Add a newline to separate functions
May 21 2019
May 20 2019
Latest change addresses most of issues.
Outstanding is adding a test when garbage GuardStyle value is fed into config.
The trial testcase used only CHECK-*-NOT which caused complaining from test harness looking for the CHECK- cases.
Looking for something similar in the other checker tests did not turn up anything.
In general it seems like clang-tidy assumes developers get the string names of the options and their values correct and does not provide a lot of (any?) checking.
Change the GuardStyle data type from a string to an enum.
Add test cases for explicitly setting valid style.
Document valid style values.
Reduce llvm-header-guard doc to point to bugprone-header-guard.
May 13 2019
llvm-header-guard is an alias to bugprone-header-guard.
The style is passed through the option 'GuardStyle'
May 6 2019
All of the real work is the header-guard checks is done by their common base class utils/HeaderGuard.
The much smaller llvm and bugprone subclasses are only concerned with the style of the fix.
Move from misc to bugprone
May 4 2019
Addressed issue with returning early and using backtick in doc.
Since misc-header-guard doc was a cut-n-paste of llvm-header-guard, that and other similar docs were changed.
May 3 2019
Apr 15 2019
A comment by whisperity in the origin WIP did not seem to be addressed.
Have you looked at IWYU https://github.com/include-what-you-use/include-what-you-use ?
The end goals of clang-scan-deps and iwyu seem similar so their implementation problems would also be similar.
The iwyu problems doc is https://github.com/include-what-you-use/include-what-you-use/blob/master/docs/WhyIWYUIsDifficult.md
Mar 29 2019
Nov 29 2018
I believe have a fix for the issue you found which can be found in my master-iro branch
Once it has made it through some more testing, I will merge it into this request.
Nov 21 2018
Thanks for giving this a try!
I do not have codemark pro, but tried the scenarios you mentioned out on the codemark.
No build problems, unfortunately.
If you can reproduce the problem, please send me the *.i
The MO/IRO refactoring i am doing may impact your thumb mo.
If you have a public dev branch for this, please send me the link.
Nov 16 2018
Since the crash happens with the iwyu tool, a bit ago i added the testcase at the iwyu project here.
CodesizeOutliner has been changed to IROutliner
cli to enable has changed for *-cso to *-ir-outliner
Generated functions prefix changed for cso_ to _iro_
Some tests add.
Most issues in review fixed.
Exception is for significant refactoring, which was deferred to the larger task for refactoring MO/IRO to use MO infra.
A proof-of-concept of this refactor is https://github.com/trixirt/llvm/tree/master-poc-mo-refactor
Nov 2 2018
Nov 1 2018
Thanks for the quick review!
Oct 31 2018
Aug 26 2018
Aug 23 2018
Aug 21 2018
I do not have commit perms, so can someone commit this for me ?
Aug 20 2018
Changes suggested by Craig.
For readability, move chooseBestLEA's variables closer to call.
Switch to using 'find' as suggested by Craig
Aug 19 2018
I don't think there are any meaningful side effects.
Any comments ?
Aug 15 2018
Jun 27 2018
Sorry for the delay, i will have something (hopefully) this weekend.
Jun 5 2018
cleanup of last patch
Jun 4 2018
Convert to AST Visitor
Jun 1 2018
This change fixes one of the few build warnings.
The extra parenthesis warning is noisy.
I agree there isn't an obvious best path.
I commented the code for this reason.
May 30 2018
May 25 2018
I need someone to commit this..
Sure that is fine.
May 24 2018
I agree on the consensus building.
As llvm is mostly an api, i assumed that it would be robust and this was a one-off.
Anything big should be done automagically, so i am going to look at sic-ing clang-tidy and/or the static analyzer at this problem.
use nothrow version of new
I will change this to try-catch
May 23 2018
Rename test to dwarf5-expansion-checksum.c
May 19 2018
May 15 2018
Improve comment in CLANG_DEFAULT_OBJCOPY to specify what is required of objcopy.
Remove the llvm- prefix from testing to allow for override of default value.
May 11 2018
May 10 2018
Apr 13 2018
Mar 27 2018
Mar 13 2018
Feb 9 2018
The symlinks were useful in the transition from gcc to clang.
Now they don't serve any purpose.
Clang uses libcompiler_rt, not libgcc.
I think it is overkill to look for fbsd 8, it was eol-ed 2015, and 9 in 2016.
A native build of clang 7 on fbsd 8 would be difficult to pull off, needing at least 1-3 intermediate clang's/gcc's
Jan 19 2018
Jan 9 2018
Oct 30 2017
Make SuitbleAlign 128 for all x86 by moving the setting to the base X86TargetInfo class
This only changes 32 bit android's value of 32
Oct 28 2017
As android 32 target is in question, adding reviewers that added the target and reviewed the addition
I believe android's SuitableAlign should be 128 like the other x86's
Oct 27 2017
Better handling of android x86_32
Change the SuitableAlign value only if the new value is bigger than the existing value
Oct 26 2017
Oct 25 2017
Jun 29 2017
Mar 13 2017
Can someone check this in ?
Feb 26 2017
Nov 6 2015
Since it has been a month between reviews, I have moved onto other things.
I would hope that this change would not get stuck on the semantics of
The e500 is broken without this change.
Oct 4 2015
improve commit log to explain long use of cntlz is why this POWER mnemonic is supported.
A testcase distilled from the original problem.
Checks that mfocrf is not used.
Since checking for something not present is brittle, add a positve check.
The second check appends -mattr+mfocrf to the compile arg.
This manually adds back the removed feature and the generated code is
checked for the mfocrf insn.
Sep 30 2015
I will whip something up.
Sep 27 2015
Added alias back to old cntlz mnemonic.
So old, stand alone assembly programs will still work.
Sep 24 2015
May 28 2015
There is also a change to support core file reading, that I believe is
May 3 2015
Handle thumb breakpoints
Detect when in thumb code and silently disable breakpoints
FreeBSD does not support thumb breakpoints.
Define a reasonable breakpoint instruction to enable when thumb breakpoints are supported
Apr 29 2015
Move ObjectFileELF::GetOsFromOSABI functionality to ArchSpec::SetArchitecture
It seems like GetOsFromOSABI's purpose is to compensate for
ArchSpec::SetArchitecture. I think it would be better to improve
SetArchitecture. I will post a change showing how this could be done
Apr 27 2015
I believe that an 'unknown' vendor is normal for some triples.
For example i386-unknown-freebsd10.1