[not currently subscribed, please Cc, thanks.]
Yes, despite the reservation in the comment above the code with the
tyop, the offending condition is indeed triggered on this
particularly cheap craptop. Figuring that out is a separate
matter...
--zeurkous.
Index: src/sys/dev/pci/if_jme.c
zeur here.
[sent this about a week ago, but it never seems to have arrived...
re-trying.]
In case anyone's still interested, here's the X patch.
Disclaimer: I'm unfamiliar with either syscons or pcvt, but those cases
seemed to contain OpenBSD-related code so I patched them, too. Apologies
if I
Alright, these ones get rid of is_stdin altogether. Isn't it better this
way?
--zeurkous.
Index: src/usr.bin/tail/extern.h
===
RCS file: /cvs/src/usr.bin/tail/extern.h,v
retrieving revision 1.12
diff -u -p -r1.12 extern.h
---
zeur here.
mewrote:
> You don't need my permission, but ok zeurkous =)
On 2nd thought, do we really need is_stdin then, though? Isn't a test
like 'tf[i].fp == stdin' enough?
--zeur.
--
Friggin' Machines!
zeur here.
"Martijn van Duren" wrote:
> I'd like to propose a slightly different diff, since POSIX states that a
> filename "-" should mean stdin
Argh, I really hate that convention... but I suppose POSIX is always
right :)
> I would like to prepare the struct
> tailfile to contain the is_stdin
zeur here.
The following {,ab}use of tail(1):
% fname);
Haai,
mark.kette...@xs4all.nl wrote:
> Sorry,
No need to apologize for sound technical decisions.
> but I really don't see the point.
That is your loss. Not that you need to see the point for *my* benefit,
'cause these patches are running like a charm here :)
> If you really need that many
> s
Haai,
> 3 character tty names
4-character actually (what is it w/ people dropping the dash?).
> isn't going to be commited.
*shrugs* That's up to you folks, of course...
It sure has been 'commited' to my local tree, and will remain so for the
foreseeable future :)
> You may as well give it up
After w(1), another casualty of my wscons patches: ps(1) gives only the
first two chars of the tty name suffix in the 'TT' field, leading to
e.g. 'C128' appearing as 'C1'. This is obviously misleading.
For just such a case, the patches below fix this by broadening said
field by 2 characters.
See below. Bit hasty, weren't we? :P
--zeur.
Index: xenocara/xserver/hw/xfree86/os-support/bsd/bsd_init.c
===
RCS file: /cvs/xenocara/xserver/hw/xfree86/os-support/bsd/bsd_init.c,v
retrieving revision 1.21
diff -u -p -r1.21 bs
rc(8) does not pass its arguments to rc.local(8); this makes the latter
unable to check for the 'autoboot' or 'shutdown' args.
Patch below fixes this by simply passing the arguments.
--zeurkous.
Index: src/etc/rc
===
RCS fil
The following completely plausible:
keyboard1.map+=keycode 101=HiFn
in the wsconsctl.conf(5) file is not passed to wsconsctl(8) intact,
leading to syntax errors being reported by the latter (and the requested
action not being taken).
The below patch fixes this by replacing the apparent UUOE wit
I noticed I broke w(1) with my previous patch set: the 'TTY' column is
no longer large enough. The below patch increases it in size by two
characters, an also gives the 'USER' field an extra character, fixing an
apparently existing wrong offset for the 'TTY' one.
--zeurkous.
Index: src/u
These final patches add a second set of Cmd keysyms, that for now have
one effect: add 128 to the screen index when a screen switch is
commanded.
Applies on top of the previous patches. Together with HiFn, this allows
addressing of 48 emulated terminals using the three 'windoze' keys on a
'modern'
These patches add a 'HiFn' keysym, useful in cases where F{13..24} input
is desired and no suitable kbd is present.
If, in opposition to some of the bolder statements below, it turns out
to be trivial to fix the Cmd&&HiFn issue, I'll gladly eat the words
concerned :)
Oh, and the reason I put the
These patches seperate the screen command keysyms into their own group,
so that the full address range (minus that pesky 255) can be used, in
conjunction with the MAKEDEV and ttys patches.
Applies on top of the small wskbd patch, and the factual wsdisplay(4)
patch.
This affects the kernel and wsc
This patch somewhat generalizes the screen cmd keysym detection. It does
depend on the relevant keysyms being contigious, though, but that
shouldn't be a problem.
Index: src/sys/dev/wscons/wskbd.c
===
RCS file: /cvs/src/sys/dev/wscons
Manual patch to go with the MAKEDEV and ttys patches. Applies on top of
the factual patch.
--- src/share/man/man4/wsdisplay.4.orig Mon Dec 24 06:25:42 2018
+++ src/share/man/man4/wsdisplay.4 Mon Dec 24 06:27:28 2018
@@ -27,7 +27,9 @@
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF TH
Irrespective of the other patches, I believe this one should certainly
be applied, as it corrects a factual error.
Index: src/share/man/man4/wsdisplay.4
===
RCS file: /cvs/src/share/man/man4/wsdisplay.4,v
retrieving revision 1.53
diff
These patches fix up ttys(5) after the MAKEDEV change.
Shouldn't this be moved into a seperate file, as per ttys.pty?
Index: src/etc/etc.amd64/ttys
===
RCS file: /cvs/src/etc/etc.amd64/ttys,v
retrieving revision 1.2
diff -u -p -r1.2
This patch changes the names for emulated ttys from tty[C-J][0-9a-f] to
tty[C-J]{0..254}. 255 is reserved for the cfg device and thus cannot
currently be used.
This enables consistent addressing for all emulated ttys, and is
not unexpected to hit bits here and there; it is also a visible change,
a
This patch enables the specification of WSDISPLAY_MAXSCREEN as an option
in a config(8) file. Without, a multiple-definition error is thrown.
This patch is expected to affect only the kernel, and then only at
config(8) time.
Index: src/sys/dev/wscons/wsdisplayvar.h
===
22 matches
Mail list logo