Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 11:45:28PM +0200, Christoph Hellwig wrote: > On Wed, Apr 21, 2004 at 04:11:49PM -0500, Branden Robinson wrote: > > Err, when pasting source code regarding an issue like this, where we > > need a clean-room reimplementation, please identify the origin of every > > line. > > The surrounding code is from XFree CVS when I last updated (about two > weeks before the license change), and the changed code is all mine, > herby placed under public domain. Thanks for the clarification! -- G. Branden Robinson| "I came, I saw, she conquered." Debian GNU/Linux | The original Latin seems to have [EMAIL PROTECTED] | been garbled. http://people.debian.org/~branden/ | -- Robert Heinlein signature.asc Description: Digital signature
Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 11:02:12PM +0200, Fabio Massimo Di Nitto wrote: > > So what about the following unutested patch: > > Sorry but i can't accept an untested patch. Would you mind to give it a > shot at least? I don't think I have the time for a full blown X build, sorry.
Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 04:11:49PM -0500, Branden Robinson wrote: > Err, when pasting source code regarding an issue like this, where we > need a clean-room reimplementation, please identify the origin of every > line. The surrounding code is from XFree CVS when I last updated (about two weeks before the license change), and the changed code is all mine, herby placed under public domain.
Re: licensing of commit to XFree86 CVS
Hi Christoph On Wed, 21 Apr 2004, Christoph Hellwig wrote: > Okay, I looked over lnx_kbd.c, there's two thing that sprang to my mind: > > - the ifndef sparc and comment below is obviously wrong, the ioctl is >implemented in Linux common code, specifically vt_ioctl.c. > - fd 0 for the ioctl seems bogus. These ioctls are on the console >device, and to my reading of the X source that would be >xf86Info.consoleFd (?) > > So what about the following unutested patch: Sorry but i can't accept an untested patch. Would you mind to give it a shot at least? Unfortunatly, I have seen the code of the original fix and i can't comment on it, other than say that you are going in the right direction. Thanks Fabio -- fajita: step one Whatever the problem, step one is always to look in the error log. fajita: step two When in danger or in doubt, step two is to scream and shout.
Re: licensing of commit to XFree86 CVS
On Wed, 21 Apr 2004, Keith Packard wrote: > > Around 22 o'clock on Apr 21, Fabio Massimo Di Nitto wrote: > > > Would you have the time to do it straight ahead? Can you give me an > > estimated ETA? > > I haven't seen the description from Branden yet, but I expect it will take > only a moment or two to implement the fix. However, I'll be away from > 21:30 until 05:30 UTC, so it may well be 06:00 or so before I have > a patch ready. eh no.. this is not good enough :P Seriously.. thanks! i didn't even expect you to tell me the time with such details. I really appreciate your fast response. Thanks again Fabio -- fajita: step one Whatever the problem, step one is always to look in the error log. fajita: step two When in danger or in doubt, step two is to scream and shout.
Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 09:50:21PM +0200, Christoph Hellwig wrote: > Okay, I looked over lnx_kbd.c, there's two thing that sprang to my mind: > > - the ifndef sparc and comment below is obviously wrong, the ioctl is >implemented in Linux common code, specifically vt_ioctl.c. > - fd 0 for the ioctl seems bogus. These ioctls are on the console >device, and to my reading of the X source that would be >xf86Info.consoleFd (?) > > So what about the following unutested patch: Err, when pasting source code regarding an issue like this, where we need a clean-room reimplementation, please identify the origin of every line. We need to strive to keep code under the X-Oz or XFree86 1.1 licenses from being posted to this list. (I realize tained code is already in SVN, but that's what we're going to fix.) -- G. Branden Robinson| Suffer before God and ye shall be Debian GNU/Linux | redeemed. God loves us, so He [EMAIL PROTECTED] | makes us suffer Christianity. http://people.debian.org/~branden/ | -- Aaron Dunsmore signature.asc Description: Digital signature
Re: licensing of commit to XFree86 CVS
Around 22 o'clock on Apr 21, Fabio Massimo Di Nitto wrote: > Would you have the time to do it straight ahead? Can you give me an > estimated ETA? I haven't seen the description from Branden yet, but I expect it will take only a moment or two to implement the fix. However, I'll be away from 21:30 until 05:30 UTC, so it may well be 06:00 or so before I have a patch ready. -keith pgpYZjD0yWDzq.pgp Description: PGP signature
Re: licensing of commit to XFree86 CVS
On Wed, 21 Apr 2004, Branden Robinson wrote: > Fabio, it's up to you how long we should wait for a reply from Mr. Dawes > before going ahead with the clean-room rewrite (or dropping the patch). > Except for this one issue I'm all ready to go with the merge -- the > whole thing. Let's keep going preparing -8 and keep this change pending. Let's say that if noone can give us a patch for when we are ready to merge into branch or Mr. Dawes will not answer within that time we will keep the fix out. Too bad but it will be rescheduled. Thanks Fabio -- fajita: step one Whatever the problem, step one is always to look in the error log. fajita: step two When in danger or in doubt, step two is to scream and shout.
