--
You received this message because you are subscribed to the Google Groups "cxx" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cxx+uns...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAP_mGKq7Pf92bjGCBhZoTRR21HZaByrnmxZcwhHqx6i5MMDJxA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAAHOzFAK8ctizb7EipBM0H3WNyfO13bgWxJcHL4YPeES-PTKhA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAF3XrKo26bK9EeHNn5n72vXhx97vOeoZjskm6r%3DozA6_7%3DGqxg%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CALB5StYigV_G75cqmr68_Wy5_cqRAt44Gu7CReHi%2BhDGb-vxsg%40mail.gmail.com.
I wouldn't necessarily have made the same choices myself, but I don't see any strong Chromium-specific reason to diverge, which is the usual bar I would use, so I think we should adopt those when they land in the public version.It appears that trunk clang enforces the ordering and non-mixing requirements that C++20 has, even in C++14 mode. If that means that the code we write today is likely to be forward-compatible to C++20, then I can live with it, I think. (Though it does increase my eagerness to jump forward past C++14 as soon as reasonable.)
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CACuR13fQXcpT1rhxUxh4M2-8zb0cE%2BJeQa1ej3JBDuO_rOeXiw%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAHtyhaT6at69SzSh0etCLz0XOGcaKMxihQkYWCwpJA9iz%2BDysQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAMGbLiEDpKp%2BZ-LdcfLJmk_BQ-%3D1AGY7zf9aM2yNzo2o6B4iAQ%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAHk1GDwhD0XsSVB%3D%3Dus8wL-sAZpKpDMLURPFeLPTaW5bDhc5CA%40mail.gmail.com.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAHk1GDxFBAKKvHKkX0hDJyHnfbhvTaweUQcg%3D3tu9bBiy8F2kw%40mail.gmail.com.
Daniel's PR was approved, so the guide updates are live now.
Doh. I'll submit a PR to add the TOC back.(The script that derives the public version of the guide doesn't know about the TOC block, and I regenerated the CL once to make a few small fixes but forgot to restore the TOC.)
On Thu, May 28, 2020 at 4:21 AM Victor Costan <pwn...@chromium.org> wrote:On Tue, May 26, 2020 at 2:22 PM Peter Kasting <pkas...@google.com> wrote:On Tue, May 26, 2020 at 12:29 PM K. Moon <km...@chromium.org> wrote:Daniel's PR was approved, so the guide updates are live now.Woah, the TOC at the top went away. That sucks, I used that a lot :/It seems like at the minimum we should probably clarify in our style guide that for the time being, Chromium targets C++14, not 17, and for further detail in the future people can always look at chromium-cpp.appspot.com. Then we should send out a mail noting the changes, particularly the references change, and follow up with the Blink style owners about formally resolving/removing the its in the Blink style guide relating to this. Am I missing any necessary steps? If not I can probably own at least some of this process.The TOC is back :)The steps above seem reasonable to me. Copy-pasting them below with numbers, for easy reference.1. Clarify in our style guide that for the time being, Chromium targets C++14, not 17, and for further detail in the future people can always look at chromium-cpp.appspot.com.2. Send out a mail noting the changes, particularly the references change.3. Follow up with the Blink style owners about formally resolving/removing the bits in the Blink style guide relating to this.PK, which steps would you be willing to own? I can try to take care of the rest.I'll try and write a CL for (1) today. Once that lands I can do (2).
If there are reasons (such as toolchain support) that we can’t yet move to C++17, can you include them in your (1)? Ideally, this would be tracked by a bug that you could link to.If there are no reasons to hold back on C++17 other than “we haven’t had the discussion yet,” then we should start the discussion.
PK
--
You received this message because you are subscribed to the Google Groups "cxx" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cxx+uns...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/cxx/CAAHOzFAjDADbi-NjdJhOSUDpgyB%3DXLUs3Pn%2BsUHJCG8U37gvhA%40mail.gmail.com.
[Sorry for cross-posting but didn't get a reply on the blink-dev@ thread.]As for mutable references: Will the rule be adjusted globally (cpplint.py) or are we supposed to adjust dependent projects manually by adding "runtime/references" to a linter allow list?
On Thu, Jun 18, 2020 at 2:09 PM Michael Lippautz <mlip...@chromium.org> wrote:[Sorry for cross-posting but didn't get a reply on the blink-dev@ thread.]As for mutable references: Will the rule be adjusted globally (cpplint.py) or are we supposed to adjust dependent projects manually by adding "runtime/references" to a linter allow list?I believe we need to remove runtime/references from depot_tools: https://cs.chromium.org/chromium/tools/depot_tools/cpplint.py?rcl=13d717d087e227dee2f928f2368c0120688b2ac9&l=225
On Thu, Jun 18, 2020 at 2:09 PM Michael Lippautz <mlip...@chromium.org> wrote:[Sorry for cross-posting but didn't get a reply on the blink-dev@ thread.]As for mutable references: Will the rule be adjusted globally (cpplint.py) or are we supposed to adjust dependent projects manually by adding "runtime/references" to a linter allow list?I believe we need to remove runtime/references from depot_tools: https://cs.chromium.org/chromium/tools/depot_tools/cpplint.py?rcl=13d717d087e227dee2f928f2368c0120688b2ac9&l=225
On Thu, Jun 18, 2020 at 11:14 AM <dan...@chromium.org> wrote:On Thu, Jun 18, 2020 at 2:09 PM Michael Lippautz <mlip...@chromium.org> wrote:As for mutable references: Will the rule be adjusted globally (cpplint.py) or are we supposed to adjust dependent projects manually by adding "runtime/references" to a linter allow list?I believe we need to remove runtime/references from depot_tools: https://cs.chromium.org/chromium/tools/depot_tools/cpplint.py?rcl=13d717d087e227dee2f928f2368c0120688b2ac9&l=225I had intentionally not yet replied to the earlier thread since I've had an open CL to do precisely this for a few days, and I was hoping to land it and then say something like "thanks, done": https://chromium-review.googlesource.com/c/chromium/tools/depot_tools/+/2248714
A(const A&);
A& operator=(const A&);
A(A&&);
A& operator=(A&&);
A(const A&);
A(A&&);
A& operator=(const A&);
A& operator=(A&&);