Re: Why is conflict check on the buildbot green?

[Jaap Keuter <jaap.keuter () xs4all nl>]

Hi,

Okay, we should at least try to ‘pick the low hanging fruit’ of this. I’ve started with these already: 37694 
<https://code.wireshark.org/review/37694>, 37692 <https://code.wireshark.org/review/37692> and 37697 
<https://code.wireshark.org/review/37697> 
It would be great if at least the obvious cases 
<https://buildbot.wireshark.org/petri-dish/builders/Ubuntu%20Petri%20Dish%20x64/builds/11982/steps/conflict%20check/logs/stdio>
 can be addressed.


Thanks,
Jaap


> On 5 Jul 2020, at 11:57, Alexis La Goutte <alexis.lagoutte () gmail com> wrote:
>
> Yes it is true ! But I don’t how to set orange with buildbot on this case, some others step (clang analyzer check, 
> pre commit...) need this.
>
> And also if it is orange, it is not specific to this change..
> Le dim. 5 juil. 2020 à 11:16, Jaap Keuter <jaap.keuter () xs4all nl <mailto:jaap.keuter () xs4all nl>> a écrit :
> Hi Alexis,
>
> "issues are not yet fixed” sounds a bit weird for a reason for marking this stage okay (green). That would be the 
> same as... ignoring all the compilation warnings.
> I’m not expecting it to be marked as error (red), but as warning (orange), calling attention to it. Because I wonder 
> how many are aware of these reported issues.
>
> Thanks,
> Jaap
>
>
> > On 5 Jul 2020, at 11:03, Alexis La Goutte <alexis.lagoutte () gmail com <mailto:alexis.lagoutte () gmail com>> wrote:
> >
> > Hi Jaap,
> >
> > It is beacuse all issue are not yet fixed (some coming from generated code... or missing spec info to known what the 
> > correct fix).
> >
> > Personally, when review code after Petri dish, I try to look different output log, because I know don’t fail it is 
> > juste warning.
> >
> > Le dim. 5 juil. 2020 à 08:53, Jaap Keuter <jaap.keuter () xs4all nl <mailto:jaap.keuter () xs4all nl>> a écrit :
> > Hi,
> >
> > Due to some recent issues with DHCPv6 the buildbot began flagging the 'conflict check' stage as failed. This drew my 
> > attention to the fact that there is a long list of warnings in there about wrong use of protocol fields, but once 
> > the DHCPv6 issues were fixed the build happily went back to green for the ‘conflict check’. Why is is not orange 
> > with the warning output?
> >
> > This is the output of a recent build:
> >
> > 'HI2Operations.latitude' exists multiple times with NOT compatible types: FT_UINT32 and FT_STRING
> > 'HI2Operations.longitude' exists multiple times with NOT compatible types: FT_INT32 and FT_STRING
> > 'HI2Operations.mcc' exists multiple times with NOT compatible types: FT_BYTES and FT_UINT32
> > 'HI2Operations.mnc' exists multiple times with NOT compatible types: FT_BYTES and FT_UINT32
> > 'HI2Operations.lai' exists multiple times with NOT compatible types: FT_BYTES and FT_UINT32
> > 'acse.result' exists multiple times with NOT compatible types: FT_INT32 and FT_UINT32
> > 'ain.notificationIndicator' exists multiple times with NOT compatible types: FT_BYTES and FT_BOOLEAN
> > 'ain.invoke' exists multiple times with NOT compatible types: FT_INT32 and FT_UINT32
> > 'ansi_683.reserved' exists multiple times with NOT compatible types: FT_UINT16 and FT_BOOLEAN
> > 'ansi_tcap.private' exists multiple times with NOT compatible types: FT_UINT32 and FT_INT32

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <wireshark-dev () wireshark org>
Archives:    https://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev
             mailto:wireshark-dev-request () wireshark org?subject=unsubscribe