Re: licensing of commit to XFree86 CVS
On Wed, 21 Apr 2004, Keith Packard wrote: > > Around 14 o'clock on Apr 21, Branden Robinson wrote: > > > Also, if there is a C programmer present who is willing to perform such > > a clean-room rewrite should it come to that, please speak up. > > I can do that. Would you have the time to do it straight ahead? Can you give me an estimated ETA? Thanks Fabio -- fajita: step one Whatever the problem, step one is always to look in the error log. fajita: step two When in danger or in doubt, step two is to scream and shout.
Re: licensing of commit to XFree86 CVS
Okay, I looked over lnx_kbd.c, there's two thing that sprang to my mind:
- the ifndef sparc and comment below is obviously wrong, the ioctl is
implemented in Linux common code, specifically vt_ioctl.c.
- fd 0 for the ioctl seems bogus. These ioctls are on the console
device, and to my reading of the X source that would be
xf86Info.consoleFd (?)
So what about the following unutested patch:
--- lnx_kbd.c~ 2004-04-21 21:46:15.633943864 +0200
+++ lnx_kbd.c 2004-04-21 21:46:27.159191760 +0200
@@ -109,15 +109,13 @@
static int
KDKBDREP_ioctl_ok(int rate, int delay) {
-#if defined(KDKBDREP) && !defined(__sparc__)
- /* This ioctl is defined in but is not
- implemented anywhere - must be in some m68k patches. */
+#if defined(KDKBDREP)
struct kbd_repeat kbdrep_s;
/* don't change, just test */
kbdrep_s.rate = -1;
kbdrep_s.delay = -1;
- if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+ if (ioctl( xf86Info.consoleFd, KDKBDREP, &kbdrep_s )) {
return 0;
}
@@ -132,7 +130,7 @@
if (kbdrep_s.delay < 1)
kbdrep_s.delay = 1;
- if (ioctl( 0, KDKBDREP, &kbdrep_s )) {
+ if (ioctl( xf86Info.consoleFd, KDKBDREP, &kbdrep_s )) {
return 0;
}
Re: licensing of commit to XFree86 CVS
Around 14 o'clock on Apr 21, Branden Robinson wrote: > Also, if there is a C programmer present who is willing to perform such > a clean-room rewrite should it come to that, please speak up. I can do that. -keith pgpmYsbU7wwMY.pgp Description: PGP signature
Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 02:17:05PM -0500, Branden Robinson wrote: > Also, if there is a C programmer present who is willing to perform such > a clean-room rewrite should it come to that, please speak up. Please > note that it's just the *change* that needs to be rewritten, not the > entire files. In fact, the previous revisions of both the .c files > predated the relicensing on 2004-02-13. I'd say stop waiting and tell me what to change :)
Re: licensing of commit to XFree86 CVS
On Wed, Apr 21, 2004 at 02:06:33PM -0500, Branden Robinson wrote: > Revision ChangesPath > 3.3162+5 -1 xc/programs/Xserver/hw/xfree86/CHANGELOG > 3.27 +5 -5 > xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c > 1.6 +5 -5 > xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c In the event Mr. Dawes does not reply to my mail, or the answers to the questions I asked him are both "no", I'd like to ask subscribers to the debian-x mailing list to not look at these, or subsequent, revisions of the above files (well, the actual code files anyway -- the CHANGELOG isn't a problem as we don't borrow from it). This will keep your brains clear for a clean-room rewrite of this very small and simple patch. Since I've already seen the code, I'm tainted, but I can easily write a description of it in my own words, so this is not a big worry. Also, if there is a C programmer present who is willing to perform such a clean-room rewrite should it come to that, please speak up. Please note that it's just the *change* that needs to be rewritten, not the entire files. In fact, the previous revisions of both the .c files predated the relicensing on 2004-02-13. Fabio, it's up to you how long we should wait for a reply from Mr. Dawes before going ahead with the clean-room rewrite (or dropping the patch). Except for this one issue I'm all ready to go with the merge -- the whole thing. -- G. Branden Robinson|Fair use is irrelevant and Debian GNU/Linux |improper. [EMAIL PROTECTED] |-- Asst. U.S. Attorney Scott http://people.debian.org/~branden/ |Frewing, explaining the DMCA signature.asc Description: Digital signature
licensing of commit to XFree86 CVS
Hi David, Back in March you made the following commit: CVSROOT:/home/x-cvs Module name:xc Changes by: [EMAIL PROTECTED] 04/03/03 10:53:42 Log message: 6. Do the Linux KDKBDREP ioctl on the correct fd. This prevents the fallback that tries to directly program the keyboard repeat rate, and the related warning messages that recent Linux kernels generate (David Dawes). Modified files: xc/programs/Xserver/hw/xfree86/: CHANGELOG xc/programs/Xserver/hw/xfree86/os-support/linux/: lnx_io.c lnx_kbd.c Revision ChangesPath 3.3162+5 -1 xc/programs/Xserver/hw/xfree86/CHANGELOG 3.27 +5 -5 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c 1.6 +5 -5 xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_kbd.c The Debian Project is interested in including this fix in our next release of XFree86 4.3.0 packages. Are the above changes under the XFree86 1.0 license? If not, would you license them under the XFree86 1.0 license? Thanks for your time. -- G. Branden Robinson| "To be is to do" -- Plato Debian GNU/Linux | "To do is to be" -- Aristotle [EMAIL PROTECTED] | "Do be do be do" -- Sinatra http://people.debian.org/~branden/ | signature.asc Description: Digital signature
