Category: Computing

Well, it was time for a new RSS feed reader – and FreshRSS was the best choice for me. The kind folk at linuxserver.io maintain a container for it, and it fits nicely into my nginx setup.

Moving to a new tool for anything will have an adjustment period, it would be nice if the user experience was identical – but there is also something to be said about trying new things. Different isn’t always bad, change can be for the better.

Setting it up on a basic linux install looks simple, but I would suggest that adopting a container version is even easier. FreshRSS bootstraps and has a guided setup for first run (first connect on the web). You can even pick a sqlite database and avoid any database setup. It supports OPML import, which many other RSS feed readers support as well – so getting your feeds in or out of FreshRSS is easy.

My initial kick the tires install, I use the no-database required sqlite configuration. This worked well, but as I already run MySQL for my wordpress installs, why not use the same data store?

This is very easy – and the MySQL is also running in a container. The linux install covers how to setup your database, but as I’m using a container version I need to make one change to the setup – because ‘localhost’ isn’t the source of the network traffic to the container. (this is the use of ‘%’ instead of ‘localhost’ as shown below)

While allowing access from any host is a bit security scary, the container is already isolated to a private network so I’ve got layers of protection already.

During the web first run setup you have to then configure the database appropriately.

That’s probably the most complicated part of the setup. Now it’s just a matter of creating a user, and importing the OPML of my feeds and away I go.

Of course the new FreshRSS is going to ‘discover’ loads of new unread articles I’ve already read – and also for feeds I’ve been ignoring, not load enough unread articles. Oh well. So for a short while I run both – marking large chunks of articles as read in the new system, and burning down ones that were missed by the new system in the old one. It’s a one time hit and it gives me a chance to tweak some of the user setup configuration.

I’m only a day or so into using FreshRSS – but I’m adapting quickly to the differences. I will shout out to the Android app FeedMe which supports FreshRSS very nicely, once you enable the mobile API.

I will say that the themes are quite nice, I’m still getting use to the layout differences between it and the previous reader I was using. No regrets in making the switch to FreshRSS.

It was way back in 2013 that Google decided to shut down Google Reader. I had a rich list of RSS feeds that I read regularly and didn’t want to lose that experience, so I decided to switch to something I could host myself. At the time TT-RSS fit the bill nicely.

In the process of moving to a docker container managed server, I switched to the linuxserver.io tt-rss image. Then at one point I went to pull an updated version of the image and discovered it had been deprecated.

This resulted in my digging around a little to see what options I had to get a docker version of TT-RSS. I did find that there is a maintained version of a container build, but it seems to be only for development purposes. It didn’t feel like it was something I wanted to make use of.

As I dug deeper, I came across some dirty laundry. This didn’t really give me a good feeling about the TT-RSS project as a whole. Then I came across someone who was quite opinionated about TT-RSS. Well, this is the internet and there are lots of opinions out there. Still, I’d formed my own opinion along the way and decided that TT-RSS is no longer for me.

What next? Well, I decided to make that a completely different post which I’ll get up soon. The nice thing about many of the RSS feed readers, TT-RSS included – is that they support the export/import of your feeds. This makes it easy to switch.

Continue reading “Bye Bye TT-RSS”

We wanted to relocate the computer the kids were using to a more visible location in the house, this is generally a good idea since you want to be able to casually keep tabs on what they are up to online. The older Dell PC we had was the classic desktop box plus paired up with an older LCD monitor. The WAF was low.

One nice thing about the current state (slow death) of Moores Law is that even older machines are totally fine for most users. My personal desktop machine is a decade old, and aside from some minor (cheap) upgrades it’s still just fine. So of course I looked on the local used market to see if there were any computers that met the criteria. I came across a nice All In One Dell computer – this combines the screen and computer as a single unit.

Adding more RAM and swapping the hard drive for a SSD breathed new life into this 5 year old machine, it’s still a little plunky but can play minecraft and browse the web like a champ. The built in microphone and camera has been very useful lately as video conferencing for school is the new reality.

My go-to OS install is Ubuntu. For most use, this is just fine. It’s what I configured this new machine to run, but I then ran into the dreaded “need X program to run for school” problem. Windows 10 came with this hardware and was an easy choice to install to get things working. Overall, the windows experience isn’t so bad. You need an existing Windows machine to build a recovery USB from, but after that you can install it easily onto a new drive. I had thankfully kept the original hard drive around, allowing me to boot Windows to get this process started.

Windows 10 really, really, wants you to get a ‘cloud’ account and control everything through that. It is possible to setup a local only account, but only after you fail to connect to the internet. Of course I took this route, disconnecting the internet via a little bit of router firewall magic.

Now, I’d like to set restrictions on use. I want to avoid the kids doing a 5am wake-up so they can get a little quality computer time (minecraft/youtube) going while I sleep until a much more reasonable time of the morning.

While again, the easy path here is for Windows to just use the cloud based account (they call these Microsoft accounts) to manage the family controls, you can setup some limits for local only accounts. You do need to break out the command line to do this, and you can only pick “on the hour” granularity. So you can let them start at 7am, or 8am for example – but not 7:15 or even 7:30.

The local controls just let you set on/off times for access. You can’t control how many hours they can use. This is useful, but so very limited. I finally caved in and created a Microsoft account and linked the user to the cloud.

There are still limitations that leave some things to be desired. If the account doesn’t actually log off, they still count this as screen time. This leads to the hour and change we allow per day, to basically be used up soon after the first session of screen time starts (because who ever logs off?).

Granting more time is easy enough, Microsoft will send you an email (to your linked parent account) asking for more time. I’ve also seen improvements over the time we’ve used it – activity reports are starting to flow in now, and you can limit time per application. Spending and even access to the e-Store are built in as well.

One last thing in this Windows 10 setup. While Edge has now moved to a Chrome based browser, using the Microsoft supplied browser wasn’t my first choice. I will say that the Family controls are better (or maybe more deeply) integrated into the Microsoft browser. I’m still a fan of Chrome and the whole Google experience, this might be mis-guided at this point in time, but it still works ok. Getting Chrome installed is easy, and keeping the computer logged into a Google account lets me pull interesting data out of the use of that browser. I do want to prevent incognito browsing, and it turns out there is a way to do that with Chrome.

Maybe I’ll pivot to the Microsoft browser at one point, it would give us even more Family control features.