- User Since
- Nov 30 2017, 10:52 AM (76 w, 5 d)
fixing support for static functions.
Mon, May 20
adding a nice inline test
Fri, May 17
Note: I think we should probably do something about the linker here too. so use lld-link instead of ld.lld in the Windows case.
adding all windows.h's in all of monorepo
Thu, May 16
Mon, May 13
@compnerd Any other test cases you'd like before the first drop? @jakehehrlich @rupprecht Do things here look good to you guys? Both formats are marked experimental, I plan to work on a merge action next.
Sat, May 11
Fri, May 10
Adding lit test cases.
Changing !tapi-tbe to !experimental-tapi-elf-v1, removing Endianness field from tapi-tbe.
@jakehehrlich What else are you looking for here? Can the patch for merging tbe files come later? @compnerd @rupprecht thoughts? I think this is a reasonable compromise: a minimal subset of what the yaml2obj and llvm-elfabi (with the addition of the Endian field) tools can already consume.
Thu, May 9
renaming things from "ifso" to "interface stubs"
cleaned up VisitNamedDecl
removed all the template functions to handle decls.
addressing a lot of @compnerd 's feedback. Still lots of templating.
Tue, May 7
Mon, May 6
Sat, May 4
Changing experimental flag to -interface-stubs-version=experimental-ifo-elf-v1
Adding support for multiple formats using -interface-stubs-version=.
Thu, May 2
Wed, May 1
- change "ifso" for intermediate files to "ifo"
- Updated schemas. I have both our own proposal and the proposal that resembles elfeabi here.
Tue, Apr 30
Mon, Apr 29
So, @compnerd and I have discussed and we'd like to propose a yaml schema that can express what we need for ifsos for the Sections and the Symbols.
Thu, Apr 25
Wed, Apr 24
First stab at trying to properly handle Weak Symbols
Updated to support properly setting OBJECT/FUNC types as well as section flags.
Tue, Apr 23
I've added an output format. The YAML coming out of -emit-ifso can now run through yaml2obj to produce a striped .o file. Those .o files can be handled by an existing linker.
Mon, Apr 22
Mar 28 2019
Mar 24 2019
Found a better way to do what I needed without this unnecessary change.
Mar 23 2019
Feb 21 2019
Landed in e7e16a7 (github) / r354495 (svn).
Feb 20 2019
Dec 9 2018
Nov 3 2018
Sep 30 2018
- changed std::function to llvm::function_ref
- added cp %t %t.o to tests
Sep 24 2018
Sep 21 2018
Sep 17 2018
Sep 15 2018
Sep 14 2018
-Cleanup: removing DecompressableSection, reusing CompressedSection.
Sep 13 2018
- using two new classes: DecompressableSection and DecompressedSection.
- When a compressed debug section is encountered in makeSection we create a DecompressableSection instead of a Section so that we can keep track of the DecompressedSize (which we need ELFT to read).
- in handleArgs if the Decompression command line arg is set then we replace all of the DecompressableSections with DecompressedSections. The difference being that the decompressed sections unset the SHF_COMPRESSED flag, undo the .zdebug* naming, and set the Section size (and alignment to DecompressedSize and DecompressedAlign. DecompressedSections Writer function also does the decompressing whereas DecompressableSection just writes the original uncompressed data.
Sep 12 2018
Sep 9 2018
Closing. Everything is in. (I think I messed up the case for "Differential Revision" in my commit)
Sep 7 2018
Sep 6 2018
So with clang-6 I get:
- added zlib::isAvailable() checks to various places before compressing.
- fixed up potential alignment issues with Chdr writes.