@fwc@j@prettygood It would probably yell at you for having a v16 cluster and fail anyway. A dump and restore makes more sense here. And if you make an improper dump and left the v12 cluster alone, you still have a working backup of it that you can dump again.
>I'm reading though the pg_upgrade instructions and some of the steps might as well be in Chinese. The general instruction are: 1. Backup the whole cluster 2. Stop PostgreSQL 3. Install the upgrade package (postgresql-server-upgrade or similar) and upgrade to newer version 4. Initialize the a new cluster with the newer version with the same settings the old cluster has 5. pg_upgrade --old-bindir=<location of old PostgreSQL binaries provided by the upgrade package> --new-bindir=<location of the new PostgreSQL binaries> --old-datadir=<location of old data files> --new-datadir=<location of new data files (from the newly initialized cluster> --check (remove --check after it checked the cluster and didn't throw any errors) 6. Start PostgreSQL again 7. su postgres -c '/var/lib/postgresql/X/analyze_new_cluster.sh' or su postgres -c 'vacuumdb --all --analyze-in-stages'
Some distros provide an easier script for DB cluster upgrades that do the new cluster init and pg_upgrade for you. RHEL for example has postgresql-setup --upgrade, Debian might also have something.
@fwc@prettygood You want to move the database from v12 to the v16 cluster without upgrading? Wouldn't a dump from v12 and restore to v16 be enough for that since you already have v16 running anyway?
>What am I supposed to see here because it loads on my end. There's no content tho so that's maybe what you mean ? I've disabled unauth access and forgot about it, sorry. See screenshots bellow.
>This has been a thing for many many years. Not in Chromium and Chrome. Those features do not exist there.
>Which one ? because as you can see in the screen recording I don't see anything you can't move. The one on the far left next to the first tab and the one on the far-right which is supposed to be for searching tabs or something. image.png image.png image.png
@mangeurdenuage@japananon@zaitcev Instead of arguing about it, here are some examples of Firefox being broken as of recently. It took almost no effort to find my posts and they are probably a few more.
You know how many bugs I encountered in Chromium and Brave every time I tried them for a month when FF was unusable for me? Zero. Not even a memleak that FF likes to do every few months after an update.
And with Brave, many of the lacking options in base Chromium were added like blocking AutoPlay. There are only two things I don't like about Brave. They are an ad company and they crypto/AI BS baked right into the browser (can be disabled in brave://flags; the equivalent of about:config). And all the annoying new buttons that are in Brave can be hidden, while the new annoying buttons with zero purpose added to FF cannot be hidden without hiding them via userChrome.css.
@fwc@prettygood You _can_ revert that one commit, but do you really want to set up yourself for debugging hell, when they eventually start to depend on new functionality not preset in 12?
That said, I would still be running version 10, if Pleroma didn't drop support for it.
@prettygood@fwc The deprecation policy just states that they remove support for software not supported upstream.
Depending on how stupid Synapse is (never tried it), I would just dump the whole DB as a backup, test the upgrade on a different system, and if all goes well, go through the upgrade process on the main system, and reindex the whole DB after the upgrade. It's always possible to rollback a version with DB backups.
@fwc@prettygood The DB version usually doesn't matter to the software connecting to it, unless it depends on some new functionality which should have been noted somewhere.
The commit that removes the support has nothing of value in the message.
@nyanide@meso You don't really attract attention from people that might want to take advantage of that clause I think. It's not like they scrape domains registered with them for "bad" content.
What caught me of guard wasn't that they had the hate speech clause. It was that hate speech wasn't defined anywhere in their terms. That's what set the alarms for me. If the listed example, or clearly defined, I wouldn't mind it as much.
@nyanide@meso Yeah, the ryona.agency thing is understandable considering they also don't want to manage their own crypto platform. I don't even know if Frantech accepts crypto without some KYC from the crypto platform. I've never tried to pay them via crypto.
And the nameless company thing... I'm not doxing myself today. :DD It was a frustrating email chain, but they worked it out. I think they are still using them to this day even.
@nyanide@meso The most notable one was cum.salon which used to be on Porkbun and got seized by them (doxing). getgle.org also had some problems with them. They called the mascot CP or something I think. Quite ironic considering their track record with actual CP/loli. A nameless company which I worked for had few problems with their support after some trolls decided it would be funny to send abuse emails to them.
Other than that, they are usually fine, but their TOS irks me the wrong way. The hate speech clause was very broad when I last checked it months ago (they might have changed it) and since they are sometimes very trigger happy with it, I don't like it.
ryona.agency almost didn't get renewed because Visa/MasterCard cards from the Russia aren't obviously allowed and their crypto options suck, but that's a different thing.
@meso Oh, that's fucking stupid. Anyway, Namecheap is probably the only good registrar left. Epik has a laughably shitty web UI that barely works, got purchased and immediately stared complaining about hate speech on Twitter and who knows how good their security really is. Njalla outright just seizes your domain whenever they want to, because it's ran by the pirate bay owners who really like antifa. And Cuckflare is well... Cuckflare.