Re: crunchgen and llvm 15
It seems some changes are being pushed without actually testing on the non-clang architectures. Can this get backed out until the correct procedure is followed? But I think it is also important to realize that not every new behaviour of clang is "systems ready". Some of what they are pushing aggressively is BS.
Re: [patch] arp.c spaces nested between tabs
On Thu, Dec 29, 2022 at 09:52:49AM +0100, Omar Polo wrote: > ksh actually supports that style of globbing :) Fair enough! TIL -- but the GNU Grep bit stands, though :) Any takers on the diff? Should be an easy one paultag -- :wq
Re: vmctl: use a space rather than tab in usage
On 2022-12-30 15:06 +01, David Demelier wrote:
> On Fri, 2022-12-30 at 14:50 +0100, Florian Obser wrote:
>> That seems reasonable. This might be the full list, do you want to do
>> all?
>>
>> usr.bin/htpasswd/htpasswd.c:fprintf(stderr, "usage:\t%s [file]
>> login\n", __progname);
>> usr.sbin/installboot/installboot.c: fprintf(stderr, "usage:\t%1$s
>> [-nv] [-r root] disk [stage1%2$s]\n"
>> usr.sbin/ldomctl/ldomctl.c: fprintf(stderr, "usage:\t%1$s
>> delete|select configuration\n"
>> usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] command [arg
>> ...]\n", __progname);
>> usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] %s %s\n",
>> __progname,
>>
>
> Thanks for the hint, there's the updated diff.
>
> Index: usr.bin/htpasswd/htpasswd.c
> ===
> RCS file: /cvs/src/usr.bin/htpasswd/htpasswd.c,v
> retrieving revision 1.18
> diff -u -p -p -u -r1.18 htpasswd.c
> --- usr.bin/htpasswd/htpasswd.c 12 Jul 2021 15:09:19 - 1.18
> +++ usr.bin/htpasswd/htpasswd.c 30 Dec 2022 14:01:38 -
> @@ -36,8 +36,8 @@ extern char *__progname;
> __dead void
> usage(void)
> {
> - fprintf(stderr, "usage:\t%s [file] login\n", __progname);
> - fprintf(stderr, "\t%s -I [file]\n", __progname);
> + fprintf(stderr, "usage: %s [file] login\n", __progname);
> + fprintf(stderr, " %s -I [file]\n", __progname);
Sorry, I should have looked close what's going on. I don't think this is
an improvement. Your original diff was better. Sorry for sending you on
a wild goose chase.
The following chunks are OK florian
> Index: usr.sbin/vmctl/main.c
> ===
> RCS file: /cvs/src/usr.sbin/vmctl/main.c,v
> retrieving revision 1.73
> diff -u -p -p -u -r1.73 main.c
> --- usr.sbin/vmctl/main.c 1 Sep 2022 15:43:07 - 1.73
> +++ usr.sbin/vmctl/main.c 30 Dec 2022 14:01:40 -
> @@ -96,7 +96,7 @@ usage(void)
> {
> extern char *__progname;
>
> - fprintf(stderr, "usage:\t%s [-v] command [arg ...]\n",
> __progname);
> + fprintf(stderr, "usage: %s [-v] command [arg ...]\n",
> __progname);
>
> exit(1);
> }
> @@ -106,7 +106,7 @@ ctl_usage(struct ctl_command *ctl)
> {
> extern char *__progname;
>
> - fprintf(stderr, "usage:\t%s [-v] %s %s\n", __progname,
> + fprintf(stderr, "usage: %s [-v] %s %s\n", __progname,
> ctl->name, ctl->usage);
> exit(1);
> }
>
--
I'm not entirely sure you are real.
Re: vmctl: use a space rather than tab in usage
On Fri, 2022-12-30 at 14:50 +0100, Florian Obser wrote:
> That seems reasonable. This might be the full list, do you want to do
> all?
>
> usr.bin/htpasswd/htpasswd.c:fprintf(stderr, "usage:\t%s [file]
> login\n", __progname);
> usr.sbin/installboot/installboot.c: fprintf(stderr, "usage:\t%1$s
> [-nv] [-r root] disk [stage1%2$s]\n"
> usr.sbin/ldomctl/ldomctl.c: fprintf(stderr, "usage:\t%1$s
> delete|select configuration\n"
> usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] command [arg
> ...]\n", __progname);
> usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] %s %s\n",
> __progname,
>
Thanks for the hint, there's the updated diff.
