Almost all of the relevant code is in wp-content/db.php. It's using Wordpress's hook that lets you use custom code to connect to the database. The code is rewriting MySQL flavored SQL into SQLite flavored SQL which is impressive but potentially flaky. I'd want to see how well it handles SQL from popular plugins before I'd start building a website using this.
If you want a simple blog with little extras, I'd expect this plugin to work just fine.
The predecessor plugin, SQLite Integration, was fairly stable with most other plugins, but plugins trying to do unconventional stuff or hit the DB directly would refuse to work. The main issue was that the plugin author stopped updating the plugin, while newer versions of WordPress didn't suppport it.
I recently deployed a web app on render.com and it looks it fits the description. I just told it to set up a PostgreSQL database, pointed it at my git repo and told it how to build/run the static site, web server, and cron job. It was pretty simple, it's got great features like logs, a console, and monitoring. The price is reasonable too.
Do you think this technique can be adapter for integration testing? So you wouldn't need a running PostgreSQL server to run tests that connect to a PostgreSQL database, or maybe it would make tests easier to run in parallel by using different SQLite files.
Tests that run against a different relational database from production make me really nervous.
The Django ORM has provided the ability to test against SQLite and deploy against PostgreSQL with the same code base (and the same tests) for years - and while it works incredibly well, I still won't use that in any system that I build. How your database behaves is such a crucial component of your application!
These days I find spinning up a real PostgreSQL database to run the tests is so easy there's essentially no reason not to do it. I use GitHub Actions for my PostgreSQL testing insurance, and it ends up just being a few extra lines in the YAML file.
I agree with Simon. I think you could run tests against SQLite locally so they're quick but then run them against Postgres in CI to ensure it works against the database you're running on. There's a lot of subtle differences in how the two databases work even though they both support a lot of the same SQL syntax.
You could also run SQLite in production. Then you don't have to test against a different database locally. :)
I'll add why Cloudflare is cheaper: it's subsidised. The fee for their own operations is $0.00, they expect you to use their other services though, which is paid.
Go daddy is definitely scammy. I’m using NameCheap for the domain and hosting of my blog and quite happy with their service and customer support. I’ve also used PorkBun for domain names. Their prices are reasonable as well
No. I've had domains registered with them for years. They have great support via online chat that I have used a couple times. Also a customer of their email hosting and it has been solid.
I've been an independent software developer for the past 3 or 4 years and I took a path that can be done by good software developer who is good at talking to clients. Talent agencies like Toptal and A.team are great to get started since you don't need to find customers. You can make $150 or more per hour which makes for a pretty comfortable living. Feel free to email me at dan@axelby.com if you'd like more specific advice or just to chat.
I've had a lot of success switching to freelancing by using Toptal and ATeam. You can go full time right away and you don't need to find clients. Feel free to email me at dan@axelby.com if you'd like to chat.
Positive Train Control is the train safety technology at issue here. It isn't ready. The "attempt at crippling the economy" is to shut down instead of operating in violation of the regulation.
The railroad companies claim to have spent billions on implementation but it's not available yet. As the article says, they're not allowed to collaborate or share implementation costs because of trust regulations. Also keep in mind the scale of the deployment. There's hundreds of thousands of cars on the track today, and putting something on each car isn't something that takes an afternoon.
The government can't make it happen by creating a regulation and a deadline.
Ummmmm, having been one of the original architects for GE on their CAD and PTC systems in the late 90's I have to disagree. The technology is pretty solid, the implementations even by competing companies are reasonable as well. The issues are political and liability, but frankly the claims are totally exaggerated on the cost, this is more about limiting future liability and changing a very ingrained old school culture. PTC helps to provide positive proof of negligence (or the lack thereof), which as you can imagine can generate potential liability for a railroad. PTC is not overriding of a train operator in all circumstances, hence the liability. And the train operator and general culture is VERY rooted, even more so than the automotive industry, which means moving these long timers off the ball takes way more than some government deadline.
As for not sharing costs, that is complete crap. The railroads are not developing the technology anyway, companies like GE and others have developed it and are continuing to develop it. Even interoperation between competing implementations is mostly moot. The railroads have to pay for the software, maintenance and on-going updates, but the development is coming from outside the railroads so that political BS claim is just that.
Not saying everything is perfect, but it frustrates me that after 15 years this isn't done and implemented, its 10 years beyond when I thought at least 2 of the biggest railroads would have it deployed at least in some form. Hell in 1998 GE had forms of it in test at 2 of the largest railroads in the country.
I don't think they have to put something on every (railroad) car. I think they have to put something on every engine, and on every switch. What they put on every switch has to talk to a network. There's a lot of switches, and some of those switches are in pretty remote places.
> Positive Track Control is the train safety technology at issue here.
I believe it's Positive Train Control.
> The government can't make it happen by creating a regulation and a deadline.
Very much true, and a big surprise to too many regulators.
You are correct, not every car needs to be modified. Generally engines are the primary modification, as is the CAD systems that must accept extra data and pass it to the trains. Things like hotbox detectors that can count trains axels, determine direction plus the switches and track occupation signals need to also be sent from the CAD (which already receives them in controlled territory) to the locomotive itself.
As a point, PTC works in both dark territory and controlled territory, because it focuses on the train itself (but it isn't as reliable in dark territory). Yes in controlled territory it has greater information but it doesn't require every switch to be modified. Even back in the late 90's every controlled switch reported back its status to CAD extremely quickly and CAD could control the switches remotely. Yes, dark territory is different, but PTC had some benefits even in dark territory. It was the long term goal to have near zero dark territory where a switch, siding or other similar factor would be able to cause a collision.
And just as a point, even when a train leaves one CAD controlled territory (say Union Pacific territory) and enters another (say ran by BNSF) the CAD operators hand off control, much like air traffic control. So the issue of interoperability is minimal as it already is in place in at least a fairly common form. Not saying it isn't without fault or couldn't be better, but the world doesn't have to change overnight, just keep making progress on each part making it better and better.
Computer aided dispatch. Essentially all trains are monitored via control centers owned by a few of the larger railroads. Passenger traffic is even handed off through these centers when crossing the territories.
And yes dark territory is what they call the areas that are unmonitored. Usually these are way out in the middle of no where on single track. Most monitoring believe it or not is hard wired and not wireless. For example, hot box detectors which are basically buried on the inside of the tracks monitor a trains bearing/axel temperatures and can report if one is over heating. This is critical because it is a source of derailments. Those boxes though usually are wired directly to lines running along the track and back to a common point. In many cases a number of lines terminate at one location and are fed to radio towers that beam that data back. In other cases it follows the fiber all the way to a rail yard or dispatch center.
I can't say now what the differences are with PTC in Europe, but I know when I was at GE we looked a lot at what they were doing because in most cases they are far more serious about their railroads.
I think dark territory means unmonitored sections of track or switches (to be more precise, I guess), that which isn't yet hooked up to the system to watch for speed and other issues that would cause accidents? New to me too :)
Also, there's no off-the-shelf version you can buy (this is all new technology), and it has to integrate with other railroads. For example, in the NY Metro area alone, you'd have to integrate the Long Island Railroad, MetroNorth, Conrail, Amtrak, New Jersey Transit, and possibly other industrial rail lines I'm not aware of. That alone will take a while to get right.
Ah, I see... so actually this is a big distinguishing factor that separates this from the other services I mentioned! Looking forward to playing around with this.
