has to say about this before landing it since he wrote the original code and if my understanding of common vs dso_local is accurate or not since I don't have much experience with the Windows specific metadata, so I'll likely land it on Monday if he's happy with it. I kept cfstring as is, without the dot as per comment and now ELF and PE/COFF share the same paths sans the DLLImport/DLLExport part.

Sat, Sep 22, 4:05 PM

Fri, Sep 21

kristina added a comment to D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..

Binary layout also looks sane (compiled with -fPIC -faddrsig -Wl,--icf=safe), just digging it through it in a dissasembler:

Fri, Sep 21, 4:48 PM
kristina updated the diff for D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..

Does this look about right? I'm not sure what's up with DSO local attribute, it doesn't seem to apply, I think this test makes sense though. ninja check-clang-codegen succeeded this time:

Fri, Sep 21, 4:35 PM
kristina added a comment to D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..

Would you be okay with me renaming cfstring to .cfstring for ELF so it's in line with ELF section naming conventions (I'm not entirely sure if that could cause issues with ObjC stuff)? And yes I think it's best to avoid that code-path altogether if it turns out to be a constant as that's likely from the declaration of the type, I'll write a FileCheck test in a moment and check that I can apply the same logic to ELF aside from the DLL stuff as CoreFoundation needs to export the symbol somehow while preserving the implicit extern attribute for every other TU that comes from the builtin that CFSTR expands to. Is what I'm suggesting making sense? If not, let me know, I may be misunderstanding what's happening here.

Fri, Sep 21, 1:20 PM
kristina added a comment to D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..

Does appear to work properly:

Fri, Sep 21, 3:25 AM
kristina added a reviewer for D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`.: compnerd.
Fri, Sep 21, 2:47 AM
kristina added a reviewer for D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`.: rsmith.
Fri, Sep 21, 2:38 AM
kristina created D52344: [Clang][CodeGen][ObjC]: Fix non-bridged CoreFoundation builds on ELF targets that use `-fconstant-cfstrings`..
Fri, Sep 21, 2:32 AM

Thu, Sep 20

kristina committed rL342705: [Docs] [Support] Correct a missed reference and use up to date .inc examples..
[Docs] [Support] Correct a missed reference and use up to date .inc examples.
Thu, Sep 20, 6:55 PM

Tue, Sep 18

kristina committed rCRT342504: [builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32.
[builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32
Tue, Sep 18, 11:58 AM
kristina committed rL342504: [builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32.
[builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32
Tue, Sep 18, 11:58 AM
kristina closed D43146: [builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32.
Tue, Sep 18, 11:58 AM
kristina added a comment to D43146: [builtins] Fix c?zdi2 on sparc64/Linux and ignore riscv32.

Landing it per request as it's approved (just going to be a second, running it on a buildbot first).

Tue, Sep 18, 11:51 AM
kristina committed rL342502: Remove svn:executable from a documentation file..
Remove svn:executable from a documentation file.
Tue, Sep 18, 11:38 AM
kristina committed rL342500: [Docs] [Support] System Library to Support Library transition along with minor….
[Docs] [Support] System Library to Support Library transition along with minor…
Tue, Sep 18, 11:07 AM
kristina closed D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..
Tue, Sep 18, 11:07 AM

Mon, Sep 17

kristina added a comment to D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..

Friendly ping.

Mon, Sep 17, 9:35 PM

Sun, Sep 16

kristina committed rL342354: [DebugInfo] Fix build when std::vector::iterator is a pointer.
[DebugInfo] Fix build when std::vector::iterator is a pointer
Sun, Sep 16, 3:23 PM
kristina closed D52142: [DebugInfo] Fix build when std::vector::iterator is a pointer.
Sun, Sep 16, 3:23 PM
kristina added a comment to D52142: [DebugInfo] Fix build when std::vector::iterator is a pointer.

Alright, I'll just run this on a single buildbot and then land it.

Sun, Sep 16, 2:56 PM
kristina added a comment to D52151: Also manages clang-X as tool for scan-build.

