I am anticipating the rebase from hell.

I have had to fix warnings in some of my in-process patches just so I can make 
progress.  Meanwhile, some juggernaut (Jon?) is fixing them the same files and 
merging them.  Then I will need to rebase my patch, and there will be dozens or 
hundreds of trivial conflicts I get to resolve by hand.

John

-----Original Message-----
From: Lankswert, Patrick 
Sent: Tuesday, August 04, 2015 11:34 AM
To: Keane, Erich
Cc: iotivity-dev at lists.iotivity.org; jonc at osg.samsung.com; Light, John J
Subject: RE: [dev] warnings!

Erich,

The specificity that I wish that we could make is "Unused variable" warning as 
warning and "Missing return from non-void function" warning as an error. Have 
you looked at this level of selection?

Pat

> -----Original Message-----
> From: Keane, Erich
> Sent: Tuesday, August 04, 2015 2:30 PM
> To: Lankswert, Patrick
> Cc: iotivity-dev at lists.iotivity.org; jonc at osg.samsung.com; Light, John 
> J
> Subject: Re: [dev] warnings!
> 
> We've had a pretty good effort throughout both Samsung and Intel, so I 
> suspect that the warnings will be near-zero in the very near future.
> 
> It would be my preference to get this to zero ASAP, then set -Werror 
> to prevent it from happening again.  I realize that the "stop immediately"
> part is annoying, but it would be better to make sure people recognize 
> them happening asap, instead of ignoring them.
> 
> I'll note that our warning count (under old rules) was 0 for quite a 
> while, but slowly eeked up to ~20 about 2 weeks ago!
> 
> -Erich
> 
> On Tue, 2015-08-04 at 18:28 +0000, Lankswert, Patrick wrote:
> > John,
> >
> > I have mixed thoughts in this space. There are a number of 
> > contributions that
> the contributor clearly did not look at the warnings.
> > In addition, there may be cases where a warning may be ok.
> >
> > Pat
> >
> > > -----Original Message-----
> > > From: Light, John J
> > > Sent: Tuesday, August 04, 2015 2:04 PM
> > > To: Lankswert, Patrick; Jon A. Cruz; Keane, Erich
> > > Cc: iotivity-dev at lists.iotivity.org
> > > Subject: RE: [dev] warnings!
> > >
> > > Pat,
> > >
> > > The problem with instructing the compiler to treat errors as 
> > > warnings is that it stops the build.  I would prefer checking for 
> > > any warnings over the build only AFTER everything builds error free.
> > >
> > > Also, might it be better to enforce the higher warning level on 
> > > resource first, and when that gets cleaned up start on service.
> > >
> > > John
> > >
> > > -----Original Message-----
> > > From: Lankswert, Patrick
> > > Sent: Tuesday, August 04, 2015 8:12 AM
> > > To: Light, John J; Jon A. Cruz; Keane, Erich
> > > Cc: iotivity-dev at lists.iotivity.org
> > > Subject: RE: [dev] warnings!
> > >
> > > John,
> > >
> > > I think that Jenkins only looks at the compiler exit code. I do 
> > > not think that we have instructed the compiler to treat warnings as 
> > > errors... yet.
> > >
> > > Pat
> > >
> > > > -----Original Message-----
> > > > From: iotivity-dev-bounces at lists.iotivity.org
> > > > [mailto:iotivity-dev- bounces at lists.iotivity.org] On Behalf Of 
> > > > Light, John J
> > > > Sent: Tuesday, August 04, 2015 11:07 AM
> > > > To: Jon A. Cruz; Keane, Erich
> > > > Cc: iotivity-dev at lists.iotivity.org
> > > > Subject: Re: [dev] warnings!
> > > >
> > > > I wasn't complaining about the new warning level.  I was 
> > > > pointing out the irony that we hadn't eliminated the warnings 
> > > > that appeared BEFORE we changed the warning level.
> > > >
> > > > I'm surprised Jenkins doesn't complain about warnings.
> > > >
> > > > John
> > > >
> > > > -----Original Message-----
> > > > From: Jon A. Cruz [mailto:jonc at osg.samsung.com]
> > > > Sent: Monday, August 03, 2015 3:35 PM
> > > > To: Keane, Erich; Light, John J
> > > > Cc: iotivity-dev at lists.iotivity.org
