Username length 16

amilojko said:
And exactly because it's trivial, it should be done right from the start.
Right is a subjective term. Your right is someone else's wrong.
  • You want something that the vast majority doesn't need.
  • Those who do need it have found ways to make it happen.
  • But you don't want to do anything for it yourself.
  • You can't even be arsed to recompile a kernel.
  • Instead, you insist that others do what you want and change things just because you want something.
  • And if they don't, you "threaten" to switch to another OS (feel free to do so).

I'm pretty much done with this silly thread. I have better things to do.

Fonz, out
 
fonz said:
Right is a subjective term. Your right is someone else's wrong.
  • You want something that the vast majority doesn't need.
  • Those who do need it have found ways to make it happen.
  • But you don't want to do anything for it yourself.
  • You can't even be arsed to recompile a kernel.
  • Instead, you insist that others do what you want and change things just because you want something.
  • And if they don't, you "threaten" to switch to another OS (feel free to do so).

I'm pretty much done with this silly thread. I have better things to do.

Fonz, out

Fonz, you are missing the point, which Crivens gets very well.

There are so many dumbasses, like me, who want to have things easy and don't want to do anything and yet have everything. Ok. Because there are OSes out there that will help you help your customers fast people will use them. Which means they won't use FreeBSD, which they should, but they won't because it takes too much time and skill to customize it. When FreeBSD was around, Ubuntu developers were just born, ok. Why do you think Windows is so popular. Because I'm lazy, I'm not that smart and I want to do as little as possible. And I didn't mean "right" as in "correct", I meant it as "from the beginning".
 
When and where has FreeBSD ever expressed the wish that it wants 'dumbasses' to use its OS, or that it wants to be 'popular'? You're entirely free (FreeBSD) to adjust FreeBSD to your wishes and submit a patch, or to move on to any OS out there that's fit for 'dumbasses' who are 'lazy' (not my words).

In short: nobody's going to do this for you, because no one feels the need to do it. Those who do feel the need to do it should do it and submit the changes to developers, who are free to reject them or send them back because the patch breaks twenty-five sub-systems.

Also note that the FreeBSD forums are aimed at end-users and admins, not so much at developers. There's only a hand-full of them active on the forums. The mailing lists would be a better place to address this issue. Be prepared to get roughly the same type of reply though. People with shopping lists and demands usually do.
 
@Crivens

This is only the start. Now you have to prove that you did not open any buffer-overflow holes in the base system
Very good point, I didn´t think about that.

So if you now have a system up and running with these new limits, please check if something weird happens when running it. If not, why not submit it and feel proud of having done it?
I know I would.
I´m in a mood for adventure, so I´m going to do this to test it out. What is this "wierd" that I´m supposed to look for? Could you come with suggestions on how to best monitor the system after the changes are made?

@Dutch

Those who do feel the need to do it should do it and submit the changes to developers
Could you please give me instructions as how to submit the changes in the "correct proper" fashion. Or point me to where this is explained?
I´m probably not going to start by just blindly submitting a patch, but rather first bring this up for discussion at the mailing lists, but if/when the time comes, it would be nice to know how.

-------------------------

After making these changes, recompiled and installed, I would like everyones ideas on stuff to test to see if the changes made had any negative ramifications that would prove that there is a real reason why NOT to have more than 16 characters, because it brakes the system.

I would like to test as much as possible, then bring it up at the lists saying "16 feels a little cramped nowadays due to longer login names. I´ve already tried recompiling and tested as much I could, and I couldn´t really find any negatives to this. How would you feel about making these changes permanent?" something or other.

I´m gonna start by installing SAMBA and joining to our domain, then try to log in via SSH and local login using a ridiculously long account name. I´m also gonna test adding a local account with a ridiculously long account name using adduser and remove it again with rmuser.

More suggestions, come on now! Let´s make sure to test this the "best" we possibly can!

