Re: [Amsn-devel] ML still working ?
I totally agree with you... Users have to use perfect softwares and not buggy drivers... Anyway, I try to know where is the bug... Why ? Because with that it's the first time I hack in the kernel level... And I like that ;) It's so bad I haven't a pwc compliant cam. Else, I already had tracked the bug and sent a patch to Luc Saillard... (Doesn't open source means bring your help to all the edifice ?) Phil P.S. Totally OT but should we apply the holly_wood patch about DPs in directories ? Le Monday 12 June 2006 22:54, Youness Alaoui a écrit : not only that, but we'd need to shift the responsability... we really should stop being so nice to the users, look at what happened lately in the forums because of AA fonts and tcl/tk 8.5... he was a stupid bastard but let's take his example why bother if the problem isn't ours ? If users complain, we just need to say it's a driver bug, not an amsn bug, don't bother us, EVEN if we know the solution to it. We had so much work to do because of pwc, I'm fed up with it now. If a user has Tcl/Tk 8.0 and has a bug, we won't say ohh, you need to modify this and this line, no, we'll just say upgrade, you can't, then it's your problem, it's a requirement, in the same way a program can require Windows XP to run instead of Windows 3.1 for example. Now, if we let Luc Saillard know about this issue, he might fix it and provide a new version.. so when people complain, instead of telling them to fuck off, we'll just say upgrade your driver... The thing is, what you said below is not true, you say we cannot assume all users will upgrade, oh yeah ? then it's their problem, it's their choice, not ours, they didn't want to upgrade, so it's not our fault... also, the users will know how to upgrade and will have root access... you know why ? because the bug is with v4l2 not with v4l, and the only pwc driver that supports v4l2 is Saillard's driver, which is unofficial and not shipped anywhere, so if a user has the problem it means he has Saillard's driver which means he installed it himself, which means he has root access and he knows how to install it. If I'm wrong, then again, it's not our problem. Do you agree with me ? KKRT ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
On Tue, 13 Jun 2006 12:44:24 -0400, Philippe Valembois - Phil [EMAIL PROTECTED] wrote: I totally agree with you... Users have to use perfect softwares and not buggy drivers... Anyway, I try to know where is the bug... Why ? Because with that it's the first time I hack in the kernel level... And I like that ;) It's so bad I haven't a pwc compliant cam. Else, I already had tracked the bug and sent a patch to Luc Saillard... (Doesn't open source means bring your help to all the edifice ?) Cool! congrats! :) Phil P.S. Totally OT but should we apply the holly_wood patch about DPs in directories ? I think so, yes, but only in /trunk, it should never go in the 0.96 branch. Please include this info in the commit log so we don't forget... Le Monday 12 June 2006 22:54, Youness Alaoui a écrit : not only that, but we'd need to shift the responsability... we really should stop being so nice to the users, look at what happened lately in the forums because of AA fonts and tcl/tk 8.5... he was a stupid bastard but let's take his example why bother if the problem isn't ours ? If users complain, we just need to say it's a driver bug, not an amsn bug, don't bother us, EVEN if we know the solution to it. We had so much work to do because of pwc, I'm fed up with it now. If a user has Tcl/Tk 8.0 and has a bug, we won't say ohh, you need to modify this and this line, no, we'll just say upgrade, you can't, then it's your problem, it's a requirement, in the same way a program can require Windows XP to run instead of Windows 3.1 for example. Now, if we let Luc Saillard know about this issue, he might fix it and provide a new version.. so when people complain, instead of telling them to fuck off, we'll just say upgrade your driver... The thing is, what you said below is not true, you say we cannot assume all users will upgrade, oh yeah ? then it's their problem, it's their choice, not ours, they didn't want to upgrade, so it's not our fault... also, the users will know how to upgrade and will have root access... you know why ? because the bug is with v4l2 not with v4l, and the only pwc driver that supports v4l2 is Saillard's driver, which is unofficial and not shipped anywhere, so if a user has the problem it means he has Saillard's driver which means he installed it himself, which means he has root access and he knows how to install it. If I'm wrong, then again, it's not our problem. Do you agree with me ? KKRT ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel -- KaKaRoTo ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
