well, it's been a while....
Dumped PCBSD as a desktop because after several upgrades I realized I was really only running FreeBSD with KDE on top. I wasn't using any PCBSD "benefits" and a somethings had gone screwy with the upgrades - particularly the Xorg updates that were broken along the way.
I also realized that I just didn't want to spend time configuring and playing with distros....this occurred to me because I attempted ArchLinux because it looked to be perfect for me....optimized for 686, and oriented much like BSD. However, it took some thought and maintenance to keep going.....so I abandoned that.
My requirements are few....firefox, python, and a shell. Add on top of that vmware server because I wanted to abandon the separate server and use just one machine. Has to be fast and light.....my desktop has a mere 8MB on board video RAM. Enter fluxbuntu.
I ran fluxbuntu for a while....it's quick and already has my fundamental requirements save for vmware, but that was pretty easy to install. However FB was based on dapper and the subversion client for dapper was too old for my repo. Two months into FB I decided to try the upgrade to Feisty and it failed in various ways. FB also had some annoying issues. There was supposed to be a Feisty release of it but that didn't happen. They are anticipating a Gusty release....maybe I'll try that.
For now I'm on Xubuntu, with vmware running my server. Having the server on vmware makes for simple backups of the entire vmware images onto dvd.
Tuesday, July 24, 2007
Tuesday, April 17, 2007
homemade DVDs in a dodge caravan
Last year we bought a dodge caravan with an in-house dvd entertainment system. Great! The kid loves the wiggles and if we can make a long trip manageable, all the better. However, it seems that all my homemade movies just wouldn't play....loss of A/V sync, choppy video, it'd take 5 minutes to read the disc, most of the time it'd just fail.
I'd done some reading that implied that I needed to use DVD+R media and set the booktype to ROM so that the player would think it was a pressed DVD. So, I tried that, but the player said there was no disc.
A bit more googling led me to a groups post that said something about menus.....none of the DVDs I create have menus....it's just easier without. So, I recreated a DVD with menus, and whaddya know? It worked. Odd that the mere presence of menus affect the ability of the player to decode video and keep sync.
UPDATE 2008-01-19: seems that the stupid player doesn't like DVD+R discs....stick with quality DVD-R!
I'd done some reading that implied that I needed to use DVD+R media and set the booktype to ROM so that the player would think it was a pressed DVD. So, I tried that, but the player said there was no disc.
A bit more googling led me to a groups post that said something about menus.....none of the DVDs I create have menus....it's just easier without. So, I recreated a DVD with menus, and whaddya know? It worked. Odd that the mere presence of menus affect the ability of the player to decode video and keep sync.
UPDATE 2008-01-19: seems that the stupid player doesn't like DVD+R discs....stick with quality DVD-R!
Monday, April 9, 2007
freebsd printing with CUPS and a Samsung ML-2010
When I was running ubuntu last year I vaguely remember using my Samsung ML-2010 printer without too much fuss.....I think I found something in the excellent ubuntu forums that suggested using the ML-4500 driver....and it worked. Didn't always print nicely, but it pretty much worked out of the box.
So, yesterday I needed to print something....searching for FreeBSD+ML-2010 yielded little to no results, and nothing useful. In fact the top hit in google groups was a post by a guy I went to university with - but nobody responded to him. So, I embarked on my own journey. CUPS was already installed. I didn't really bother trying LPD cause I thought it'd be neat to print from any computer in the house via this box.
dmesg showed that I had /dev/ulpt0 - perfect. I started up CUPS, used my browser to view the web interface at port 631, and ran through the setup wizard to add my printer. But nothing samsung listed? So, I search the web for what to do next....turns out I need a PPD file for my printer.....samsung's website has one for linux....so I tried that.....completed the work in the web interface, print a test page.....and it complains. Missing some kind of filter? I guess I needed more than the PPD file.
So, keep searching.....splix a third party samsung driver.....but there's no freebsd port in /usr/ports :( As it turns out, there is one in CVS as well as several postings in the news group about creating the port files. After following the instructions in the PR report 111034
I had the port files.....but the Makefile had issues.....missing operator on line 34......there were spaces where there should have been tabs....copy paste error on my part. Forgot Makefiles were so damned picky about tabs.
So, I built the port, installed it....re-created the printer using the CUPS web interface. Print a test page and I get "Unsupported format 'application/postscript'!". Odd. Let me try from the command line......
/usr/local/bin/lp -dditdy-printer ubuntu and debian that claim this error was due to a missing file on those distributions with CUPS 1.2.3 - the file pstoraster.convs. Well, this file ain't on my system....."locate pstoraster" led me to /usr/ports/print/cups-pstoraster. Hmm....a seperate port.....wonder why something that seems so fundamental wouldn't be installed....? Well, I installed it, restarted CUPS tried the test page and no unsupported format complaint.
