Re: CVS commit: src/sys/coda

[christos%astron.com@localhost (Christos Zoulas)]

In article <20131123234556.1B4AC14A45E%mail.netbsd.org@localhost>,
Mindaugas Rasiukevicius  <rmind%netbsd.org@localhost> wrote:
>christos%zoulas.com@localhost (Christos Zoulas) wrote:
>> On Nov 23, 11:24pm, rmind%netbsd.org@localhost (Mindaugas Rasiukevicius) 
>> wrote:
>> -- Subject: Re: CVS commit: src/sys/coda
>> 
>> | mp = mp == TAILQ_END(&mountlist) ? NULL : mp;
>> | 
>> | Over:
>> | 
>> | return mp;
>> | 
>> | Everybody understood NULL, so why obfuscate the code?
>> 
>> The point (as I understand it) is so that if you want to change to another
>> ADT where end != NULL (are there any now, that we have banned CIRCLEQ?)
>> using the END macro lets you s/FOOQ/BARQ/ and have it mostly work.
>
>Since 1990s (or even earlier) none of the lists in queue(3) had _END(),
>apart from the circle queue (for a very obvious reason it makes sense).
>Now that CIRCLEQ is banned - you added _END() for LIST and TAILQ.  I do
>not follow the logic. :)

*I* did appear do be adding it from the commit message, but in reality
I was syncing with OpenBSD/FreeBSD.

>I would remove _END() macros to keep the way it always was.

Perhaps that's a good idea now that it is always NULL.

>If the goal
>is to improve the interface, then now is a good time to design a new API,
>but we already have a long thread on tech-userlevel for this..

Yes, so let's leave it the way it is now, and we'll make all the changes
in one go.

christos