I will do it tomorrow evening... Phil Le Tuesday 13 June 2006 20:17, Youness Alaoui a écrit : On Tue, 13 Jun 2006 12:44:24 -0400, Philippe Valembois - Phil [EMAIL PROTECTED] wrote: I totally agree with you... Users have to use perfect softwares and not buggy drivers... Anyway, I try to know where is the bug... Why ? Because with that it's the first time I hack in the kernel level... And I like that ;) It's so bad I haven't a pwc compliant cam. Else, I already had tracked the bug and sent a patch to Luc Saillard... (Doesn't open source means bring your help to all the edifice ?) Cool! congrats! :) Phil P.S. Totally OT but should we apply the holly_wood patch about DPs in directories ? I think so, yes, but only in /trunk, it should never go in the 0.96 branch. Please include this info in the commit log so we don't forget... Le Monday 12 June 2006 22:54, Youness Alaoui a écrit : not only that, but we'd need to shift the responsability... we really should stop being so nice to the users, look at what happened lately in the forums because of AA fonts and tcl/tk 8.5... he was a stupid bastard but let's take his example why bother if the problem isn't ours ? If users complain, we just need to say it's a driver bug, not an amsn bug, don't bother us, EVEN if we know the solution to it. We had so much work to do because of pwc, I'm fed up with it now. If a user has Tcl/Tk 8.0 and has a bug, we won't say ohh, you need to modify this and this line, no, we'll just say upgrade, you can't, then it's your problem, it's a requirement, in the same way a program can require Windows XP to run instead of Windows 3.1 for example. Now, if we let Luc Saillard know about this issue, he might fix it and provide a new version.. so when people complain, instead of telling them to fuck off, we'll just say upgrade your driver... The thing is, what you said below is not true, you say we cannot assume all users will upgrade, oh yeah ? then it's their problem, it's their choice, not ours, they didn't want to upgrade, so it's not our fault... also, the users will know how to upgrade and will have root access... you know why ? because the bug is with v4l2 not with v4l, and the only pwc driver that supports v4l2 is Saillard's driver, which is unofficial and not shipped anywhere, so if a user has the problem it means he has Saillard's driver which means he installed it himself, which means he has root access and he knows how to install it. If I'm wrong, then again, it's not our problem. Do you agree with me ? KKRT ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Why not contacting Luc Saillard to inform him of the bug ? Phil Le Monday 12 June 2006 18:40, Harry Vennik a écrit : I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net
Re: [Amsn-devel] ML still working ?
So what does upstream say about this bug? On Mon, 2006-06-12 at 18:40 +0200, Harry Vennik wrote: I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
of course that's a good thing to do, but it won't fix it for users with an older pwc, and we cannot assume all users will upgrade. some might even not know how to do it, or not have root access to the system. Op maandag 12 juni 2006 18:49, schreef Philippe Valembois - Phil: Why not contacting Luc Saillard to inform him of the bug ? Phil Le Monday 12 June 2006 18:40, Harry Vennik a écrit : I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list
Re: [Amsn-devel] ML still working ?
but we can still say: go update or don't bother us. Anyway it is important to let upstream know about this On Mon, 2006-06-12 at 19:16 +0200, Harry Vennik wrote: of course that's a good thing to do, but it won't fix it for users with an older pwc, and we cannot assume all users will upgrade. some might even not know how to do it, or not have root access to the system. Op maandag 12 juni 2006 18:49, schreef Philippe Valembois - Phil: Why not contacting Luc Saillard to inform him of the bug ? Phil Le Monday 12 June 2006 18:40, Harry Vennik a écrit : I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net
Re: [Amsn-devel] ML still working ?