But the job sat in the queue. PR-99460 solved that issue: in printers.conf change deviceURI from usb:/dev/ulpt0 to file:/dev/ulpt0. CUPS won't be able to query the device but printing should work.....so I ran a test page after restarting cups and it appears to have completed (I did the test remotely, will have to wait until I'm home to see if it worked!).
Ugh....that was long and painful!
UPDATE:
I got home to find pages printed with errors listed on them. I can't remember the specific error, but suffice it to say that I'd used splix-1.0.1, and they'd recently released splix-1.0.1-1 specifically to workaround a firmware bug in the ML-2010. However, building this new splix might be an issue.....to start the tarball is half the size of the previous rev.
I did a diff on the two versions, and there were a few files changed as well as rastertospl is missing in the latest? I tried to build 1.0.1-1, but it complains about missing some tool ppdpo? Apparently that's part of the cupsddk, which I couldn't build either. OK, dig deeper.
A closer look at the diff shows that only one relevant file changed.....src/spl2.cpp had a two line change.....so I copied the file over to the 1.0.1 folder, repackaged the tarball, put it in /usr/ports/distfiles and tweaked the port's distinfo file to have matching MD5 and SHA256 signiatures and matching file size. Rebuilt and installed, printed a test page and it worked!
So, yesterday I needed to print something....searching for FreeBSD+ML-2010 yielded little to no results, and nothing useful. In fact the top hit in google groups was a post by a guy I went to university with - but nobody responded to him. So, I embarked on my own journey. CUPS was already installed. I didn't really bother trying LPD cause I thought it'd be neat to print from any computer in the house via this box.
dmesg showed that I had /dev/ulpt0 - perfect. I started up CUPS, used my browser to view the web interface at port 631, and ran through the setup wizard to add my printer. But nothing samsung listed? So, I search the web for what to do next....turns out I need a PPD file for my printer.....samsung's website has one for linux....so I tried that.....completed the work in the web interface, print a test page.....and it complains. Missing some kind of filter? I guess I needed more than the PPD file.
So, keep searching.....splix a third party samsung driver.....but there's no freebsd port in /usr/ports :( As it turns out, there is one in CVS as well as several postings in the news group about creating the port files. After following the instructions in the PR report 111034
I had the port files.....but the Makefile had issues.....missing operator on line 34......there were spaces where there should have been tabs....copy paste error on my part. Forgot Makefiles were so damned picky about tabs.
So, I built the port, installed it....re-created the printer using the CUPS web interface. Print a test page and I get "Unsupported format 'application/postscript'!". Odd. Let me try from the command line......
/usr/local/bin/lp -dditdy-printer ubuntu and debian that claim this error was due to a missing file on those distributions with CUPS 1.2.3 - the file pstoraster.convs. Well, this file ain't on my system....."locate pstoraster" led me to /usr/ports/print/cups-pstoraster. Hmm....a seperate port.....wonder why something that seems so fundamental wouldn't be installed....? Well, I installed it, restarted CUPS tried the test page and no unsupported format complaint.
But the job sat in the queue. PR-99460 solved that issue: in printers.conf change deviceURI from usb:/dev/ulpt0 to file:/dev/ulpt0. CUPS won't be able to query the device but printing should work.....so I ran a test page after restarting cups and it appears to have completed (I did the test remotely, will have to wait until I'm home to see if it worked!).
Ugh....that was long and painful!
UPDATE:
I got home to find pages printed with errors listed on them. I can't remember the specific error, but suffice it to say that I'd used splix-1.0.1, and they'd recently released splix-1.0.1-1 specifically to workaround a firmware bug in the ML-2010. However, building this new splix might be an issue.....to start the tarball is half the size of the previous rev.
I did a diff on the two versions, and there were a few files changed as well as rastertospl is missing in the latest? I tried to build 1.0.1-1, but it complains about missing some tool ppdpo? Apparently that's part of the cupsddk, which I couldn't build either. OK, dig deeper.
A closer look at the diff shows that only one relevant file changed.....src/spl2.cpp had a two line change.....so I copied the file over to the 1.0.1 folder, repackaged the tarball, put it in /usr/ports/distfiles and tweaked the port's distinfo file to have matching MD5 and SHA256 signiatures and matching file size. Rebuilt and installed, printed a test page and it worked!
Tuesday, April 3, 2007
MOre FreeBSD upgrade woes - large FAT32 partitions
After the upgrade to FreeBSD 6.2 I could no longer mount my 250G FAT32 partition? mount_msdosfs was complaining that the partition was too large. Turns out there's a kernel configuration options MSDOSFS_LARGE that needs to replace the standard MSDOSFS option in the GENERIC kernel. I guess PC-BSD had it enabled in the GENERIC kernel. So, I rebuilt the kernel....kinda wanted to stick with the GENERIC kernel since I don't want to add anything more to the plate of things I have to remember to maintain, but in the end I guess it's for the better since I was able to disable a whack of stuff in the kernel and ideally it will run quicker/smoother.