Just a minor suggestion, I think it would make it more clear as before LLVM 7, Clang did not have a version number with the main executable. GCC is slightly less consistent with their formats as they usually ship as host compilers with various suffixes, but with Clang it can only be clang, clang-7 or 8 (or 9 in the future) although it's still generally invoked as clang unless you have multiple installations.

Sun, Sep 16, 1:15 PM
kristina added a reviewer for D45741: Python bindings: Fix handling of file bodies with multi-byte characters: kristina.
Sun, Sep 16, 12:12 PM
kristina added a comment to D45741: Python bindings: Fix handling of file bodies with multi-byte characters.

Would you mind re-uploading these patches with full context (with diff -U99999). 3 lines of context around changes makes this very difficult to review. Also I would suggest testing for Python version and using appropriate semantics (using encode in case of Python3). Not entirely sure if the bindings are supposed to be compatible with Python3, but I don't see any harm. However explicit tests that avoid changing functionality of existing Python2.7 code would be better in my opinion.

Sun, Sep 16, 12:12 PM
kristina added a comment to D52151: Also manages clang-X as tool for scan-build.

Nitpick: This regex is far too broad to cover the rare use case where you'd be invoking clang as clang-N. I think something like clang(?:\-[789])? would be more suitable?

Sun, Sep 16, 11:54 AM
kristina added a reviewer for D52149: add support functions for hard/soft float multilib distinction: kristina.

This looks fine from a quick glance but would you mind quoting the differential revision in Clang (or making one) which is blocked by this change? I currently do not see this linked to any other proposed change.

Sun, Sep 16, 11:41 AM
kristina added a comment to D52143: Make initializeOutputStream() return false on error and true on success..

My two cents in, false on error makes more sense to me as it's not the same as integer return codes where 0 traditionally indicates success, it makes sense for a verb named function to return false on failure and true on success, or alternatively use a different way of reporting the status back.

Sun, Sep 16, 11:28 AM
kristina added a comment to D52142: [DebugInfo] Fix build when std::vector::iterator is a pointer.

Hi, I know it's a minor change but it should be easy to add a test case for it (a unit test would be appropriate here), especially considering this isn't currently causing issues in any of the current builds as far as I can tell. Also while I'm not insisting, a more-descriptive explanation of the change be nice as it currently, just a sentence or two, so it's more clear on the mailing lists. Thank you.

Sun, Sep 16, 11:22 AM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Hi, I'll leave it in "Needs Review" state until there's some sort of clear consensus on the mailing lists, as far as I understand it you're backed by Debian CI infra, which means you at least have buildbots to ensure this back-end doesn't break or cause regressions in other parts of LLVM. I think as far as that part goes it is fine, but as Chandler explained, it's best to have an existing non-experimental back-end endorse you or getting some form of (even informal) downstream endorsement from Canonical or Debian that can be as simple as confirming you are able and have been maintaining this back-end downstream and that it has kept pace with the current development of LLVM. And as you mentioned yourself, that part is best left to the mailing lists. I would suggest discussing this over with at least one existing non-experimental backend owner at the very least regardless, you can find the list in CODE_OWNERS.txt. If there's some community interest and some form of endorsement (or heavy endorsement from one of the major downstream vendors), I'd be happy to re-review this. Leaving it open as I'll watch the lists and see how this goes and huge apologies for my hasty initial reply.

Sun, Sep 16, 10:57 AM

Sat, Sep 15

kristina abandoned D52093: [ToolChains] Linux.cpp: Use the same style across all all multi-arch includes. NFC..
Sat, Sep 15, 6:49 PM · Restricted Project
kristina added a comment to D52093: [ToolChains] Linux.cpp: Use the same style across all all multi-arch includes. NFC..

Yeah but what about the rest of them, most are completely on their own line, some no newlines, mipsel64 was split across two lines while others weren't, and switch is meant to have default as the topmost case IIRC. I mean I'm if you don't think it's more readable that way I'll abandon it. Though IMHO reading wise it's much easier to read as a list down with separators than as a mix of various newline/no-newline styles.

