[sane-devel] Schedule for release of sane-backends 1.0.22
I've been in contact with a couple of backend authors, who expressed concerns about this release. Let's back the schedule by two weeks, and give things some time to stabilize. Revised timetable: Jan 30, 2011: Feature freeze (only bugs + documentation updates) Feb 06, 2011: Code freeze (only horrible bugs + documentation updates) Feb 13, 2011: Release allan -- The truth is an offense, but not a sin
[sane-devel] Schedule for release of sane-backends 1.0.22
Nicolas Martin nicolas0martin at gmail.com wrote: Hi, What is weird here, is that the effect of overlap in memcpy does not seem to produce the same effect on different systems ! Using memcpy() on overlapping memory areas is undefined behaviour per spec. glibc has several memcpy() implementations optimized for different hardware capabilities. A recent implementation for Atom/Core2 CPUs performs the copy from tail to head (giving a nice 4x/5x speedup on these CPUs), which breaks with overlapping memory areas. Using memmove() is the correct way to handle overlapping areas, as documented. See http://lwn.net/Articles/414467/ for more (links/comments). HTH, JB. -- Julien BLACHE http://www.jblache.org jb at jblache.org GPG KeyID 0xF5D65169
[sane-devel] Schedule for release of sane-backends 1.0.22
Hi, Thanks for this pointer, very interesting and commented So we have to be double careful when using memcpy though, even if it acts faster than its cousin memmove ! Nicolas Le samedi 08 janvier 2011 ? 09:43 +0100, Julien BLACHE a ?crit : Nicolas Martin nicolas0martin at gmail.com wrote: Hi, What is weird here, is that the effect of overlap in memcpy does not seem to produce the same effect on different systems ! Using memcpy() on overlapping memory areas is undefined behaviour per spec. glibc has several memcpy() implementations optimized for different hardware capabilities. A recent implementation for Atom/Core2 CPUs performs the copy from tail to head (giving a nice 4x/5x speedup on these CPUs), which breaks with overlapping memory areas. Using memmove() is the correct way to handle overlapping areas, as documented. See http://lwn.net/Articles/414467/ for more (links/comments). HTH, JB.
[sane-devel] Schedule for release of sane-backends 1.0.22
On Fri, Jan 7, 2011 at 5:03 AM, Tomas Pospisek tpo_deb at sourcepole.ch wrote: On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noah kitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. I can also confirm that the MP450 still works, and that I believe is a generation 1 machine. Gernot -- ISP Asahi-Net: http://asahi-net.jp/en/ No.1 in Japan by customer satisfaction (Nikkei News, 7 July 2010)
[sane-devel] Schedule for release of sane-backends 1.0.22
After pulling commit e72c0f548eebfae0171ab8a2e2a74174c1bd96ae, the pixma backend works correctly with my MX850; there is no corruption, even on widths indivisible by 32. Thanks! - Dustin DeWeese On 01/06/2011 04:19 PM, Nicolas Martin wrote: AFAIK, I've currently only heard about this pending issue for the pixma backend, reported by Dustin DeWeese: https://alioth.debian.org/tracker/?func=detailatid=410366aid=312874group_id=30186 Maybe Dustin could give us more info about his current test status, using the latest git ? Nicolas Le jeudi 06 janvier 2011 ? 21:03 +0100, Tomas Pospisek a ?crit : On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noahkitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. Aren't we waiting for Nicholas' word on the issue? *t -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password to sane-devel-request at lists.alioth.debian.org
[sane-devel] Schedule for release of sane-backends 1.0.22
Thanks Dustin for your feedback. The issue of memmove vs memcpy appears in the last phase of image processing, to reduce the scanned line width, in order to get an exact scanned area width, and get rid of the Pixma modulo 32 line width. Il this case, the cropped line is copied above the scanned line, that explains the overlap situation, but the cropped line is always smaller, so no risk of buffer overflow. When the width is an exact multiple of 32, the cropped and original lines are exactly identical. What is weird here, is that the effect of overlap in memcpy does not seem to produce the same effect on different systems ! Looks some systems have memcpy to copy correctly above the original line, but not others. Anyway, the memmove here handles correctly the overlap situation, so this issue should be solved now. Nicolas Le vendredi 07 janvier 2011 ? 00:46 -0500, Dustin DeWeese a ?crit : After pulling commit e72c0f548eebfae0171ab8a2e2a74174c1bd96ae, the pixma backend works correctly with my MX850; there is no corruption, even on widths indivisible by 32. Thanks! - Dustin DeWeese On 01/06/2011 04:19 PM, Nicolas Martin wrote: AFAIK, I've currently only heard about this pending issue for the pixma backend, reported by Dustin DeWeese: https://alioth.debian.org/tracker/?func=detailatid=410366aid=312874group_id=30186 Maybe Dustin could give us more info about his current test status, using the latest git ? Nicolas Le jeudi 06 janvier 2011 ? 21:03 +0100, Tomas Pospisek a ?crit : On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noahkitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. Aren't we waiting for Nicholas' word on the issue? *t -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password to sane-devel-request at lists.alioth.debian.org
[sane-devel] Schedule for release of sane-backends 1.0.22
AFAIK, I've currently only heard about this pending issue for the pixma backend, reported by Dustin DeWeese: https://alioth.debian.org/tracker/?func=detailatid=410366aid=312874group_id=30186 Maybe Dustin could give us more info about his current test status, using the latest git ? Nicolas Le jeudi 06 janvier 2011 ? 21:03 +0100, Tomas Pospisek a ?crit : On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noah kitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. Aren't we waiting for Nicholas' word on the issue? *t -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password to sane-devel-request at lists.alioth.debian.org