not only that, but we'd need to shift the responsability... we really should stop being so nice to the users, look at what happened lately in the forums because of AA fonts and tcl/tk 8.5... he was a stupid bastard but let's take his example why bother if the problem isn't ours ? If users complain, we just need to say it's a driver bug, not an amsn bug, don't bother us, EVEN if we know the solution to it. We had so much work to do because of pwc, I'm fed up with it now. If a user has Tcl/Tk 8.0 and has a bug, we won't say ohh, you need to modify this and this line, no, we'll just say upgrade, you can't, then it's your problem, it's a requirement, in the same way a program can require Windows XP to run instead of Windows 3.1 for example. Now, if we let Luc Saillard know about this issue, he might fix it and provide a new version.. so when people complain, instead of telling them to fuck off, we'll just say upgrade your driver... The thing is, what you said below is not true, you say we cannot assume all users will upgrade, oh yeah ? then it's their problem, it's their choice, not ours, they didn't want to upgrade, so it's not our fault... also, the users will know how to upgrade and will have root access... you know why ? because the bug is with v4l2 not with v4l, and the only pwc driver that supports v4l2 is Saillard's driver, which is unofficial and not shipped anywhere, so if a user has the problem it means he has Saillard's driver which means he installed it himself, which means he has root access and he knows how to install it. If I'm wrong, then again, it's not our problem. Do you agree with me ? KKRT On Mon, 12 Jun 2006 13:24:10 -0400, Sander Hoentjen [EMAIL PROTECTED] wrote: but we can still say: go update or don't bother us. Anyway it is important to let upstream know about this On Mon, 2006-06-12 at 19:16 +0200, Harry Vennik wrote: of course that's a good thing to do, but it won't fix it for users with an older pwc, and we cannot assume all users will upgrade. some might even not know how to do it, or not have root access to the system. Op maandag 12 juni 2006 18:49, schreef Philippe Valembois - Phil: Why not contacting Luc Saillard to inform him of the bug ? Phil Le Monday 12 June 2006 18:40, Harry Vennik a écrit : I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will work. - If we force pwc to v4l instead of v4l2 it will also work. The first one seems quite ugly to me, so IMO forcing to v4l is the way to go. Op zondag 11 juni 2006 17:26, schreef Vivia Nikolaidou: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam).
Re: [Amsn-devel] ML still working ?
A buggy driver is not our fault, that is very clear. But if we know a work-around, we MAY use that to make things easier for our users. Of course we should not spend 3 days of programming to work around a bug that is not our fault, but a 10-minute fix is okay. And where the line is in between, that is hard to tell. Anyway, just to be sure, I installed the very latest snapshot of the pwc driver, and the bug is still there, so we can report it without the risk of a yell like 'USE THE LATEST SNAPSHOT' Op maandag 12 juni 2006 22:54, schreef Youness Alaoui: not only that, but we'd need to shift the responsability... we really should stop being so nice to the users, look at what happened lately in the forums because of AA fonts and tcl/tk 8.5... he was a stupid bastard but let's take his example why bother if the problem isn't ours ? If users complain, we just need to say it's a driver bug, not an amsn bug, don't bother us, EVEN if we know the solution to it. We had so much work to do because of pwc, I'm fed up with it now. If a user has Tcl/Tk 8.0 and has a bug, we won't say ohh, you need to modify this and this line, no, we'll just say upgrade, you can't, then it's your problem, it's a requirement, in the same way a program can require Windows XP to run instead of Windows 3.1 for example. Now, if we let Luc Saillard know about this issue, he might fix it and provide a new version.. so when people complain, instead of telling them to fuck off, we'll just say upgrade your driver... The thing is, what you said below is not true, you say we cannot assume all users will upgrade, oh yeah ? then it's their problem, it's their choice, not ours, they didn't want to upgrade, so it's not our fault... also, the users will know how to upgrade and will have root access... you know why ? because the bug is with v4l2 not with v4l, and the only pwc driver that supports v4l2 is Saillard's driver, which is unofficial and not shipped anywhere, so if a user has the problem it means he has Saillard's driver which means he installed it himself, which means he has root access and he knows how to install it. If I'm wrong, then again, it's not our problem. Do you agree with me ? KKRT On Mon, 12 Jun 2006 13:24:10 -0400, Sander Hoentjen [EMAIL PROTECTED] wrote: but we can still say: go update or don't bother us. Anyway it is important to let upstream know about this On Mon, 2006-06-12 at 19:16 +0200, Harry Vennik wrote: of course that's a good thing to do, but it won't fix it for users with an older pwc, and we cannot assume all users will upgrade. some might even not know how to do it, or not have root access to the system. Op maandag 12 juni 2006 18:49, schreef Philippe Valembois - Phil: Why not contacting Luc Saillard to inform him of the bug ? Phil Le Monday 12 June 2006 18:40, Harry Vennik a écrit : I think it simply isn't an amsn bug, but a driver bug. The streaming support on pwc seems to be somewhat buggy. That is why clearing the V4L2_CAP_STREAMING capability flag will make it work. On v4l the streaming support is not available at all, so it will always work. And if indeed it is a driver bug, we can only provide a work-around, not a real fix. Harry Op maandag 12 juni 2006 18:26, schreef Philippe Valembois - Phil: Hi, for me the better solution is to find the bug and eliminate it Phil P.S. I don't have marks anymore only curses, so I consider I am almost in holidays ! Le Monday 12 June 2006 18:17, Harry Vennik a écrit : But that's not much of an argument here, because any solution will require patching libng or even its v4l2 driver. Anyway, by doing it Youness' way we will avoid an ugly thing like if(driver == pwc) { . } by adding a real feature. The only disadvantage being all reports about 'webcam not released when using v4l2' that we certainly will get. Harry Op zondag 11 juni 2006 23:44, schreef Philippe Valembois - Phil: No without patching libng... Phil Le Sunday 11 June 2006 22:45, Youness Alaoui a écrit : can't we make it use the driver you choose ? for example, the user can choose v4l or v4l2 from the list of devices, so if he chooses v4l, the v4l driver will be used, if he chooses v4l2, then the v4l2 driver will be used... and have a label (we already have one, right ?) for pwc users that would say use the v4l device, not the v4l2 one... is it possible ? KKRT On Sun, 11 Jun 2006 16:37:37 -0400, Harry Vennik [EMAIL PROTECTED] wrote: Problem found. There are 2 work-arounds: - If we somehow unset the V4L2_CAP_STREAMING flag for pwc, it will
Re: [Amsn-devel] ML still working ?