Monday, April 2, 2007
CD/DVD burners lost after FreeBSD upgrade
So, after an upgrade to 6.2 k3b no longer detects my DVD/CD burners? After upgrading all my ports and packages (and breaking a few in the mean time), it turns out that atapicam wasn't loaded....must have been in the GENERIC kernel for PC-BSD. Whatever. Adding atapicam_load="YES" to my /boot/loader.conf should fix that.
Intel E7221/i915/i810 video on FreeBSD
My PC-BSD 1.3 installation was based upon FreeBSD 6.1. My Dell SC420 has Intel E7221 on-board video which is based upon the i915 which uses the i810 driver. However, dmesg told me that it was using the isa-vga driver. Things were sluggish....after a while, switching windows would take a few seconds. my Xorg log said that /dev/agpgart wasn't found? Sure enough, nothing. After much searching, it seemed my best be would be to upgrade my kernel and source as the latest i810 driver had made some significant improvements.
So, I upgraded to 6.2 - unfortunately this didn't help. I found several posts that said to hack vga_pci.c to always create an agp device. But vga_pci.c doesn't even seem to be in the source anymore?
So, I upgraded to 6.2 - unfortunately this didn't help. I found several posts that said to hack vga_pci.c to always create an agp device. But vga_pci.c doesn't even seem to be in the source anymore?
Thursday, March 15, 2007
JVC Everio and making a DVD
Last year when I was looking to buy a camcorder I discovered that there was basically three kinds on the market - miniDV, dvd, and hard disk. I didn't want to have to spend hours on end transcoding video from a miniDV cam as well as having to deal with adding a firewire card, etc. Nor did I want to have to waste a lot of money buying the dvd-ram media that a dvd camcorder would require. For me it's all about making the most of my time. So I bought a JVC Everio GZ-MG20 hard disk camcorder. The everio records video in mpeg2 format and audio in ac3. Perfect - already in dvd ready format. As well, the fact that the transfer mechanism is USB and the hard disk shows up as a mass drive, makes it even simpler.
Now - when the hard disk almost filled up I decided it was time to make some dvds. So, I installed the windows software that came with the unit - power producer I think. I fed it the whack of files....created some menus, and expected it to simply author a dvd structure for me. But instead it started transcoding the video - it was going to take several hours?!?! But the whole point of recording in mpeg2 is so that it wouldn't be necessary to transcode.
So, I searched high and low and tried several dvd authoring packages - but all insisted on transcoding. Until I found TMPGEng DVD Author from Pegasys software. It let me create a few menus and create a dvd structure and it only took as long as the vob structure took to write to disk.
I have two issues with this software: first, it only runs on windows.....now that my desktop is BSD.....the second issue is that I would have liked for the chapter names of each of the videos to be set to the filename automatically, or at least an option to specify.....I run a script over the video files that renames them to the date that they were created....for this purpose. TMPGEnc automatically names each chapter as chapter 1, chapter 2, etc. Manually renaming 100 files is tedious at best, so I have settled for letting it name them this way. I also keep raw archives of the videos on dvd and my server, so the filenames/dates will be preserved there.
There were a couple of other nice options it provided - there is the ability to create a DVD without menus, as well as having the first action (when inserted into a dvd player) to play all tracks sequentially (when using menus).
So, in the end, for windows, TMPGEnc DVD Author seems to be the best solution for authoring DVDs. The reason for this post is as a precursor to some posts I hope to make regarding my trials and tribulations of authoring DVDs on BSD/Linux.
Now - when the hard disk almost filled up I decided it was time to make some dvds. So, I installed the windows software that came with the unit - power producer I think. I fed it the whack of files....created some menus, and expected it to simply author a dvd structure for me. But instead it started transcoding the video - it was going to take several hours?!?! But the whole point of recording in mpeg2 is so that it wouldn't be necessary to transcode.
So, I searched high and low and tried several dvd authoring packages - but all insisted on transcoding. Until I found TMPGEng DVD Author from Pegasys software. It let me create a few menus and create a dvd structure and it only took as long as the vob structure took to write to disk.
I have two issues with this software: first, it only runs on windows.....now that my desktop is BSD.....the second issue is that I would have liked for the chapter names of each of the videos to be set to the filename automatically, or at least an option to specify.....I run a script over the video files that renames them to the date that they were created....for this purpose. TMPGEnc automatically names each chapter as chapter 1, chapter 2, etc. Manually renaming 100 files is tedious at best, so I have settled for letting it name them this way. I also keep raw archives of the videos on dvd and my server, so the filenames/dates will be preserved there.
There were a couple of other nice options it provided - there is the ability to create a DVD without menus, as well as having the first action (when inserted into a dvd player) to play all tracks sequentially (when using menus).
So, in the end, for windows, TMPGEnc DVD Author seems to be the best solution for authoring DVDs. The reason for this post is as a precursor to some posts I hope to make regarding my trials and tribulations of authoring DVDs on BSD/Linux.
Subscribe to:
Posts (Atom)