Hi everyone
Today I got a notification from Chrome saying they turned off Timimi because it is no longer supported by Chrome. I was able to force it on again, but does anyone know if this being rectified, or if there is an alternative that is not a pain in the butt to get running?
This is the problem with Manifest 3. Lots of extensions are effected.
Riz announced awhile back that he was leaving TiddlyWiki to spend more time with other software, or something like that. So unless he comes out of retirement, probably not much help from that quarter.
The 3 solutions that work really well (though actually I haven’t tested them under Manifest 3) are
You just keep a copy of TiddlyStow (TS) on your desktop, and open it whenever you want to open a TW. It even gives you a menu of the last 10 or so files that you have opened. The downside is that if you have any external files with relative paths, you won’t be able to see them. There are workarounds to that problem, possibly.
The TW5-browser-nativesaver (TBNS) uses a plugin. So the main complication is setting up your TW5 file initially. After that, there are a couple popup question boxes you have to answer in the affirmative every time you reload. Also you need to tell it to use the internal database of your browser, or it will ask you for a save destination every time you save. But once you have it working it understands relative paths, so it’s more natural that way.
The tool I mostly use now is rclone. You just download rclone from rclone.org. Go to the directory where you have put the rclone executable (typically rclone.exe on Windows). Give a command like:
./rclone.exe serve webdav ~/data/Wikis --addr :8080 --vfs-cache-mode full
Replace ~/data/Wikis with the path to your folder of TW files.
Now you can navigate in your browser to 127.0.0.1:8080 and then click through to your TW file, double click and open. You can have as many TW files as you want open and they all share the same port number. Once you have your TW files up you can add them to your bookmarks so you can find them quickly next time.
This method should work with any modern browser no matter what. It doesn’t require a separate helpler file or a plugin. It does require you to start up the program whenever you reboot. I imagine it’s possible to automate this by putting a shortcut in the startup folder, but I’ve never done this to verify.
There’s also a way to open it up to your entire network, so if you’re on a wifi router you trust all your devices can access your files. It’s not something you would want to do in a coffee shop or an airport, but probably ok at home.
Just wanted to note that Timimi is still working fine with Firefox, if you’d prefer to switch browsers rather than switch saving solutions.
I think my TiddlyWiki days are done. At the very least I will never again be able to recommend it to a non-developer.
I will try TiddlyStow and maybe TiddlyDesktop as last ditch attempts.
Native saver’s example-wiki.html does not walk me through step by step but assumes I know how it works. Also, I added the plugin to a file and now I am getting red JavaScript error windows.
And you are saying Rclone forces me to use command line every time. Also a deal breaker for me. I work with ten or more TWs a day, most of them open at the same time. I can’t be opening the command line and switching them constantly. Having to use the command line as a non-techi is why I never warmed up to node.js.
Someone needs to find Riz and ask him to bail us out for this emergency. Or else he needs to hand over Timimi to someone who will fix it and maintain it.
I have other software options for taking notes. But I use TW to publish materials. That will require some major shifts in my thinking, my workflow, etc. This is a very bad day for me.
Could I just run Timimi indefinitely even though it is not a supported extension? Or will it eventually not work on Chrome? Or will they permanently block it from working somehow?
Thanks etardiff.
Unfortunately my employer had me switch to Chrome, and I use TW and Google products back and forth. Back and forth between browsers would be a big pain, at least for the work-related TWs I use.
If the wikis are co-located, all below some directory, then you can just serve up that directory. You could conceivably serve up your entire hard-drive, as long as you stay local (127.0.0.1).
Windows also has “junctions” which are like links. So you might be able to have a directory that you serve up that has sub-directories which are actually junctions to your files wherever they are. Haven’t tried this, since I spend very little time with Windows these days.
Just to be clear, you don’t run this every time you start a TW file. You run this once when you boot up your computer. It serves up your files as webdav. One instance can serve up an indefinite number of TW files, as long as they are in subdirectories of your serving directory.
If you need a second directory running, you can run another instance of Rclone with a different port number (e.g. 8090).