On Sun, 11 Jun 2006, Youness Alaoui wrote: On Sun, 11 Jun 2006 00:06:12 -0400, Youness Alaoui [EMAIL PROTECTED] wrote: yes ? no ? YES. SO WHERE THE FUCK IS EVERYBODY :@ I just woke up. Good morning! Seriously, we have an empty TODO-list, so this can explain all the inactivity. How about RC1 now? ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Youness Alaoui wrote: On Sun, 11 Jun 2006 00:06:12 -0400, Youness Alaoui [EMAIL PROTECTED] wrote: yes ? no ? YES. SO WHERE THE FUCK IS EVERYBODY :@I just woke up. Good morning!Seriously, we have an empty TODO-list, so this can explain all theinactivity. How about RC1 now? ___Amsn-devel mailing listAmsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel-- Maxime VERZIER (www.billets-orientes.com / www.verzier.info) ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Well, i think phil already started fixing it... He made me do some stuff with the webcam and give him the output but he ran out of time and i haven't seen him since on irc.Here is the webcam debug output when i open the v4l1 channel of my pwc webcam in the amsn prefs : http://verzier.info/pub/amsn-pwc-v4l1-output(strange thing is that is talks about v4l2 whereas i clicked v4l... Phil was looking embarassed but as he left right after that, i don't know what i might need.) So phil or anyone else, whenever you got the time, just ask me on irc and i would be more then pleased to be the cobaye to solve that problem.On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam).Thanx!! :)(BTW, bugfixes can be made between 2 RC releases)Here's the diff - hope you (or someone else) can find a fix for that:--- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511)@@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev,captureItem-fmt); }+captureItem- dev.v-startvideo(captureItem-dev.handle, 25, 1);+ Tcl_SetObjResult(interp,Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK;@@ -506,6 +508,8 @@ return TCL_ERROR; }+capItem-dev.v-stopvideo(capItem-handle);+ // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) {-if ((capItem-image_data =capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data =capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width,fmt.height ); # endif___Amsn-devel mailing listAmsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel-- Maxime VERZIER (www.billets-orientes.com / www.verzier.info) ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
you have this, and the GUI thing where the 'cancel login' doesn't appear in the menu when we are logging in... in case network error or whatever, at least we can cancel... very annoying... KKRT On Sun, 11 Jun 2006 12:39:47 -0400, Max [EMAIL PROTECTED] wrote: Well, i think phil already started fixing it... He made me do some stuff with the webcam and give him the output but he ran out of time and i haven't seen him since on irc. Here is the webcam debug output when i open the v4l1 channel of my pwc webcam in the amsn prefs : http://verzier.info/pub/amsn-pwc-v4l1-output (strange thing is that is talks about v4l2 whereas i clicked v4l... Phil was looking embarassed but as he left right after that, i don't know what i might need.) So phil or anyone else, whenever you got the time, just ask me on irc and i would be more then pleased to be the cobaye to solve that problem. On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel -- KaKaRoTo ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Ooops I was working on porting my webcam driver to V4L2 :s I go immediately to IRC... Phil Le Sunday 11 June 2006 18:39, Max a écrit : Well, i think phil already started fixing it... He made me do some stuff with the webcam and give him the output but he ran out of time and i haven't seen him since on irc. Here is the webcam debug output when i open the v4l1 channel of my pwc webcam in the amsn prefs : http://verzier.info/pub/amsn-pwc-v4l1-output (strange thing is that is talks about v4l2 whereas i clicked v4l... Phil was looking embarassed but as he left right after that, i