Author: zinovy.nis
Date: Wed Apr 11 22:41:24 2018
New Revision: 329873
URL: http://llvm.org/viewvc/llvm-project?rev=329873&view=rev
Log:
[clang-tidy] [modernize-use-auto] Get only a length of token, not the token
itself
Modified:
clang-tools-extra/trunk/clang-tidy/modernize/UseAutoCheck.cpp
Test?
On Thu, Apr 12, 2018 at 8:41 AM, Zinovy Nis via cfe-commits
wrote:
> Author: zinovy.nis
> Date: Wed Apr 11 22:41:24 2018
> New Revision: 329873
>
> URL: http://llvm.org/viewvc/llvm-project?rev=329873&view=rev
> Log:
> [clang-tidy] [modernize-use-auto] Get only a length of token, not the tok
Also, i guess you now can remove linking to clangTooling that was
required to unbreak the build.
I'm guessing clangLex contains Lexer::MeasureTokenLength() ?
On Thu, Apr 12, 2018 at 10:22 AM, Roman Lebedev wrote:
> Test?
>
> On Thu, Apr 12, 2018 at 8:41 AM, Zinovy Nis via cfe-commits
> wrote:
>>
Yes, due to to incorrect token length returned by the old code, the test
was passing. I fixed it.
Hm, you are right, clangTooling is not requirted anymore.
чт, 12 апр. 2018 г. в 10:25, Roman Lebedev :
> Also, i guess you now can remove linking to clangTooling that was
> required to unbreak the b
On Thu, Apr 12, 2018 at 10:30 AM, Zinovy Nis wrote:
> Yes, due to to incorrect token length returned by the old code, the test was
> passing. I fixed it.
Let me rephrase.
If i revert just the
clang-tools-extra/trunk/clang-tidy/modernize/UseAutoCheck.cpp change,
will the corresponding tests in
clan
Yes, a test that should fail, did not fail with old code. And it's a
problem)
чт, 12 апр. 2018 г. в 10:35, Roman Lebedev :
> On Thu, Apr 12, 2018 at 10:30 AM, Zinovy Nis wrote:
> > Yes, due to to incorrect token length returned by the old code, the test
> was
> > passing. I fixed it.
> Let me r
On Thu, Apr 12, 2018 at 7:44 AM Zinovy Nis via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: zinovy.nis
> Date: Wed Apr 11 22:41:24 2018
> New Revision: 329873
>
> URL: http://llvm.org/viewvc/llvm-project?rev=329873&view=rev
> Log:
> [clang-tidy] [modernize-use-auto] Get only a length
I composed a test and you are right - for multi-token type names new code
returns wrong result. So this new code was a premature optimization. Sorry.
I'll see how to fix it in a better way and create a review on it. Thanks
for poiting.
чт, 12 апр. 2018 г. в 20:01, Alexander Kornienko :
> On Thu,
On Thu, Apr 12, 2018 at 10:47 PM, Zinovy Nis via cfe-commits
wrote:
> I composed a test and you are right - for multi-token type names new code
> returns wrong result. So this new code was a premature optimization. Sorry.
> I'll see how to fix it in a better way and create a review on it. Thanks f