Half a Page of Scribbled Linux

I have an old AMD K5 500MHz machine (named Bluto) that I have been using for wireless routing. It’s been running Redhat 6.2 for years. The cablemodem runs to my main machine (named Hoover), then via ethernet to Bluto. This all works fine, except that Internet access for the house depends on Hoover. Since this is my main machine, it is subject to tinkering, and when I run into trouble, there goes the Internet. I thought I’d buy a dedicated router to solve this problem. I got a LinkSys WRT54G and hooked it up. Unfortunately, the signal strength couldn’t match that of my old Orinoco card with the extension antenna. There were some other issues also including grabbing the DHCP assigned host name from Comcast for Hoover’s use. So I sent the router back and set about using Bluto as the router for the entire network.

One of the reasons I never upgraded Bluto from 6.2 was that the monitor attached to it is a Gateway CrystalSan 1024NI. This is a vintage 1990 monitor from the era when monitors didn’t always sync well with the input signal. Redhat 7.x used a text mode that caused the monitor to freak out. When I ran the Fedora installer, I was pleasantly surprised to see that problem was gone. I’ve since found out Redhat 9 works, too.

So I started the upgrade process on Bluto. BTW, if anyone has an old CrystalScan, set the scan rates to 48KHZ horizontal and 72Hz vertical to fix it to 800×600 resolution. It’s happy with that.

The installation went fine with one exception. The motherboard in Bluto has 5 PCI slots and 2 ISA slots. In the ISA slots are a 3Com 3c509 Ethernet card that connects to Hoover and the Orinoco PCMCIA to ISA bridge card that the Orinoco wireless Ethernet PCMCIA card plugs into. Fedora didn’t recognize the ISA cards at all. Neither did Redhat 9. In Redhat 6.2, there was a utility called pnpdump that would talk to the ISA bus and read in card data. You’d use the output of pnpdump to create /etc/isapnp.conf and edit this file to set up IRQs, and so forth. Fedora and Redhat 9 don’t have a pnpdump anymore. Not only that, but the standalone isapnp program has been moved into the kernel. So now I’m stuck figuring out how to create /etc/pnpdump.conf.

I have an email account that will be going away within a year. It supports large amounts of storage and IMAP. I’ve been looking for a replacement for it. First, I tried FastMail. This supports IMAP, but storage is only 10MB. After the first email with 4MB of family picnic photos, this limit proved much too small.

There’s been a lot of press about Gigabyte email accounts due to Google’s GMail. GMail isn’t publicly available yet, but some competetitors are. I signed up for a SpyMac account. This has web and POP3 access, but no IMAP, though a message on their site says it’s under development. In the meantime, I also checked out Walla. It’s a free Gigabyte account, but I couldn’t find out anything about access. After signing up, I found out it’s web only.

Installation of cdbkup done with ./configure --with-snardir=/mnt/sbackup --with-dumpgrp=root. /mnt/sbackup is a 10 GB partition I created for use with scdbackup. It’s a staging area for burning DVDs. the dumpgrp was necessary because cdbkup used a default group of operator, that I don’t have on my system.

The autodetect feature couldn’t tell how big my DVD was, so it used the default of 650 MB. Then after the message Collecting input data... it just exited.

I’ve been backing up my file systems to Travan TR-4 tape with a Seagate TapeStor SCSI unit for a few years now. Last week I was running low on tapes and a Froogle for them returned prices from $20 – $30 each! Given that these tapes hold 4 GB uncompressed and DVDs hold 4.7 GB, and media are a couple bucks at most, it was time to retire the tape drive.

So I started to look for a program that would back up to DVD. My first try was scdbackup. The system backup worked just fine. Then I tried to create a level 0 back up (to allow incremental back ups) using sdvdbackup_sys -conf_dir /root/backup_system -level -create_configuration. This hung during generating -content_list_adr. After running some tests creating a level 0 back up, it seems that scdbackup will fail collision detection if it is given an empty directory (such as /opt) to back up and it hangs during creation of content_list_adr if told to back up /dev.

I keep a monthly budget and have for years. This keeps me from bouncing checks and scrambling when the yearly automobile insurance is due, but I don’t have a clear picture of whether I’ll be playing golf every day after I retire or eating dog food. GnuCash has a retirement planning feature on its to-do list, and nothing else out there (for Linux, at least) provides this yet either. So I’m doing it using the old spreadsheet method with Gnumeric. But it’s really complicated.

After reading more about phpGedView and sifting through their online forums, it became clear that the authors expect phpGedView to be the web-based viewer of genealogical data and that GDBI will be used for editing this data. So I canned GRAMPS and pulled down gdbi-7-pgv.jar, the GDBI JAR file for phpGedView. A JAR is a Java Archive file, so of course Java is required to run it: java -jar gdbi-7-pgv.jar.

I won’t repeat what’s already in the documentation, but these fields required by GDBI could use some explaining: The URL is the location of your phpGedView website with /gdbi.php appended to it. So it should look something like http://www.example.com/phpGedView/gdbi.php. The GEDCOM is the name of the GEDCOM you uploaded to your site, e.g., my-family.ged. Leave the Read-only and Blank Editor boxes unchecked.

Make whatever changes you want, then just exit the program. To update the site, sign in to phpGedView. At the top-right of the page where is says logged in as (username), click on Admin. Choose the third item on the right side of the table Accept / Reject Changes.

At the bottom, under Import Gedcom, select my-family.ged – Import Gedcom. At the bottom of the next screen, select Continue Adding. On the next screen (Import GEDCOM), below Do you want to empty the dataset?, select the Yes button.

Sourceforge periodically sends a list of the top projects. On the most recent list, phpGedView caught my eye. It creates a web site of your genealogy data so you can share it with others in your family tree. My sister maintains my side of our family tree and my wife does hers. Once I started reading, it was apparent that this software, as well as its peers, is more for viewing than editing. So I started looking into open source genealogy software. I’d like to have something that can be run on multiple platforms, or at least exchanges data with other platforms. Currently, our family trees are in Family Tree Maker 6.0. This is a Windows program. Fortunately, it exports to GEDCOM, the most popular file format for geneaology data.

The first contender is GRAMPS. It is available as an RPM that installed with no trouble on my RedHat9 box. It imported my three data files that I exported from FTW in GEDCOM format with no trouble, except for not understanding telephone numbers (the PHON field). You may be wondering why I have three files. One is the original my sister made in FTW 4, the other is the most recent copy of my sister’s data and the third is my wife’s. As you can imagine there’s a fair amount of overlap. Under the Tools/Database Processing section of the menu in GRAMPS is a duplicate person finder. I ran this and it did find the duplicates. At the bottom of the dialog box is a Merge button. I selected a duplicate, hit merge and it popped up another dialog showing the fields of both records. I picked the lower-ID’s record’s fields and hit Merge and Edit. No joy:

Traceback (most recent call last):

File "/usr/share/gramps/MergeData.py", line 354, in on_merge_clicked

self.update(self.p1,self.p2,old_id)

File "/usr/share/gramps/plugins/Merge.py", line 268, in on_update

self.redraw()

File "/usr/share/gramps/plugins/Merge.py", line 243, in redraw

list.append((c,p1,p2.getId()))

AttributeError: 'str' object has no attribute 'getId'