Sat, Sep 15, 11:38 AM · Restricted Project

Fri, Sep 14

kristina added a reviewer for D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it.: dblaikie.
Fri, Sep 14, 12:06 PM
kristina updated the diff for D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..

Correct index to account for moved page.

Fri, Sep 14, 11:31 AM
kristina updated the diff for D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..

My mistake when combining diffs.

Fri, Sep 14, 11:18 AM
kristina updated the diff for D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..

Moved page.

Fri, Sep 14, 11:05 AM
kristina created D52107: [Docs] [Support] System Library to Support Library transition along with minor corrections to reflect it..
Fri, Sep 14, 11:04 AM
kristina added a reviewer for D52093: [ToolChains] Linux.cpp: Use the same style across all all multi-arch includes. NFC.: sammccall.
Fri, Sep 14, 7:23 AM · Restricted Project
kristina added a reviewer for D52093: [ToolChains] Linux.cpp: Use the same style across all all multi-arch includes. NFC.: Restricted Project.
Fri, Sep 14, 6:21 AM · Restricted Project
kristina created D52093: [ToolChains] Linux.cpp: Use the same style across all all multi-arch includes. NFC..
Fri, Sep 14, 6:21 AM · Restricted Project
kristina committed rL342231: [Driver] Fix missing MultiArch include dir on powerpcspe.
[Driver] Fix missing MultiArch include dir on powerpcspe
Fri, Sep 14, 5:43 AM
kristina committed rC342231: [Driver] Fix missing MultiArch include dir on powerpcspe.
[Driver] Fix missing MultiArch include dir on powerpcspe
Fri, Sep 14, 5:43 AM
kristina closed D52066: [Driver] Fix missing MultiArch include dir on powerpcspe.
Fri, Sep 14, 5:43 AM
kristina closed D52066: [Driver] Fix missing MultiArch include dir on powerpcspe.
Fri, Sep 14, 5:43 AM
kristina accepted D52066: [Driver] Fix missing MultiArch include dir on powerpcspe.

In the future please make sure you supply all patches with context (ie. -U99999), without context they may be harder to review and more importantly patch may mess up.

Fri, Sep 14, 5:31 AM

Thu, Sep 13

kristina updated subscribers of D50858: [M680x0] Add ELF and Triple info.

Also, can I add that since this is a new target, it requires a code owner, could I ask that this is brought up with @chandlerc with regards to that on the mailing lists or even on IRC, since you will be taking responsibility for this (experimental) target and understand the implications of that?

Thu, Sep 13, 5:17 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Apologies, I couldn't find it when I tried. It seems good, and I'm happy to re-review it and land it once it's in a build-able/testable state. My only remaining issue is addressing the failed builds, if you can revise it, I'll rerun it and make sure it's all good since I understand this is a blocking patch for most of the other ones. Again I do apologize and I'm sorry it went unreviewed for so long.

Thu, Sep 13, 5:08 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Also, sorry for the comment flurry, but would you mind bringing this up on llvm-dev as a proposed target? There isn't anything on there, and I was too hasty to accept it since adding a new target is a fairly significant change. I'd be happy to re-review it once the build is fixed and once there's some feedback from the development mailing list.

Thu, Sep 13, 4:54 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Also if possible can you resubmit all of these with context, it doesn't look like they have any and it's possible patch missed something when I applied them to run test builds. Let me know when that's been done, the current patches break the build, unless it's an issue on my end with my buildbots or misapplied patch due to missing context, it should be revised.

Thu, Sep 13, 3:05 PM
kristina added a comment to D51921: [VFS] vfs::directory_iterator yields path and file type instead of full Status.

Fine with me, I wasn't really suggesting a revision just making a remark more or less. Sorry if it came off that way.

Thu, Sep 13, 2:14 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Did you by any chance leave out a file from this patch?

Thu, Sep 13, 1:04 PM
kristina requested changes to D50858: [M680x0] Add ELF and Triple info.

Build failure, x86_64 Linux, Modular, ThinLTO, clean cache:

Thu, Sep 13, 12:51 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

I think i lost track, was there some consensus on this 'support for m680x0' somewhere on the mailing list?
I don't dis-welcome this, just asking.