don't know what i might need.) So phil or anyone else, whenever you got the time, just ask me on irc and i would be more then pleased to be the cobaye to solve that problem. On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Op zondag 11 juni 2006 18:39, schreef Max: Well, i think phil already started fixing it... He made me do some stuff with the webcam and give him the output but he ran out of time and i haven't seen him since on irc. Here is the webcam debug output when i open the v4l1 channel of my pwc webcam in the amsn prefs : http://verzier.info/pub/amsn-pwc-v4l1-output Probably it's really v4l2, see my next comment. (strange thing is that is talks about v4l2 whereas i clicked v4l... Phil was looking embarassed but as he left right after that, i don't know what i might need.) Known bug (at least to me) If there is v4l2, it will take v4l2, no matter if you choose v4l or v4l2. This is because only the device name is used when processing the choice made. So phil or anyone else, whenever you got the time, just ask me on irc and i would be more then pleased to be the cobaye to solve that problem. On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
Le Sunday 11 June 2006 20:48, Harry Vennik a écrit : Op zondag 11 juni 2006 18:39, schreef Max: Well, i think phil already started fixing it... He made me do some stuff with the webcam and give him the output but he ran out of time and i haven't seen him since on irc. Here is the webcam debug output when i open the v4l1 channel of my pwc webcam in the amsn prefs : http://verzier.info/pub/amsn-pwc-v4l1-output Probably it's really v4l2, see my next comment. (strange thing is that is talks about v4l2 whereas i clicked v4l... Phil was looking embarassed but as he left right after that, i don't know what i might need.) Known bug (at least to me) If there is v4l2, it will take v4l2, no matter if you choose v4l or v4l2. This is because only the device name is used when processing the choice made. I discovered it too when I helped Elias I am trying to fix it removing duplicate entries... Phil So phil or anyone else, whenever you got the time, just ask me on irc and i would be more then pleased to be the cobaye to solve that problem. On 6/11/06, Vivia Nikolaidou [EMAIL PROTECTED] wrote: On Sun, 11 Jun 2006, Max wrote: I see at least one thing missing (imho) in the TODO-list and it is a bug fix for the pwc webcam's drivers. It was working perfectly before but it stopped working after 6511 (i checked). (i'am talking about the not released webcam). Thanx!! :) (BTW, bugfixes can be made between 2 RC releases) Here's the diff - hope you (or someone else) can find a fix for that: --- utils/linux/capture/capture.c (revision 6510) +++ utils/linux/capture/capture.c (revision 6511) @@ -480,6 +480,8 @@ captureItem-rgb_buffer = ng_malloc_video_buf(captureItem-dev, captureItem-fmt); } + captureItem-dev.v-startvideo(captureItem-dev.handle, 25, 1); + Tcl_SetObjResult(interp, Tcl_NewStringObj(captureItem-captureName,-1)); return TCL_OK; @@ -506,6 +508,8 @@ return TCL_ERROR; } + capItem-dev.v-stopvideo(capItem-handle); + // If a converter was used, close it and release the rgb_buffer if (capItem-handle) { ng_process_fini(capItem-handle); @@ -607,7 +611,7 @@ // - High resolution // - Low resolution for (dim_idx = resolution;;) { -if ((capItem-image_data = capItem-dev.v-getimage(capItem-dev.handle)) == NULL) { +if ((capItem-image_data = capItem-dev.v-nextframe(capItem-dev.handle)) == NULL) { # ifdef DEBUG fprintf(stderr,Capturing image failed at %d, %d\n, fmt.width, fmt.height); # endif ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
[Amsn-devel] ML still working ?
yes ? no ? -- KaKaRoTo ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel
Re: [Amsn-devel] ML still working ?
On Sun, 11 Jun 2006 00:06:12 -0400, Youness Alaoui [EMAIL PROTECTED] wrote: yes ? no ? YES. SO WHERE THE FUCK IS EVERYBODY :@ -- KaKaRoTo ___ Amsn-devel mailing list Amsn-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amsn-devel