Re: CVS commit: src/usr.sbin/wake
On Jan 12, 4:26am, dholland-sourcechan...@netbsd.org (David Holland) wrote: -- Subject: Re: CVS commit: src/usr.sbin/wake | On Mon, Jan 11, 2010 at 06:31:29PM -0500, Christos Zoulas wrote: | | In the long run (as I've said before) we should replace our lint with | | something more modern... | | Once we get the llvm tools, I am fine with that. | | There's also splint, or sparse. Splint: License: - GPLv2: Cons:- does not understand gcc constructs - wierd flags and annotations - no c99 support - does not seem to be really supported: 3.1.2 release 2007-07-12 3.1.0 release 2003-04-21 Pros:- seems to be comprehensive and does many more tests than our lint Sparse: License: - Some weird and extra long transmeta license Cons:- Seems to do less than our lint, too simplistic - no c99 support - no annotations Pros:- Seems to understand gcc, is actively maintained. So splint looks like it has some value added, but it will need *a lot* of work to be integrated. Unless I am missing something, neither is appropriate in its current state to replace lint (without a few man-months of work), and I don't think that anybody will step up to do it. Comments? christos
Re: CVS commit: src/usr.sbin/wake
On Sun, Jan 10, 2010 at 11:23:19PM +, Christos Zoulas wrote: Please explain to me how you propose we fix lint. Realistically, lint should be made to know a list of standard functions with usually-uninteresting return values. That would solve 95+% of the practical problem. In the long run (as I've said before) we should replace our lint with something more modern... -- David A. Holland dholl...@netbsd.org
Re: CVS commit: src/usr.sbin/wake
chris...@astron.com (Christos Zoulas) wrote: Log Message: remove stupid casts. if lint complains about return values that are ignored, lint should fe fixed, and not code being cluttered. Please explain to me how you propose we fix lint. Or to be more precise, how does the code convey intent about purposefully ignoring the return value of a function so that lint can ignore it. FYI this change is against the current style guide, so if you are going to be pushing for more changes like that, I suggest that you get consensus about changing the style guide first. Otherwise I suggest that you revert the change, because it only serves to produce lint warnings for someone that was linting cleanly before. I would, in essence, agree with this. Not because of lint, but such void cast, primarily, demonstrates to developer/reader that function returns some value and we deliberately ignore it. However, functions like printf(), memset() and friends are common and well known cases where return value is usually ignored. We do not cast them in kernel code (nor I think we should). If lint complains about ignoring of memset() return value - I am really not interested. :) -- Mindaugas
Re: CVS commit: src/usr.sbin/wake
On Jan 11, 8:44pm, dholland-sourcechan...@netbsd.org (David Holland) wrote: -- Subject: Re: CVS commit: src/usr.sbin/wake | On Sun, Jan 10, 2010 at 11:23:19PM +, Christos Zoulas wrote: | Please explain to me how you propose we fix lint. | | Realistically, lint should be made to know a list of standard | functions with usually-uninteresting return values. That would solve | 95+% of the practical problem. | | In the long run (as I've said before) we should replace our lint with | something more modern... Once we get the llvm tools, I am fine with that. christos
Re: CVS commit: src/usr.sbin/wake
On Mon, Jan 11, 2010 at 06:31:29PM -0500, Christos Zoulas wrote: | In the long run (as I've said before) we should replace our lint with | something more modern... Once we get the llvm tools, I am fine with that. There's also splint, or sparse. -- David A. Holland dholl...@netbsd.org
CVS commit: src/usr.sbin/wake
Module Name:src Committed By: mbalmer Date: Sun Jan 10 19:40:12 UTC 2010 Modified Files: src/usr.sbin/wake: wake.c Log Message: remove stupid casts. if lint complains about return values that are ignored, lint should fe fixed, and not code being cluttered. To generate a diff of this commit: cvs rdiff -u -r1.11 -r1.12 src/usr.sbin/wake/wake.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
Re: CVS commit: src/usr.sbin/wake