[sane-devel] Schedule for release of sane-backends 1.0.22
Is the generational patch that Gernot discovered in git right now? allan On Thu, Jan 6, 2011 at 4:19 PM, Nicolas Martin nicolas0martin at gmail.com wrote: AFAIK, I've currently only heard about this pending issue for the pixma backend, reported by Dustin DeWeese: https://alioth.debian.org/tracker/?func=detailatid=410366aid=312874group_id=30186 Maybe Dustin could give us more info about his current test status, using the latest git ? Nicolas Le jeudi 06 janvier 2011 ? 21:03 +0100, Tomas Pospisek a ?crit : On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noah kitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. Aren't we waiting for Nicholas' word on the issue? *t -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password ? ? ? ? ? ? ?to sane-devel-request at lists.alioth.debian.org -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password ? ? ? ? ? ? to sane-devel-request at lists.alioth.debian.org -- The truth is an offense, but not a sin
[sane-devel] Schedule for release of sane-backends 1.0.22
On Thu, 6 Jan 2011, m. allan noah wrote: Is the generational patch that Gernot discovered in git right now? Yes, it is in. Sorry for the confusion and thanks Nicolas, Gernot Julien. *t On Thu, Jan 6, 2011 at 4:19 PM, Nicolas Martin nicolas0martin at gmail.com wrote: AFAIK, I've currently only heard about this pending issue for the pixma backend, reported by Dustin DeWeese: https://alioth.debian.org/tracker/?func=detailatid=410366aid=312874group_id=30186 Maybe Dustin could give us more info about his current test status, using the latest git ? Nicolas Le jeudi 06 janvier 2011 ? 21:03 +0100, Tomas Pospisek a ?crit : On Wed, 5 Jan 2011, Julien BLACHE wrote: m. allan noah kitno455 at gmail.com wrote: I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. I think Gernot has confirmed that the one-line fix does fix the machines he has access to. It does also fix my own pixma mp600. We do not know though about generation 1 machines. Aren't we waiting for Nicholas' word on the issue? *t -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password ? ? ? ? ? ? ?to sane-devel-request at lists.alioth.debian.org -- sane-devel mailing list: sane-devel at lists.alioth.debian.org http://lists.alioth.debian.org/mailman/listinfo/sane-devel Unsubscribe: Send mail with subject unsubscribe your_password ? ? ? ? ? ? to sane-devel-request at lists.alioth.debian.org -- The truth is an offense, but not a sin
[sane-devel] Schedule for release of sane-backends 1.0.22
On Sun, Jan 2, 2011 at 8:38 AM, Julien BLACHE jb at jblache.org wrote: avision is not yet back to working order as it was prior to Rene's last code drop that broke everything. Same for epson2, there are a number of bugs that still need to be fixed to get it back to a usable state. pixma has a bug that affects one or even two generations of devices; fix is being worked on but it's not there yet. J- I think we need give the backend authors some help with more specifics. Alessandro thinks only the Perfection 610 is still broken, and Mike has only an ADF and a network issue outstanding for avision. I've not heard from Nicholas. Do you have more detail, maybe debian bug numbers that have are still broken with the most recent git? allan -- The truth is an offense, but not a sin
[sane-devel] Schedule for release of sane-backends 1.0.22
m. allan noah kitno455 at gmail.com wrote: Hi, J- I think we need give the backend authors some help with more specifics. Alessandro thinks only the Perfection 610 is still broken, The 640u is reportedly still broken (#583166), Olaf says it doesn't like the area setting; the exchange was Cc:ed to both Alessandro and the list so everybody should be up to speed on this one :) and Mike has only an ADF and a network issue outstanding for avision. There's also the broken support for the various versions of HP5300C that stopped working after Rene's last update; this is the biggest Avision issue btw. The HP5300C was quite popular and sold for a long time. I also have a report for the AV120 (#548731). I've not heard from Nicholas. The pixma fix is still in work, IIRC we should hear back from Gernot in the coming weeks. JB. -- Julien BLACHE http://www.jblache.org jb at jblache.org GPG KeyID 0xF5D65169
[sane-devel] Schedule for release of sane-backends 1.0.22
There is quite a bit of updated code (and many months) since our last release. In particular, there have been some bug fixes to regressions in the avision and epson2 backends, and some major additions to other backends (genesys, pixma, etc). If those developments are stable, I think it's time for a release. Timetable: Jan 16, 2011: Feature freeze (only bugs + documentation updates) Jan 23, 2011: Code freeze (only horrible bugs + documentation updates) Jan 30, 2011: Release Speak up now if this timetable needs to change. I'd rather have stability than a speedy release :) allan -- The truth is an offense, but not a sin
[sane-devel] Schedule for release of sane-backends 1.0.22
m. allan noah kitno455 at gmail.com wrote: Hi, release. In particular, there have been some bug fixes to regressions in the avision and epson2 backends, and some major additions to other avision is not yet back to working order as it was prior to Rene's last code drop that broke everything. Same for epson2, there are a number of bugs that still need to be fixed to get it back to a usable state. backends (genesys, pixma, etc). If those developments are stable, I pixma has a bug that affects one or even two generations of devices; fix is being worked on but it's not there yet. I'd delay any new release until epson2 and pixma are back in working order, at least. JB. -- Julien BLACHE http://www.jblache.org jb at jblache.org GPG KeyID 0xF5D65169