Index: usr.bin/htpasswd/htpasswd.c
===
RCS file: /cvs/src/usr.bin/htpasswd/htpasswd.c,v
retrieving revision 1.18
diff -u -p -p -u -r1.18 htpasswd.c
--- usr.bin/htpasswd/htpasswd.c 12 Jul 2021 15:09:19 - 1.18
+++ usr.bin/htpasswd/htpasswd.c 30 Dec 2022 14:01:38 -
@@ -36,8 +36,8 @@ extern char *__progname;
__dead void
usage(void)
{
- fprintf(stderr, "usage:\t%s [file] login\n", __progname);
- fprintf(stderr, "\t%s -I [file]\n", __progname);
+ fprintf(stderr, "usage: %s [file] login\n", __progname);
+ fprintf(stderr, " %s -I [file]\n", __progname);
exit(1);
}
Index: usr.sbin/installboot/installboot.c
===
RCS file: /cvs/src/usr.sbin/installboot/installboot.c,v
retrieving revision 1.16
diff -u -p -p -u -r1.16 installboot.c
--- usr.sbin/installboot/installboot.c 8 Nov 2022 12:08:53 -
1.16
+++ usr.sbin/installboot/installboot.c 30 Dec 2022 14:01:39 -
@@ -38,8 +38,8 @@ char *stage2;
static __dead void
usage(void)
{
- fprintf(stderr, "usage:\t%1$s [-nv] [-r root] disk
[stage1%2$s]\n"
- "\t%1$s [-nv] -p disk\n",
+ fprintf(stderr, "usage: %1$s [-nv] [-r root] disk
[stage1%2$s]\n"
+ " %1$s [-nv] -p disk\n",
getprogname(), (stages >= 2) ? " [stage2]" : "");
exit(1);
Index: usr.sbin/ldomctl/ldomctl.c
===
RCS file: /cvs/src/usr.sbin/ldomctl/ldomctl.c,v
retrieving revision 1.40
diff -u -p -p -u -r1.40 ldomctl.c
--- usr.sbin/ldomctl/ldomctl.c 24 Oct 2021 21:24:18 - 1.40
+++ usr.sbin/ldomctl/ldomctl.c 30 Dec 2022 14:01:39 -
@@ -125,13 +125,13 @@ main(int argc, char **argv)
__dead void
usage(void)
{
- fprintf(stderr, "usage:\t%1$s delete|select configuration\n"
- "\t%1$s download directory\n"
- "\t%1$s dump|list|list-io\n"
- "\t%1$s init-system [-n] file\n"
- "\t%1$s create-vdisk -s size file\n"
- "\t%1$s panic|start [-c] domain\n"
- "\t%1$s console|status|stop [domain]\n",
+ fprintf(stderr, "usage: %1$s delete|select configuration\n"
+ " %1$s download directory\n"
+ " %1$s dump|list|list-io\n"
+ " %1$s init-system [-n] file\n"
+ " %1$s create-vdisk -s size file\n"
+ " %1$s panic|start [-c] domain\n"
+ " %1$s console|status|stop [domain]\n",
getprogname());
exit(EXIT_FAILURE);
Index: usr.sbin/vmctl/main.c
===
RCS file: /cvs/src/usr.sbin/vmctl/main.c,v
retrieving revision 1.73
diff -u -p -p -u -r1.73 main.c
--- usr.sbin/vmctl/main.c 1 Sep 2022 15:43:07 - 1.73
+++ usr.sbin/vmctl/main.c 30 Dec 2022 14:01:40 -
@@ -96,7 +96,7 @@ usage(void)
{
extern char *__progname;
- fprintf(stderr, "usage:\t%s [-v] command [arg ...]\n",
__progname);
+ fprintf(stderr, "usage: %s [-v] command [arg ...]\n",
__progname);
exit(1);
}
@@ -106,7 +106,7 @@ ctl_usage(struct ctl_command *ctl)
{
extern char *__progname;
- fprintf(stderr, "usage:\t%s [-v] %s %s\n", __progname,
+ fprintf(stderr, "usage: %s [-v] %s %s\n", __progname,
ctl->name, ctl->usage);
exit(1);
}
Re: vmctl: use a space rather than tab in usage
That seems reasonable. This might be the full list, do you want to do
all?
usr.bin/htpasswd/htpasswd.c:fprintf(stderr, "usage:\t%s [file] login\n",
__progname);
usr.sbin/installboot/installboot.c: fprintf(stderr, "usage:\t%1$s [-nv] [-r
root] disk [stage1%2$s]\n"
usr.sbin/ldomctl/ldomctl.c: fprintf(stderr, "usage:\t%1$s delete|select
configuration\n"
usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] command [arg ...]\n",
__progname);
usr.sbin/vmctl/main.c: fprintf(stderr, "usage:\t%s [-v] %s %s\n", __progname,
On 2022-12-30 14:45 +01, David Demelier wrote:
> Most utilities seem to use a unique space between the colon and the
> program name, vmctl uses a tab that will expand to two spaces once
> invoked.