Thu, Sep 13, 12:43 PM
kristina added a comment to D50858: [M680x0] Add ELF and Triple info.

Do you have commit access? If not, I can land it for you, need a second to test it out with more configurations first though.

Thu, Sep 13, 12:31 PM
kristina accepted D50858: [M680x0] Add ELF and Triple info.

LGTM, thank you for not forgetting to update the modulemap!

Thu, Sep 13, 12:18 PM
kristina added a reviewer for D50858: [M680x0] Add ELF and Triple info: kristina.
Thu, Sep 13, 12:16 PM
kristina added a comment to D51921: [VFS] vfs::directory_iterator yields path and file type instead of full Status.

Added inline comments regarding code style and internal/external naming.

Thu, Sep 13, 4:45 AM

Wed, Sep 12

kristina committed rL342089: [Support] sys::fs::directory_entry includes the file_type..
[Support] sys::fs::directory_entry includes the file_type.
Wed, Sep 12, 3:09 PM
This revision was not accepted when it landed; it landed in state Needs Review.
Wed, Sep 12, 3:09 PM
kristina added a comment to D51918: [Support] sys::fs::directory_entry includes the file_type..

Sam I don't have access to amend your revision, it might be related to you being an
administrator on Phab. Just sending you the patch for the Windows Path.inc.

Wed, Sep 12, 2:51 PM
kristina added a comment to D51918: [Support] sys::fs::directory_entry includes the file_type..

Shall I just fix up the Windows stuff and add another diff if you don't mind?

Thanks very much! I'm more reliant on my IDE than I thought!

Please do upload a new version if it's not too much trouble (I added one that might work, but I can't verify)

Wed, Sep 12, 2:41 PM
kristina added a comment to D51918: [Support] sys::fs::directory_entry includes the file_type..

Shall I just fix up the Windows stuff and add another diff if you don't mind?

Wed, Sep 12, 1:58 PM
kristina requested changes to D51918: [Support] sys::fs::directory_entry includes the file_type..

Windows build failure after grafting the patch:

Wed, Sep 12, 1:46 PM
kristina added a comment to D51918: [Support] sys::fs::directory_entry includes the file_type..

@sammccall I'll run the build on a Windows machine if that gives you any peace of mind, just preparing to.

Wed, Sep 12, 12:55 PM
kristina accepted D51918: [Support] sys::fs::directory_entry includes the file_type..

Thank you!

Wed, Sep 12, 12:34 PM
kristina added a comment to D50359: Add a new library, libclang-cxx.

To elaborate, consider a scenario where a customer wants to build the clang driver, diagtool and just for the sake of the argument change-namespace from extras. Clang tools do not provide nearly the same level of control as most LLVM tools, so in that scenario, you'd have to maintain your own CMakeLists.txt for clang tools(/extra) to neuter unwanted artifacts like clang-tidy or libclang.so (the C API re-exporter) etc.

Wed, Sep 12, 12:29 PM
kristina added a comment to D50359: Add a new library, libclang-cxx.

I'm in support as long as it's a configuration option defaulting similar to LLVM's one. Should likely follow the same naming convention as LLVM, ie. clang-shlib. Clang has a lot of tools it ships with especially if you consider extras, I think this is one of the cases where an exception for libClang-8.so can be made as long as it's not a default. It would make builds of things like clang-tidy much faster without requiring a fully shared library build.

Wed, Sep 12, 11:32 AM
kristina added a comment to D51918: [Support] sys::fs::directory_entry includes the file_type..

Style consistency.

Wed, Sep 12, 11:13 AM
kristina requested changes to D51918: [Support] sys::fs::directory_entry includes the file_type..

Would it be okay to use IT instead of it, as you're already changing those parts, since that follows the general variable naming convention? If possible can same be done for cur_dir and alike?

Wed, Sep 12, 11:01 AM
kristina added a reviewer for D51918: [Support] sys::fs::directory_entry includes the file_type.: kristina.
Wed, Sep 12, 11:00 AM

Mon, Sep 10

kristina added a comment to D51899: Remove extraneous ".a" suffix from baremetal clang_rt.builtins when compiling for baremetal..

