Re: [DynInst_API:] dyninst fails to compile with GCC 7

[Ray Chen <rchen@xxxxxxxxxx>]

Duly noted.  I'll put something up on the webpage today.

- Ray

On Thu, Feb 9, 2017 at 12:08 PM, William Cohen <wcohen@xxxxxxxxxx> wrote:
> On 02/09/2017 11:41 AM, Bill Williams wrote:
>> Will--
>>
>> This is not the only issue with Dyninst 9.3.0 and proper compatibility with toolchains, but it gets to go on the list. I'll go through the build log and add explicit fallthrough comments for 9.3.1, which should be coming soon. Thanks for the detailed report. (Though at least some of these cases have fallthrough comments that are simply ill-formatted by gcc7 standards; it seems that gcc could stand to be a bit looser in its matching here.)
>>
>> Reminder for everyone: while the mailing list is a fine place for bug reports, our GitHub issue tracker (https://github.com/dyninst/dyninst/issues) is canonical, and reporting there is the best way to ensure you see all follow-ups on a bug. (It's also the best way to ensure that we don't fix the bug without creating a proper issue.)
>>
>> --bw
>
> Thanks, Bill.  I didn't realize the https://github.com/dyninst/dyninst/issues was now the preferred place to report issues. I didn't see mention of it on http://www.dyninst.org/, but I  didn't spend a lot of time looking around.  It might be good to have something on dyninst.org to point people at the github repo and issue tracker pages. I will use it the github issue the future.
>
> Thanks,
>
> -Will
>
>>
>>
>> ________________________________________
>> From: Dyninst-api <dyninst-api-bounces@xxxxxxxxxxx> on behalf of William Cohen <wcohen@xxxxxxxxxx>
>> Sent: Thursday, February 9, 2017 9:45 AM
>> To: dyninst-api@xxxxxxxxxxx
>> Subject: Re: [DynInst_API:] dyninst fails to compile with GCC 7
>>
>> On 02/09/2017 10:41 AM, William Cohen wrote:
>>> Hi All,
>>>
>>> I attempted to rebuild the dyninst rpm for Fedora rawhide and it failed (https://koji.fedoraproject.org/koji/taskinfo?taskID=17680868).  Fedora Rawhide now uses GCC 7 to compile code.  There are a number of additional check that are performed and the resulting warnings are treated as errors causing the RPM build to fail.  The build log for x86_64 (https://kojipkgs.fedoraproject.org//work/tasks/894/17680894/build.log) shows a number of warning/errors for "-Wimplicit-fallthrough".  SystemTap had a similar issue that was fixed (https://sourceware.org/git/gitweb.cgi?p=systemtap.git;a=commit;h=575fe913f6b6b81421f7bfda7c72ab8c15ecb8e3).
>>>
>>> -Will
>>> _______________________________________________
>>> Dyninst-api mailing list
>>> Dyninst-api@xxxxxxxxxxx
>>> https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api
>>>
>>
>> There is some information about porting to gcc 7 at:
>>
>> https://gcc.gnu.org/gcc-7/porting_to.html
>> https://gcc.gnu.org/gcc-7/changes.html
>>
>> -Will
>>
>> _______________________________________________
>> Dyninst-api mailing list
>> Dyninst-api@xxxxxxxxxxx
>> https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api
>>
>
> _______________________________________________
> Dyninst-api mailing list
> Dyninst-api@xxxxxxxxxxx
> https://lists.cs.wisc.edu/mailman/listinfo/dyninst-api