In article 20100110194012.3693a17...@cvs.netbsd.org, Marc Balmer source-changes-d@NetBSD.org wrote: -=-=-=-=-=- Module Name: src Committed By: mbalmer Date: Sun Jan 10 19:40:12 UTC 2010 Modified Files: src/usr.sbin/wake: wake.c Log Message: remove stupid casts. if lint complains about return values that are ignored, lint should fe fixed, and not code being cluttered. Please explain to me how you propose we fix lint. Or to be more precise, how does the code convey intent about purposefully ignoring the return value of a function so that lint can ignore it. FYI this change is against the current style guide, so if you are going to be pushing for more changes like that, I suggest that you get consensus about changing the style guide first. Otherwise I suggest that you revert the change, because it only serves to produce lint warnings for someone that was linting cleanly before. christos
CVS commit: src/usr.sbin/wake
Module Name:src Committed By: mbalmer Date: Sun Jan 3 17:58:14 UTC 2010 Modified Files: src/usr.sbin/wake: wake.8 wake.c Log Message: Indicating the interface over which to send Wake on LAN frames is optional, if the interface is not given on the command line, wake(8) will use the first ethernet interface that is up and running. To generate a diff of this commit: cvs rdiff -u -r1.5 -r1.6 src/usr.sbin/wake/wake.8 cvs rdiff -u -r1.9 -r1.10 src/usr.sbin/wake/wake.c Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
CVS commit: src/usr.sbin/wake
Module Name:src Committed By: wiz Date: Sun Jan 3 19:04:26 UTC 2010 Modified Files: src/usr.sbin/wake: wake.8 Log Message: Capitalize Ethernet. To generate a diff of this commit: cvs rdiff -u -r1.6 -r1.7 src/usr.sbin/wake/wake.8 Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files.
Re: CVS commit: src/usr.sbin/wake
Am 27.06.2009 um 05:27 schrieb Simon Burge: Marc Balmer wrote: Am 27.06.2009 um 00:08 schrieb Marc Balmer: Am 26.06.2009 um 20:22 schrieb David Laight: On Fri, Jun 26, 2009 at 01:38:32PM -0400, Christos Zoulas wrote: I think this utility should be renamed wakeonlan because wake is too generic. FWIW I agree ... I disagree. Totally. and I forgot to mention why: the command name reflects what the command does, not how it does it. wakeonlan would be misleading, because a LAN could also be a token ring, where this would nor work... (ok, exaggerating a bit, but you get the idea) should we rename ssh to sshovertcpip ;) tcpdump to tcpdumponlan? wake is a lot more generic term than ssh or tcpdump. When I first saw the subject CVS commit: src/usr.sbin/wake it wasn't in any way obvious that this sent wake-on-lan packets to remote hosts. My first guess was something that woke up processes blocked in the kernel that are otherwise unkillable - try man -k wake for example. I deliberately choose that name when I first wrote the command (I use it on other OSs as well) and of course I first talked to my sponsors before I added it to NetBSD. Waking processes from the command line sounds like a strange use case to me. Cheers, Simon.
Re: CVS commit: src/usr.sbin/wake
On Fri, Jun 26, 2009 at 01:56:25PM -0500, David Young wrote: On Fri, Jun 26, 2009 at 07:22:04PM +0100, David Laight wrote: On Fri, Jun 26, 2009 at 01:38:32PM -0400, Christos Zoulas wrote: I think this utility should be renamed wakeonlan because wake is too generic. FWIW I agree ... I think that wake is a good command: short, imperative, mnemonic. I agree with Young David here. Thomas
Re: CVS commit: src/usr.sbin/wake
On Sat, Jun 27, 2009 at 08:00:03AM +0900, Izumi Tsutsui wrote: Now you see an example that discussion after commit is annoying ;-p Wake turbulence! -- David A. Holland dholl...@netbsd.org
Re: CVS commit: src/usr.sbin/wake
m...@msys.ch wrote: I think this utility should be renamed wakeonlan because wake is too generic. FWIW I agree ... I disagree. Totally. and I forgot to mention why: the command name reflects what the command does, not how it does it. wakeonlan would be misleading, because a LAN could also be a token ring, where this would nor work... (ok, exaggerating a bit, but you get the idea) should we rename ssh to sshovertcpip ;) tcpdump to tcpdumponlan? It looks you missed the point. - christos said wake is ambiguous, not all command should be verbose - ssh and tcpdump are third party commands, so you can't refer them for examples Now you see an example that discussion after commit is annoying ;-p --- Izumi Tsustui