> > > > Subject: Re: [dev] warnings!
> > > >
> > > > FYI, the increased warnings had already flagged at least one bug 
> > > > where network send errors were being missed due to a check that 
> > > > looked at a signed value being less than zero. Catching such a 
> > > > bug in code review that should have been flagged by the compiler 
> > > > was the reason the lack of -Wextra was noticed in the first place.
> > > >
> > > > First cleanups fixed Ubuntu 14.04 from 9.0k warnings down to 
> > > > 0.6k
> warnings.
> > > > Similar passes for 12.04 will follow, but as it stands builds 
> > > > were only seeing about 2k warnings.
> > > >
> > > > The velocity on warning cleanup should fairly quickly get to the 
> > > > point again where it is easy to spot significant issues as (or
> > > > before) they are
> > > introduced.
> > > >
> > > >
> > > > On 08/03/2015 03:00 PM, Keane, Erich wrote:
> > > > > As far as the 'sea of non-critical warnings', you aren't wrong.
> > > > > However, NOW is sorta the best time to do this, since we are 
> > > > > between releases, and it gives us as much time as possible to 
> > > > > fix them.  All patches to fix warnings are looked on quite 
> > > > > favorably :)
> > > > >
> > > > > -Erich
> > > > >
> > > > >
> > > > > On Mon, 2015-08-03 at 21:55 +0000, Light, John J wrote:
> > > > >> I?ve noticed a great increase in the number of warnings during build.
> > > > >> There have been more warnings in recently merged code, but 
> > > > >> this lastest increase seems to be the result of ratcheting up 
> > > > >> the warning threshold.
> > > > >>
> > > > >>
> > > > >>
> > > > >> I suspect we are well into the territory where critical 
> > > > >> warnings won?t be seen because they will be lost in a sea of 
> > > > >> non-critical
> warnings.
> > > > >>
> > > > >>
> > > > >>
> > > > >> Leaving that aside, I have a coding question in this new regime.
> > > > >>
> > > > >>
> > > > >>
> > > > >> A C file I am modifying but didn?t write has the following code:
> > > > >>
> > > > >>
> > > > >>
> > > > >>     OCPersistentStorage ps = {};
> > > > >>
> > > > >>     ps.open = client_fopen;
> > > > >>
> > > > >>     ps.read = fread;
> > > > >>
> > > > >>     ps.write = fwrite;
> > > > >>
> > > > >>     ps.close = fclose;
> > > > >>
> > > > >>     ps.unlink = unlink;
> > > > >>
> > > > >>
> > > > >>
> > > > >> This gets a warning about each line, like:
> > > > >>
> > > > >>
> > > > >>
> > > > >>     warning: missing initializer for member 
> > > > >> 'OCPersistentStorage::open'
> > > > >>
> > > > >>
> > > > >>
> > > > >> I can eliminate the warnings by coding it thus:
> > > > >>
> > > > >>
> > > > >>
> > > > >>     OCPersistentStorage ps = { client_fopen, fread, fwrite, 
> > > > >> fclose, unlink };
> > > > >>
> > > > >>     OCRegisterPersistentStorageHandler(&ps);
> > > > >>
> > > > >>
> > > > >>
> > > > >> But this seems more fragile since the ordering matters.
> > > > >>
> > > > >>
> > > > >>
> > > > >> Is there a C initialization method I?m missing?
> > > > >>
> > > > >>
> > > > >>
> > > > >> John
> > > > >>
> > > > >>
> > > > >> _______________________________________________
> > > > >> iotivity-dev mailing list
> > > > >> iotivity-dev at lists.iotivity.org 
> > > > >> https://lists.iotivity.org/mailman/listinfo/iotivity-dev
> > > > >
> > > > > _______________________________________________
> > > > > iotivity-dev mailing list
> > > > > iotivity-dev at lists.iotivity.org 
> > > > > https://lists.iotivity.org/mailman/listinfo/iotivity-dev
> > > > >
> > > >
> > > > --
> > > > Jon A. Cruz - Senior Open Source Developer Samsung Open Source 
> > > > Group jonc at osg.samsung.com
> > > _______________________________________________
> > > > iotivity-dev mailing list
> > > > iotivity-dev at lists.iotivity.org
> > > > https://lists.iotivity.org/mailman/listinfo/iotivity-dev

Reply via email to