Bug #1316

Some minor issues

Added by free almost 6 years ago. Updated over 5 years ago.

Status:ClosedStart date:
Priority:NormalDue date:
Assignee:-% Done:

0%

Category:-
Target version:-

Description

Hello, just want to write about some minor (maybe not?) issues that
I've noticed while using DragonFly v2.2.0 GUI Release (DVD version)
last week.
I don't know whether it would be better to post them separately, but
I decided to post them in one message.

0. I was installing DragonFly to the ad0s3 slice, and there were no such
device in the /dev folder, so the installer couldn't do its job until
I created the needed devices using MAKEDEV script.
In the end of the installation process, when I selected the packages
to install menu item (and there where all packages selected) and
accepted to install everything selected, I got some errors, but now I
I can't find them in the install.log file, though I'm attaching it.

1. "more" is not working well with "urxvt", unlike "less", "more" is
behaving like the -E option is provided even without providing the
-E option (the -E options tells it to quit when it reaches EOF).
Here is where the source: /usr/src/contrib/less-4/option.c [694]:
public int
get_quit_at_eof()
{
if (!less_is_more)
return quit_at_eof;
/* When less_is_more is set, the -e flag semantics are different.
*/
return quit_at_eof ? OPT_ON : OPT_ON_PLUS;
}
The thing is that unlike "xterm", the "urxvt" terminal emulator is
clearing everything printed by the "more" command after it exits, so,
for example, the "whatis" command won't work under "urxvt" if your
PAGER
is set to "more". For example, you type "whatis sh" and get nothing.
It's like "urxvt" is somehow differencing the output of "more"-like
commands, and keep it in another buffer, for example, when I type:
$ echo "test" | less
then type "q" to exit "less", the text "test" is disappearing, but when
I type another time:
$ echo "test2" | less
I can see the not only the latest "test2" text, but also the "test"
text above, until I press "q" to exit.

2. In /usr/src/contrib/less-4 the source files are +x (why?).

3. Sometimes the X is failing to start (I'm starting it from a user, not
root), but when I delete ~/.Xsession-l and ~/.Xsession-h 0-byte length
files, and then try again, it starts. I can't understand why, when and
how these two files are being created.

4. My system was used to boot (from the first day of installation till
today) not so fast, as this was my first experience with DragonFly and
BSD at all, I decided that this must be the way that it is. It was
"hanging" about a minute or maybe more after echo-ing the line where
the system is saying me that it's setting scrnmap, font8x16 and so on.
Then, after ~1 minute, the logon prompt was coming, then ~1 minute
hanging was occurring another time after typing the startx command.
I don't know what is the real reason that today my system started
booting
without that hangings (ya-hoo!), but the only change that I remember
is that I lowed down the count of ttys with "on" status in the
/etc/ttys file. Now I have only 4 ttys (ttyv0-ttyv3) instead of 8.
And also I converted them to insecure.

5. In the file /usr/share/skel/dot.shrc we can find interesting trick when
setting PS1: `hostname | sed 's/\..*//'`
Wo-hoo! Just use hostname -s
I'm not sure about sh, but in bash we can use just \h instead of that,
for
example my .bashrc file contains this line:
PS1="\u@\h:\w$ " instead of PS1="`whoami`@`hostname -s`:`pwd`$ "
The result is something like this: username@dflyhost:/usr/src$ _

6. There is a package in the mirrors named "free", while it's possible to
install it, it's not possible to use it, because it is working only
under
NetBSD and is calling us (DragonFly users) kids :) When I tried to
remove
the OS check from the script, the output was something not what I've
expected, so, I think it's working only under NetBSD and its package
should be removed from the mirrors or maybe better be fixed if somebody
can do that.

7. Some other minor things:

There is a typo in the NAME section of "man shlock" (scipt instead of
script).

There is a strange text in "man motd" at the end:
/home is full, please clean up your directories.

In the /etc/master.passwd the line containing root entry has name
"Charlie &". Who's Charlie? :)
And, there is a 2nd UID=0 user named toor, under name
"Bourne-again Superuser", it is a disabled to login user, but is it OK?

"man 5 auth.conf" needs some updates.

"man crypt" says that "bdes" and "pgp" are not in the distribution, but
"bdes" is there. Why pgp is isn't? ;) Oh, by the way, can I expect
porting to DragonFly some disk encryption technologies (there are two
available for FreeBSD currently)?

8. And, finally, just a question, I'm new to git and I have very limited
internet channel so I just can't do full git clone of DragonFly
sources,
but instead, I have the sources of the 2.2.0 version in my DVD
(src-all.tgz), I extracted it under my /usr directory, and now i would
like to synchronize the sources downloading just diffs using git, how
can I tell git to do that?

Thank you very much.

free_coder

History

#1 Updated by swildner almost 6 years ago

Separately would have been better since the different issues you
reported could be handled better in the bug tracker then (mails sent to
bugs@ automatically go into the tracker). Some answers follow.

This bug has been fixed in master, but unfortunately we haven't rolled
2.2.1 yet so it will continue to bite.

Matt, what's the plan for 2.2.1?

Overlooked when importing I guess.

Fixed in master!

Yes, it's an example motd (note the section heading "EXAMPLES"). I agree
that it is confusing, though, since it's an example we inherited from
FreeBSD, so I've "updated" it.

"Charlie Root" (the "&" gets replaced by the user name) seems the be the
traditional full name of the superuser in BSD derived OSs.

Yes, it's an alternative root account for those who prefer sh (Bourne
Shell) instead of csh.

Sascha

#2 Updated by swildner almost 6 years ago

I meant, "...has been fixed in master and the 2.2 branch..."

Sascha

#3 Updated by justin almost 6 years ago

On Tue, March 17, 2009 5:13 pm, free_coder wrote:

Yeah, the Makefile for the package should be adjusted to install only on
NetBSD, since a check inside the script bombs out without it actually
being on NetBSD. This is usually handled through Makefile checks, so you
can't build a useless program on a given platform.

It hasn't been actively maintained since 2005, probably before DragonFly
was even using pkgsrc. You can try submitting a note or patch to
.

Also available in: Atom PDF