Thanks. My phone is on 14 and won’t get another update, oh well.

latest release of android

Does that mean 15?

Command line mplayer has been plenty for me.

Might be worth patching fail2ban to recognize the scrapers and block them in iptables.

I’m satisfied. Virmach has had ups and downs but their network is fast too.

If you want something Hetzner-like in North America, you might look at OVH. They have a data center in Beauharnois (BHS), Canada, which is near Quebec. It’s not too different from using a server in the Northeast US.

That article is almost 4 months old and yeah, the internet in EU is better than in the US. But you can get US plans with plenty of bandwidth. I’ve been happy with buyvm.net and there are many others. Hang out on lowendspirit.com for a while to get a sense of things.

Usually you write the book with a text formatter and package the results in to an epub, so IDK if it’s common to edit the epub directly.

The epub is just a zip file containing a metadata file and a bunch of simplified HTML files (one per chapter). So if you’re comfortable editing HTML, or better yet writing scripts, you can probably slap together something simple that unpacks the epub, strips those images out of the pages, and re-packages the epub.

Github: Microsoft code hosting site that feeds all your code through AI training and tries to lock you in through their pull request and related machinery. Once used a motto like “social coding”, but let go of that when they realized Facebook for nerds didn’t sound that great. Software is mostly proprietary besides Git itself.

Gitlab: 1) a Github competitor (gitlab.com, code hosting site with somewhat similar features; 2) the software for that site, huge and bloaty and slow, written with Ruby on Rails. You can self host it if you want, but yecch.

Forgejo: Git front end software, fork of Gitea and/or Gogs. Small and fast and written in Go. Fewer features than Github or Gitlab. If you want to self-host, I’d use this or some variant. Quite easy to install and run.

Gitweb: comes with git, pretty rudimentary but has old school attractiveness at least for me. Really just a browsing interface. No pull requests or anything like that.

Git, just plain Git: if you are self-hosting a project for yourself and maybe a few friends/collaborators, it’s fine to just use git with no web stuff, and push/pull by ssh. You’d manually install account credentials for your friends. This is really the simplest, but NO fluffy UI or other creature comforts.

Fossil: amazingly small and fast alternative to all the above (fossil-scm.org) but uses its own VCS (Fossil) that doesn’t interoperate with Git. I think the author said he might convert it over sometime. It’s written in C! Uses sqlite as repo backend instead of the file system like git uses. Has built in wiki, bug tracking, documentation viewer, etc. and used about 2MB of ram last time I tried it, ridiculously small (Gogs used around 40MB and Gitlab uses gigabytes).

Sourceforge (sf.net), very old school code hosting site, not of much relevance any more. They released an old old version of the software a long time ago and that got forked to become Savannah.

Savannah (savannah.gnu.org) hosting site for GNU and related software. Also savannah.nongnu.org for non-GNU stuff in the same spirit. I don’t know the exact criteria for putting stuff on nongnu but I think it’s on a project-approval basis, rather than letting everyone upload whatever they want.

Darcs (darcs.net), another alternative to git, better in some ways, written in Haskell, lost most of its users after a self-inflicted footbullet around 5y ago. There was a hosting site (darcsweb?) for it but that looks to be gone now.

There are a few more of them too, none of much importance these days even though some were interesting.

I’m on a 4gb machine right now and it’s tolerable if I don’t do too many things at once, but Google Docs bogs in particular bogs it down.

Upgrade that box or repurpose it for something else. Web bloat has made 2gb machines useless for browsing and 4gb marginal, if the user needs Google docs, put in 8gb or more.

Very reliable hard drives don’t exist whatever the price. You need RAID. But, look at backblaze drive reliability statistics to identify some obvious problem drives to avoid. It would help if you said what you are trying to do with the drives, what capacity you want, etc.

If you look closely, a .epub is just a zip archive of a bunch of very simple html pages. So extracting the zip and viewing the HTML with a browser works fine.

I’ve never understood the fuss about Mint, but I thought it was more of a tinkerer’s distro? I’ve been using Debian, which has its warts, but seems to want to minimize loose ends (not always successfully). Some more explanation of Mint’s benefits could be useful.

If the drive has bad sectors that it can’t read right now, it likely had other sectors that were marginal and got copied (remapped) to new spare sectors before they became unreadable. So there is still potentially recoverable data in the remapped sectors, and not much you can do about it.

Basically, writing zeros to the disk is about as good as you can hope for. If your data is s00per seekrit to the point where you can’t stand the possibility of any bits at all being recovered, you basically have to melt the drive. Otherwise, zero it and send it in.

Next time, set up encryption ahead of time, so your new drives never see plaintext. Some drives have a “secure erase” feature that is basically a crappy version of this built into the drive.

If you had something hog memory and a lot of other stuff got paged out of ram as a result, that can slow things down. Try running “top” and see how much swap space is in use. If it’s more than a little bit, once you have enough memory available by shutting off whatever was hogging it, try “swapoff” (pages the stuff back in, which can take a little while) followed by “swapon” to re-enable swap.

You mean RPL? There are lots of interpreters for that already. That’s the sort of thing to write because you want to write it, not because other people want it, unless you’re part of a community that is asking for it.

If you just want to implement an interpreter because language implementation interests you, I suggest Lisp or Scheme rather than something like RPL. The book “Structure and Interpretation of Computer Programs” explains how to do that and it is great. You can read it online at mitpress.mit.edu/sicp .

I use TB under Debian and there is a tray icon and an arrival notification, poll time of maybe a few minutes, seems fine. Showing the # of messages in the tray icon could be sort of handy I guess, though I had never thought about it before and didn’t miss it. Basic features = shut off the “email contains remote content” banner or “spam filter thinks this email is spam” (I can recognize spam for myself). I just want a preference that permanently disables remote content without throwing banners at me. And eliminate the client side spam filtering completely since I have that on the server side, and can manually flag any that gets through. Plus various other stuff like that. Yes, get rid of the calendar and contacts stuff. Biggest feature needing significant code changes: make message search not suck.

This doesn’t really seem like something that needed to be done. Thunderbird already has too many features. It needs less, not more. A bunch of stuff in the email client part is also badly designed. That needs fixing, preferably upstream, but I wouldn’t think of that as feature enhancement.

I don’t bother with a proxy host or with LetsEncrypt, though I guess you could use LetsEncrypt perfectly well. Back when I was doing this, LetsEncrypt didn’t exist and you had to actually pay for public certificates, so using locally generated free ones saved money. It also had a minor(?) security advantage in that if the private server key somehow leaked, it wouldn’t let people impersonate our internet domain.

For the private CA I simply used the crappy CA.pl script that comes with OpenSSL or did at the time. There are much better ways to do it, especially at any kind of scale, but CA.pl sufficed dealing with a few development machines.

Proxy host out on the public internet? Usually I just use a local private CA for this, and install the CA root in my browser.