Even to do this, I would need a walkthrough by phone or Zoom while on the computer. Even these simple instructions have big gaps for someone like me. For example, the first sentence should read “In Command Prompt, go to the directory…”, at least I figured that out. Then I went online to refresh my memory on how to navigate to the folder in Command Prompt.
Here is how far I got: (tws is the directory with rclone, and me is the subdirectory with a TiddlyWiki file to test it out)
You guys don’t even remember the levels of competency that you have all passed through years ago that I have either not done or learned once to try out node.js and have since long forgotten. Probably I have my relative paths wrong in the command. The filepath is OneDrive/tws/me. But that is not the error it mentions.
I am letting this go for tonight. Gonna start investigating Obsidian again…
./ might be a muscle memory slip of a Linux user. Since this is an .exe file, you don’t need this part in Windows. As long as you have rclone.exe either in current directory or in PATH, it should work without the ./ prefix
Thanks vuk. Tried it, and now it is saying I need to get my admin to allow it. So everything is a moot point until at least tomorrow.
My fault. I was speaking winlux. (Like Spanglish, for computers) The command should be just rclone.exe (assuming you put rclone in the same dir). In Linux you always have to start with ./command.
No one has every accused me of passing through levels of competency …
You can run windows terminal as admin.
I’m kind of surprised that it requires this though.
BTW, did TiddlyStow work?
Another way is to use TiddlyHost, but with the browser storage plugins (I use btheado’s opinionated plugin along with the official browser storage.)
The advantage here is that you don’t have to worry about losing work from a disconnect. I believe (but couldn’t swear) that I’ve gone all day without a save and was still able to work. It’s only when you reload or save that you have to re-connect. There may be other things that trigger your TW page to check if you’re still online, but I’m not sure what they are.
Just brought up a version of a TW from TiddlyHost on a phone without logging in from nearly two weeks ago! (Firefox, but presumably similar for Chrome). This will depend on how your particular device handles memory, but it’s interesting that it can do it at all.
Timimi is an approach to tiddlywiki saving that is in my mind superior to all others for local file:// single file wiki’s some months ago this was slated as an impending problem. All it needs is someone familiar with browser plugins.
Lets face it if we want deep integration with our computers we need this “middle wear” to be built and maintained by this community, we are not dummies, we can work this out.
TiddlyServer was a good alternative to Timimi before it came a long.
Timimi is an approach to tiddlywiki saving that is in my mind superior to all others for local file:// single file wiki’s
… until you try to use it in Ubuntu. Which has browsers installed via snap. Thus making Timimi’s installer assumption (that browser files are supposed to be in ~/.config) false. This is my first suspect why I couldn’t make it work.
This is where when a browser middleware is needed we need a version for each browser/OS and maintain it as a community. It simply needs to be done and such issues addressed.
I would like to add the functionality that Timimi does for TiddlyWiki is available in other services PWA’s and applications/sites such as implemented by Google and Microsoft to name a few. I do think sometimes they don’t honor these methods we use to save locally, because they know their way around it and happy to force the alternatives to go through these regular breakages, all in the name of browser hardening. They are using their upgrade release cycle to break solutions that are not inside there community and revenue streams.
I’m really hoping Snap will go away. It blocked my password manager, so I was already looking for another solution and never realised it also interfered with Timimi.
Background: When you install Linux you also get a package distribution system with access to thousands of applications, such as Chrome. Recently The Ubuntu brand of Linux changed the way applications are installed, so that each one has all the dependencies it needs, but is also isolated from all the other applications. Maybe this analogy is broken, but imagine every application came in it’s own Docker container. After that, extensions that needed access to resources – executables – outside the installation container were broken.
What I did was to uninstall the distribution’s version and instead install the version from Mint, which hasn’t done the Snap thing. This also means I have to do manual updates.
I think the way other apps have gotten around this issue, and the Manifest 3 issues, is to basically run a little server. Then the extension communicates with the server rather than directly with an executable, which might be a security risk. I’m thinking of Joplin and Zotero. Hmm. Probably Obsidian too. All of these have some concept of where their data is located. For TW to have a similar solution, the user would have to define where their data lives, rather than the free-for-all we got used to with Timimi.
So running Rclone basically does the same thing – runs its own server, working from a user-defined data source (directory).
