svn commit: r314654 - in head/cddl: lib/drti lib/libavl lib/libctf lib/libdtrace lib/libnvpair lib/libumem lib/libuutil lib/libzfs lib/libzfs_core lib/libzpool sbin/zfs sbin/zpool usr.bin/ctfconver...

[Bryan Drewery]

On 3/4/2017 1:04 PM, Rodney W. Grimes wrote:
>>
>>> On Mar 4, 2017, at 12:46, Ngie Cooper (yaneurabeya) <yaneurabeya at gmail.com> wrote:
>>>
>>>
>>>> On Mar 4, 2017, at 12:42, Ngie Cooper (yaneurabeya) <yaneurabeya at gmail.com> wrote:
>>>
>>> ...
>>>> I care about this for the reasons brewery posted, but I also care because it slows down my terminal output and it bloats my disk with typescript logs that contain unnecessary information.
>>>
>>> And it?s not just my personal laptop. It bloats log files in CI services, slows down grep?ing logfiles, increases data retention requirements, etc.
>>>
>>> We (at $work) do thousands of builds a week. Saving a few MB here and there can go a long way over the course of weeks, months, etc.
>>
>> Sidenote: it really peeves me how viewing a log in Jenkins will crash my browser when it tries to render the text. Maybe now it will work a bit better.
> 
> That is a very bad reason, if this disk space saving and not having
> to see all the ../.. scroll by on your terminal as make buildworld
> runs is your motivation for this I would say we need to stop
> now and have a more reasonable and serious discussion about why
> this is actually a good idea.
> 
> Please go fix the bug in your browser and hold off on these
> changes (in either direction) until a proper discussion, possibly
> real tests to see if there is any savings in cycles worthy
> of this sized change.
> 
>> -Ngie
> 

Your reason is "a very bad reason" being "my gutt is telling me this
change is Bad(tm)." vs the objective list of reasons SRCTOP is better
than '../..'. (quotes from you).

Changing to SRCTOP is not about speed or saving grand or trivial amounts
of space.  It's about making Makefiles simpler to write and understand
and making the output simpler to read and use.

I don't think it is worth the effort to discuss every little thing like
this.  There has been a discussion among people who maintain the build
and all agreed with it.

I was telling my wife just last night that if I had put up every little
change I did for review/discussion in the build over last year that I
likely would be delayed by years due to bikeshedding trivial aspects of
them, and would have just given up.  I still have a lot left to do as
well.  Among all of the changes, the only real dissent I've had in the
end is from one person saying that some of the work slows down the
kernel build by a few seconds when building on a small system over a
slow NFS link.  That's a valid complaint though vs this thread which is
starting to feel like a bikeshed.

-- 
Regards,
Bryan Drewery

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 473 bytes
Desc: OpenPGP digital signature
URL: <http://lists.freebsd.org/pipermail/svn-src-head/attachments/20170304/59b2f150/attachment.sig>