kern/50827: [kernel] [patch] new feature: add sane record locking

Stephen Hurd shurd at sasktel.net
Thu Oct 27 02:40:04 PDT 2005


Mark Linimon wrote:

>On Wed, Oct 26, 2005 at 09:52:56PM -0600, Stephen Hurd wrote:
>  
>
>>I've already gotten over my bitterness on the point, so if you want to 
>>mark it closed or whatever, that's fine... I don't plan on pursuing it 
>>any further as feature advocacy isn't a hobby of mine.
>>    
>>
>
>As I am emphasizing to people as I try to try to reclassify and/or refile
>some of these really old PRs, FreeBSD simply does not have the volunteer
>power to get every single suggested change into the system.  There are
>currently exactly 4,950 PRs extant and only a few hundred developers,
>all but a tiny handful of whom are volunteers.  Even if each of them
>were to devote their full time to FreeBSD, there still would not be enough
>time to meet this demand.
>  
>
I understand completely.

>It's a volunteer project.  Persistence (in some cases dogged persistence)
>is the only way changes get incorporated.  It can take many months and
>many high-quality PRs being sent in before some existing committer starts
>noticing them and wondering if it's time for that person to get a mentor
>and a commit bit.
>  
>
Yeah, that was the source of my bitterness originally.  When I first 
whipped up the patch, it was to fill a very specific need which I felt 
(and still do actually) many others would be happy to have a solution 
even if it was FreeBSD specific.  It seemed as though there was a Not 
Invented Here issue which was simply the result of there being no simple 
way to discover who the "right" person to talk to about the patch was.  
I was pretty blind to the workload of the volunteers and expected 
*something* even it it was just a "that's cool, but we can't include it 
for <reason>" or a "I notice you didn't update the manpage... we can't 
very well add undocumented new features" stuff like that.  
Unfortunately, I have minimal time to spend hacking on FreeBSD 
(actually, that's a good part of the reason I use FreeBSD... I don't 
*need* to spend time hacking on it) and I'll never have the persistence 
- dogged or otherwise - to get small patches like this one added.

It was actually because of this patch that I never even bothered trying 
to submit a largeish patch to doscmd... I made a half-hearted attempt to 
find someone who was "in charge" of doscmd to go over the changes (a 
single post to -emulation iirc) then just dropped the whole thing when 
there was no response.

I'll never submit many PRs - high-quality or otherwise, so for a while I 
was in my little pointless "nobody cares about ME" snit.  As I said, I 
came to realize that my expectations were unrealistic and that if I'm 
not going to take the time to become part of the community, I can't 
expect the community to take time to deal with me unless I'm heading in 
the same direction someone else is.  I've since restricted my PRs to bug 
reports... following up with anyone who replies, but just dropping the 
subject if nobody bites.  My recent natd PR is an example of that.  It'd 
take me about 20 minutes to whip up a patch to fix the issue (and I'm 
sure it would take others less time) but I no longer expect a response.  
As a result, I'm pleased when I get one, but I'm less likely to fix 
something... I'll just figure out a workaround, submit a PR, and add it 
to my notes.

I'm easing back into a collaborative frame of mind though... I just 
submitted a new port, and took maintainership of another (and the 
response on the update/maintainership was fast and satisfying).  I'm 
still not sure if I'll ever be at a point where I'll feel I should 
submit a patch against a core bit of the OS again, but I'll probably 
whip up that natd patch if nobody has done it by the time I find myself 
with a bit of spare time... I may even dig out that old doscmd patch and 
take 'er for a whirl.

I think what my opinion developed to is that unless you're willing to 
spend a large amount of your free time hacking FreeBSD, there's no point 
in spending a little bit of time hacking it.  In essence, if you spend a 
small amount of time, that time will have been wasted unless you're 
willing to spend a large amount of time hacking other bits to get the 
first bit noticed, reading and raising a stink in the mailing lists, 
generally pushing your hacks in the face of anyone who happens to walk by.

In re-reading this, I seem to come off as still bitter and the text 
feels like an accusation or at least a complaint... but it's not.  I 
have a vague feeling that something isn't as perfect as it could be, but 
I have no clue as to what that something would be, and no reasonable 
suggestion as far as what to do about it.  We live in an imperfect 
world... and while FreeBSD may not have reasonable record locking, 
neither does anyone else in the *nix world and aside from that one issue 
(which I happen to have a patch for :-) I've not had any serious 
complaint about FreeBSD (unlike almost every other UNIX, BSD, and Linux 
distro I've had to use).



More information about the freebsd-bugs mailing list