/Sebulon
 
@ /Sebulon

I don't think you will have many issues with very long names. I think most programs don't have an issue with it. In this case the OS is the bottleneck. (MySQL has 16 character username limit actually too). This is not a new patch either. People have had to do this for a long time.
But it takes a while to do, an hour or more. Sometimes the rebuild with fail and then you are screwed. You have to spend hours and hours dealing with it.

It would be, in my opinion, easy and very pro-active for the FreeBSD team to make the change and then it would be there permanently. There is a whole list of things I would like to see in FreeBSD but like the DutchDeamon said I better submit something to them. Whining on a user forum is not going to do anything.

Like why does anyone build sendmail and ftpd without SSL support? Should never be done. (unless its a anonymous FTP). Those are little twinkie things that make admins' lives much easier and are not hard to do. But you have to do it for every server you install.
 
amilojko said:
I wouldn't say it's a "stupid" detail, I'd say it's a "trivial" detail, but it takes an hour to recompile the source to support longer usernames.
And exactly because it's trivial, it should be done right from the start.
This is one of the things that comes up every now and then, and ends up with the status quo. Changing this counts as an ABI change, so it would only happen in a new major release. Since 9.0 is in the process of releasing, that means that this change would have to wait for 10.0 (or 10-CURRENT if you want to run prerelease software).

In the other BSD's I looked at just now, NetBSD is 16, OpenBSD is 32, and DragonFly is 16.

I tilted at this particular windmill years ago (under BSD/OS and later FreeBSD) and during the time I needed it (at a previous job)*, it was one of a number of things that I changed in the base system. As I recall, there are a couple places which had it hardcoded at 16 - I think top might have been one of them. But that was over 10 years ago, and things have probably changed.

I have a number of limits I change in the base system (ARGMAX, among others). I accept the fact that I need to compile everything myself.

I don't think you're going to get anything useful asking for this change, both because as I mentioned above, you wouldn't see it until 10.0, and because this isn't the place to reach the developers and attempt to convince them of the need for this feature. And if you do want to post to the developer's mailing list, I'd suggest researching the reasoning behind OpenBSD's implementation of this change. You may find other things there that will help your proposal.

* I implemented a single-signon environment which had (among other things) a number of VMS systems, an IBM mainframe, Netware (yes, it was that long ago) and BSD/OS (and later FreeBSD). As the VMS systems were the senior members, everything else had to adapt to the usernames on that system, which were up to 16 characters in length (before FreeBSD 3.0, the limit was 12).
 
Sebulon said:
@Crivens
Very good point, I didn´t think about that.
Sorry, it comes with the job ;)
I´m in a mood for adventure, so I´m going to do this to test it out. What is this "wierd" that I´m supposed to look for? Could you come with suggestions on how to best monitor the system after the changes are made?
Oh, the usual things. Kernel panics, sudden crashes of programs like login, the box being rooted and spewing kiddie pr0n all over the place, nuclear junk starting to show up in your backjard ...
More suggestions, come on now! Let´s make sure to test this the "best" we possibly can!

/Sebulon
Maybe a tool like KLEE could be of help.
 
amilojko said:
Dutch! Please don't edit the formatting in my post. It is part of my expression. ;)
Your post was edited because it didn't comply to the forum rules, which you have been sent a copy of when you signed up. It's part of our policy to keep things tidy here. Those rules apply to everybody, even you.

Fonz
 
Sebulon said:
@Dutch

Could you please give me instructions as how to submit the changes in the "correct proper" fashion. Or point me to where this is explained?
I´m probably not going to start by just blindly submitting a patch, but rather first bring this up for discussion at the mailing lists, but if/when the time comes, it would be nice to know how.

The best starting point would be freebsd-questions. They should be able to point you to the relevant mailing list (possibly freebsd-hackers) or even directly to the developer(s) in charge of this piece of code.

If you have a clean bit of code (so a tidy, documented patch that doesn't need discussion, just testing), you could drop it using a PR.
 
Back
Top