Sorry, didn't notice this was recently added, from looking at Darwin and Android embedded tests, this doesn't really match up with either of them. The hosted tests for Android using libbuiltins look like this:

Mon, Sep 10, 6:01 PM
kristina added a reviewer for D51899: Remove extraneous ".a" suffix from baremetal clang_rt.builtins when compiling for baremetal.: t.p.northover.

Added the main ARM code owner as a reviewer, I'm not entirely sure why you're getting this behavior since regression tests should have weeded it out.

Mon, Sep 10, 5:33 PM
kristina added a comment to D51899: Remove extraneous ".a" suffix from baremetal clang_rt.builtins when compiling for baremetal..

The library shouldn't be an object file (it could be if you changed some stuff around but typically isn't), libbuiltins is a combination of multiple translation units, if you look at the source. I'm not sure how it's turning into an object file for you and even if it was you can pass an archive to the driver on the commandline and it would treat it like a group of objects. What is the generated output you're getting when you build it for bare metal?

Mon, Sep 10, 5:24 PM
kristina added a comment to D51899: Remove extraneous ".a" suffix from baremetal clang_rt.builtins when compiling for baremetal..

Please submit your diffs with context in the future (-U99999). Also this seems like it would break a lot, .a is just a generic extension for an object archive file, I don't think the extension is the issue unless you're using some sort of an exotic linker.

Mon, Sep 10, 5:15 PM
kristina closed D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Hi @kristina .

Sure, I didn't mean to do that broader change here. Apologies if it read that way.

Would it be acceptable to add an empty --sysroot= to the test? I can post the change for review in another diff.

Thanks a lot.

Yes, you can submit another diff for this specific test, just use "Update Diff" and add yours on top which should open it up for re-review or use "Commandeer Revision" (and then submit another diff since it was previously closed). Up to your judgement.

Please don't perform necromancy on already committed and closed differentials (unless the commit was reverted, of course).
Do open new differentials.

Mon, Sep 10, 11:12 AM
kristina added a comment to D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Hi @kristina .

Sure, I didn't mean to do that broader change here. Apologies if it read that way.

Would it be acceptable to add an empty --sysroot= to the test? I can post the change for review in another diff.

Thanks a lot.

Mon, Sep 10, 10:58 AM
kristina added a comment to D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Thanks for reopening this @kristina.

I suggest passing --sysroot= to make sure we see the expected behaviour when the sysroot is actually empty.

Note that this would not really test the scenario where DEFAULT_SYSROOT is empty and no --sysroot appears in the command line. I'm not sure if we really want to test that case (but if we do, I think we will have to move that case into a test of its own and add a feature in lit.cfg.py that describes that clang does not have any built-in default sysroot).

Thoughts?

Mon, Sep 10, 10:10 AM
kristina added inline comments to rL341655: Differential Revision: https://reviews.llvm.org/D50246.
Mon, Sep 10, 9:26 AM
kristina reopened D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Seems to be causing a test failure for someone (per comment in rL341655), reopened it for now.

Mon, Sep 10, 9:07 AM
kristina accepted D51835: [ADT] Support converting to lowercase string in toHex.

LGTM, given that there are already good use cases for it.

Mon, Sep 10, 7:35 AM

Sun, Sep 9

kristina added a comment to D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert)..

This is a testcase for the crash, patch fixes it, though I'll amend it to work on MCStreamer level as per what @rnk suggested. I'm going to sound really stupid asking but where should the testcase be, is it a FileCheck based one? I just made two targets, one with pre-patch toolchain and one with current (my working tree) toolchain which handles this just fine. First one crashes, second one doesn't, not sure where tests like that go.

Sun, Sep 9, 4:46 PM
kristina added a comment to D51835: [ADT] Support converting to lowercase string in toHex.

