Date: Wed, 13 Dec 2000 14:22:06 -0700 From: "Justin T. Gibbs" <gibbs@scsiguy.com> To: Mike Smith <msmith@FreeBSD.org> Cc: cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/conf files options src/sys/dev/pci pci_user.c eisa_pci.c pci.c pci_pci.c pcireg.h pcivar.h src/sys/pci pci_compat.c Message-ID: <200012132122.eBDLM6s33677@aslan.scsiguy.com> In-Reply-To: Your message of "Wed, 13 Dec 2000 13:05:16 PST." <200012132105.eBDL5G309996@mass.osd.bsdi.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>> > - Add prototypes and re-layout the core PCI modules in line with
>> > current coding standards (not a major whitespace change, just moving
>> > the module data to the top of the file).
>>
>> Don't you mean, "reformat to my personal code style"?
>
>Actually, no, I meant that functions should be prototyped, and that the
>module declarations and method tables should be at the top of the file,
>using the prototypes rather than the actual implementations of the
>functions.
Yes, I had prototyped the entire file as well. Moving the method
table wasn't necessary in my opinion, but regardless, that was not
the source of my complaint.
>However, since I'm in the process of effectively rewriting all of this
>code, yes, it's probably going to end up in my own personal variation of
>style(9). I started off trying to maintain the original 8-space indents
>for new code (which was a mistake, since it now means some substantial
>whitespace diffs against work I've already committed), but once it became
>clear how much needed to be reworked I decided I might as well do it in a
>format I'm comfortable working with.
You are not the sole maintainer of this code, nor are you the only
one working in it. Why should you get your style instead of some
other developer getting their style? To avoid these kinds of pissing
wars, we have adopted a single "preferred style".
>> This is not about your style being better than another, but that we
>> have a mandated style for a reason. You've shown in the past the ability
>> to follow our coding standard, so why are you bucking it now?
>
>I'm not; I'm just applying the commonsense interpretation of the style
>guidelines.
The first sentence of the style guide is this:
DESCRIPTION
This file specifies the preferred style for kernel source files in the
FreeBSD source tree.
You have taken code that was in the "preferred style" for the kernel,
and modified it into something other than that style. I don't know
how you can interpret it otherwise.
>If I was just tinkering with a few lines of this code, I'd
>stick with the way it was originally formatted. Since, however, I'm in
>the process of almost entirely rewriting it, and since I expect to be
>maintaining it for quite some time to come, it makes a lot of sense to me
>to do it in a style that isn't going to inhibit my work.
And as I've already noted, this does inhibit the work of others. The changes to
make PCI a parent class of cardbus are no less trivial than the work
you are doing. In fact, the power state changes you made are very
similar to mine save the introduction of suspend and resume methods
at the PCI level. I've modified everything from how the probe is performed
to how resources are allocated so that code can be shared. Again, you
are not, nor will you likely be, the sole person interested, capable,
or willing to work on this code. This is a colaborative project which
is why we have chosen a common style.
>In retrospect, I should probably have waved big red flags at you when you
>mentioned you were working on the cardbus rationalisation, however it
>wasn't clear at that point that what I wanted to do with PCI was going to
>necessitate such vigorous change. However one of the objectives of this
>work is to throw away most, if not all of the "cardbus bus" and just
>attach PCI to the cardbus bridge. I'm sorry for the inconvenience this
>has inadvertently caused you.
Its not inconvenience, but gratuitous change and indifference to
what it takes to work collaboratively that bothers me. BTW, I'm
90% of the way to making PCI plus an extrememly thin cardbus sub-class
attach to pccbb. I'd hate to see you repeat this work.
>On the other hand, I'll note that so far all the technical review I've
>recived has been criticism of my choice of 4-space indents over tabs.
>I find this pretty miserable; the code works just as well (and is just as
>readable) either way.
I will bring up my technical dissagreements with the code on -arch
(and yes, I do have some). How we manage our tree as a group is a
more general issue that needed to be addressed to -committers. It
really bothers me that a member of Core is pushing an agenda that
will make our tree look more and more like Linux. I spend too much
of my work day dealing with that "standardless" code base as it is.
I don't want FreeBSD to go the same way.
--
Justin
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200012132122.eBDLM6s33677>