>
> e.g.
>
> pfctl: unknown command line argument: tata ...
> usage: pfctl [-deghNnPqrvz] [-a anchor] [-D macro=value] [-F modifier]
> [-f file]
>
> tar foobar
> tar: f argument missing
> usage: tar {crtux}[014578befHhjLmNOoPpqsvwXZz]
>
> vmctl oops
> unknown argument: oops
> usage:vmctl [-v] command [arg ...]
>
> The style(9) manual page seems to recommend a unique space though:
>
> The getprogname(3) function may be used instead of hard-coding the
> program name.
>
>fprintf(stderr, "usage: %s [-ab]\n", getprogname());
>exit(1);
>
> Index: main.c
> ===
> RCS file: /cvs/src/usr.sbin/vmctl/main.c,v
> retrieving revision 1.73
> diff -u -p -u -p -r1.73 main.c
> --- main.c1 Sep 2022 15:43:07 - 1.73
> +++ main.c30 Dec 2022 13:39:37 -
> @@ -96,7 +96,7 @@ usage(void)
> {
> extern char *__progname;
>
> - fprintf(stderr, "usage:\t%s [-v] command [arg ...]\n",
> __progname);
> + fprintf(stderr, "usage: %s [-v] command [arg ...]\n",
> __progname);
>
> exit(1);
> }
> @@ -106,7 +106,7 @@ ctl_usage(struct ctl_command *ctl)
> {
> extern char *__progname;
>
> - fprintf(stderr, "usage:\t%s [-v] %s %s\n", __progname,
> + fprintf(stderr, "usage: %s [-v] %s %s\n", __progname,
> ctl->name, ctl->usage);
> exit(1);
> }
>
--
I'm not entirely sure you are real.
vmctl: use a space rather than tab in usage
Most utilities seem to use a unique space between the colon and the
program name, vmctl uses a tab that will expand to two spaces once
invoked.
e.g.
pfctl: unknown command line argument: tata ...
usage: pfctl [-deghNnPqrvz] [-a anchor] [-D macro=value] [-F modifier]
[-f file]
tar foobar
tar: f argument missing
usage: tar {crtux}[014578befHhjLmNOoPpqsvwXZz]
vmctl oops
unknown argument: oops
usage: vmctl [-v] command [arg ...]
The style(9) manual page seems to recommend a unique space though:
The getprogname(3) function may be used instead of hard-coding the
program name.
fprintf(stderr, "usage: %s [-ab]\n", getprogname());
exit(1);
Index: main.c
===
RCS file: /cvs/src/usr.sbin/vmctl/main.c,v
retrieving revision 1.73
diff -u -p -u -p -r1.73 main.c
--- main.c 1 Sep 2022 15:43:07 - 1.73
+++ main.c 30 Dec 2022 13:39:37 -
@@ -96,7 +96,7 @@ usage(void)
{
extern char *__progname;
- fprintf(stderr, "usage:\t%s [-v] command [arg ...]\n",
__progname);
+ fprintf(stderr, "usage: %s [-v] command [arg ...]\n",
__progname);
exit(1);
}
@@ -106,7 +106,7 @@ ctl_usage(struct ctl_command *ctl)
{
extern char *__progname;
- fprintf(stderr, "usage:\t%s [-v] %s %s\n", __progname,
+ fprintf(stderr, "usage: %s [-v] %s %s\n", __progname,
ctl->name, ctl->usage);
exit(1);
}
Re: libcbor v0.10.0
On 2022/12/30 02:06, Theo Buehler wrote: > I understand that it is a libcbor major bump. Why is the libfido2 bump > needed? We can run into a problem with bumps with inter-library dependencies in base. If software from packages uses functions from both libraries, if you _don't_ bump major for libfido2, it will pull in the new fido2 lib (using new cbor abi), and depending on what ld.so decides, one or other version of libcbor, that can't satisfy abi requirements of bith new libfido2 and old binary. It resolves itself when new packages are installed, but that takes time. Now, I don't know if that actually happens with packages using fido2/cbor; could be that they only want libfido2 functions and don't use libcbor functions themselves, in which case there's no problem. It happens very often when X libraries are bumped, but they're used by more packages and there's higher likelihood of functions from both libraries actually being called (or structs being passed around and potentially copied incomplete, etc).