(Really nit-picky comment: I would have done it by keeping an offset aside to add for lower case A-F to a-f and kept the original table as is. Implications of removing static may cause an insignificant increase in code size with bad linkers, just speculating, feel free to ignore)

Sun, Sep 9, 9:00 AM

Sat, Sep 8

kristina added a reviewer for D51830: Add a way to make scripted breakpoints: clayborg.

Please add code reviewers to your differentials when you make them especially when they're this big, at the very least the code owner for that particular project should be added. I added as a reviewer @clayborg as he is the code owner for LLDB.

Sat, Sep 8, 9:07 AM
kristina added a reviewer for D33118: Another version of patch for D25987: RKSimon.
Sat, Sep 8, 9:01 AM
kristina added a comment to D25987: [X86] New pattern to generate PSUBUS from SELECT.

@yulia_koval Are you still working on this?

Sat, Sep 8, 8:59 AM
kristina added a comment to D33118: Another version of patch for D25987.

Please add code reviewers to your differential, a good place to start would be CODE_OWNERS.txt to see who is responsible for what, that will make sure you have at least one assigned reviewer for your code and please try to be more descriptive in the differential message, when it's blank it's hard to relate it to others and D25987 should be mentioned in the the body for Phabricator to pick it up and link it to the previous differential.

Sat, Sep 8, 8:55 AM

Fri, Sep 7

kristina added a reviewer for D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert).: rnk.

Alright, I'll handle it in MCStreamer directly instead, if SMLoc is provided will try to print an error with more context seeing as it could be from -cc1as.

Fri, Sep 7, 2:09 PM
kristina added a comment to D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert)..
In D51695#1227779, @rnk wrote:

I'd suggest fixing this in MCStreamer::EmitCFIStartProc by having it recover in some way that doesn't lead to future assertions. Also, it should be possible to write a standalone .s test case for this.

Fri, Sep 7, 1:38 PM
kristina committed rL341680: [Bindings][Go] Fixed go.test failure due to C-API argument type mismatch..
[Bindings][Go] Fixed go.test failure due to C-API argument type mismatch.
Fri, Sep 7, 10:37 AM
kristina closed D51721: [Bindings][Go] Fixed go.test failure due to C-API argument type mismatch..
Fri, Sep 7, 10:37 AM
kristina added a comment to D51721: [Bindings][Go] Fixed go.test failure due to C-API argument type mismatch..

Sure, one second.

Fri, Sep 7, 10:19 AM
kristina added a comment to D51721: [Bindings][Go] Fixed go.test failure due to C-API argument type mismatch..

I haven't ever filed it, just locally fixed it as I was messing around with trying to build llgo correctly (check the mailing lists for llgo).

Fri, Sep 7, 7:57 AM
kristina retitled D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert). from [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using integrated-as. to [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert)..
Fri, Sep 7, 7:19 AM
kristina updated the summary of D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert)..
Fri, Sep 7, 7:06 AM
kristina added a reviewer for D51695: [MC] [AsmParser]: Ensure a new CFI frame is not opened within an existing one when using cc1as (causing a crash due to failed assert).: debug-info.
Fri, Sep 7, 7:06 AM
kristina added a reviewer for D51778: Introduce LLVM_HOT: chandlerc.
Fri, Sep 7, 6:54 AM
kristina added a comment to D51721: [Bindings][Go] Fixed go.test failure due to C-API argument type mismatch..

Had the same issue when trying to build llgo and had to enable Go bindings, fixed in similar way locally.

Fri, Sep 7, 6:48 AM
kristina closed D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Closed by rC341655.

Fri, Sep 7, 6:11 AM
kristina committed rC341655: Differential Revision: https://reviews.llvm.org/D50246.
Differential Revision: https://reviews.llvm.org/D50246
Fri, Sep 7, 6:05 AM
kristina committed rL341655: Differential Revision: https://reviews.llvm.org/D50246.
Differential Revision: https://reviews.llvm.org/D50246
Fri, Sep 7, 6:05 AM
kristina added a comment to D50246: [RISCV] Add support for computing sysroot for riscv32-unknown-elf.

Going to land it as it's approved by code owner.

Fri, Sep 7, 6:05 AM