Aug 19 2017
Aug 18 2017
Aug 17 2017
Aug 15 2017
Aug 14 2017
Aug 3 2017
Aug 2 2017
Yup, I've tried reverting D36026 locally, and after doing so check-lit no longer dirties my source tree. I'll comment there with next steps. IMHO reverting and then applying the changes you suggested might be a good way to go!
Aug 1 2017
Jul 31 2017
Looks like this fixes the broken build and should be closed.
Jul 29 2017
Mar 1 2017
Extended test to verify that two DIEs with equal values refer to the same abbrev, and third DIE with another value refers to another abbrev.
- Discarded changes in dwarfgen API
- Modified test: verify generated DWARF abbrevs using MemoryBuffer and verify dump format using raw_string_ostream
- Small style changes:
DW_FORM_implicit_const == getForm()
getForm() == DW_FORM_implicit_const
Feb 28 2017
Feb 27 2017
Feb 25 2017
Jan 10 2017
- Classify DW_FORM_implicit_const as DWARFFormValue::FC_Constant
- Get rid of unnecessary assert()
Jan 9 2017
- DWARFAbbreviationDeclaration::AttributeSpec struct now uses 1 64bit field for both ByteSize and implicit const attribute value (depends on Form == DW_FORM_implicit_const).
- Fixed few mistypes and reordered switch statements to be more logical.
Jan 8 2017
Fix mistype in unittest.
Jan 7 2017
Oct 31 2016
Made test more strict: it now checks that variable i is atomic.
- Reverted CodeView-related changes, replaced with TODO comments instead
- Made DW_TAG_atomic_type available only with DwarfVersion >= 5 (otherwise do not emit it)
- Updated test to match all type tree (also changed dwarf version in test IR to 5)
Oct 26 2016
Oct 19 2016
This patch is conflating two set of changes:
(1) NFC: rename all occurrences of unsigned for alignment purposes in the frontend with uint32_t
(2) shrink all debug-info-related alignment variables from uint64_t -> unint32_t.
I think this patch should only be doing the changes in (2).
- Removed all changes non-related to CGDebugInfo: will post them in separate NFC patch which will switch all unsigned alignment entries to uint32_t.
Oct 18 2016
Can anyone take a look at this please? =)
Oct 17 2016
- Use uint32_t directly for alignment instead of creating typedef ofr it
- Get rid of DebugInfo dependency in AST
- Use uint32_t directly instead of new typedef
- Place uint32_t alignment member after all uint64_t members in DIType class
After letting this sit for a few days, I think that using an unsigned (there doesn't seem to be much precedent for uint32_t) is the way to go. It's more readable than an opaque DIAlignment type (whose size isn't obvious from the name).
I support this.
One question: What happens when we read bitcode that contains an alignment field with a value larger > 2^32?
Does the bitcode reader throw an error? Do we even care to support a non-breaking upgrade path for this?
Oct 14 2016
- Use DIAlignment type instead of uint64_t for alignment in DebugInfo
- Use DIAlignment type instead of uin64_t
- Fix mistype in comment
- HasAlignmentFlag in BitcodeReader is now not static
Use DIAlignment type instead of uint64_t for alignment in DebugInfo.
Oct 10 2016
- Fix indentation in comment.
- Fixed the comment for DILocalVariable in BitcodeWriter to be more self-descriptive.
- Fix mistype in SourceLevelDebugging.rst (attribute() -> attribute__()).