On 09/18/2012 01:44 PM, Clint Byrum wrote:
> Excerpts from Alan Robertson's message of 2012-09-17 12:23:29 -0700:
>> On 09/17/2012 01:05 PM, Dejan Muhamedagic wrote:
>>> Hi Alan,
>>>
>>> On Mon, Sep 17, 2012 at 10:28:52AM -0600, Alan Robertson wrote:
>>>> I still need to get it working for RHEL 6.2 - and for that matter,
>>>> probably on RHEL5. But feel free to try it on newer OSes. There is
>>>> another known issue with these older versions of glib - and I know we
>>>> need to get it going.
>>>>
>>>> But, if you add the -Werror flag, you'll run into one other issue, which
>>>> I also know the workaround for.
>>> Did you consider using the opensuse build service? It supports a
>>> number of different platforms and it has been working quite well
>>> for some time. https://build.opensuse.org/
>>>
>> I did think about it. The last time I looked at it I couldn't make
>> sense of how to use it. Do you want to tackle that for us?
>>
>> I think that makes more sense than trying to replicate all the platforms
>> it compiles for.
>>
> I have 0 experience with suse's build service, so thats fine but it will
> take me a lot longer to get it up than a simple jenkins instance, which I
> had up and ready to build debs before I hit send on my last email.
>
> Will poke at it a bit now though.
Let Dejan and me know what you need from us. Dejan's a great guy.
But having someone else supply all the horsepower and the build
environments hopefully will work out - given that the nanoprobes ideally
would run everywhere. I think the build service will also cover some of
the BSDs too... We'll have to do something different for Windows...
I'll find my patches, formalize them, and maybe even eliminate the need
for one of the functions that's "new" compared to RHEL 6.1 and friends...
But we will still need to turn off some of the flags on some platforms
because of use of libraries that are broken in those particular (older)
versions. (grumble...)
Thanks much for diving into this!
>> One of the unfortunate aspects of this project, is that the nanoprobes
>> will likely need to run on far more platforms than would make sense for
>> most projects. Experience tells me that my insistence on giving as
>> aggressive a set of warning flags as I can combined with -Werror will be
>> the source of many of our compile difficulties.
>>
>> Nevertheless, I still like it as a policy. I think that initial pain in
>> getting things right pays off - despite the initial pain.
>>
> Most good projects these days are running their daily CI builds on
> the newest compilers with all warnings as errors, as it improves code
> quality. I don't think it will help with portability between OS's that
> much, but it won't hurt. It should help with things like porting to
> non-x86 arches, which always seems to cause issues.
I did that with Linux-HA - from the beginning (that is - back last
millenium) and I liked it. Not that I'm opinionated ;-)
--
Alan Robertson <[email protected]> - @OSSAlanR
"Openness is the foundation and preservative of friendship... Let me claim
from you at all times your undisguised opinions." - William Wilberforce
_______________________________________________
Assimilation mailing list - Discovery-Driven Monitoring
[email protected]
http://lists.community.tummy.com/cgi-bin/mailman/listinfo/assimilation
http://assimmon.org/
