Screenshots of the new web-based GUI

Just wanted to share with you two screenshots of the new web-based version that I have been working on.

The first one is the storages page with mocked data:

The second one is the dialog to add a new storage. A new feature that is already working for all supported storages is the icon buttons next to the Bucket and Directory inputs, which allow you to select a bucket or a directory from the list.

15 Likes

Uhoh, that looks smooth and beautiful! :star_struck:

For add storage you shouldn’t have the “more button”. That looks awkward: why are the other storages less important than the first ones?
I think just having a larger (wider) dialog should be fine, as i also imagine that different storages have different heights (because different parameters are needed for setting them up).

Looks really great! Love it!

Do you really mean “web-based”? Or do you mean browser based?

Loved it! :star_struck:

I especially liked the graph that shows the size “evolution” and the “check” status!

1 Like

@gchen looks awesome!

out of interest, will you be using output from existing commands or do you have a new version that exposes some api endpoints e.g. json api spec?

thanks

This looks very promising!!!

Is there an ETA on this or a rough expectation on when something will be released? I would be interested in testing it if needed :smiley:

1 Like

soon

1 Like

Previously the target was the beginning of September, but this seems a bit unrealistic due to a busy summer. I’ll try to get something out by the end of September.

@iluke the new GUI version will just parse the output from the CLI just like the old GUI version.

2 Likes


When you want to reply to someone, you should reply to that particular post (button 1). You can reply to posts above the “last one”.
When you want to reply to the topic, not to someone specifically you should use the reply button #2.

For example, right now i replied to your particular post (using the reply button above @towerbr’s icon) as you ca see here:
image
and also here (after i replied i did another screenshot and then edited this):

You can also change the reply type if you press the wrong button:
image

3 Likes

Isn’t this the same as the current GUI version in which there are performance problems when listing all the files belonging to a revision? (because parsing a string takes extra time after the command finished?)

I was thinking that you intend to directly call the methods used by CLI, instead of calling CLI and then parsing. :confused:

Parsing a few MB of text should not take a significant amount of time as it’s not a lot of work for a System sold in this millenium.

I guess if the old GUI was slow it was probably because some GUI framework related bullshit issue caused by OOP focused devs with no clue about writing efficient code. My bet is that updating a List/TreeView is the real bottleneck of the old GUI - well or maybe memory management, which is the next typical issue as soon as you reach tens of millions of allocs …

IMO I think it’s a good idea for the GUI to be only a frontend for the CLI version as this way, there’s always only a single duplicacy.exe doing the actual work. And if there’s some bug/problem/issue this is the single one you have to look at.

1 Like

Exactly. The TreeView in wxWidgets doesn’t support on-demand expanding of folder nodes and most of the time is spent in loading all data into the TreeView. Maybe there is way but I would rather spend time working on the new GUI.

This is exactly why I stick with this model. The CLI version is always there to rescue if there is anything wrong with the GUI version.

7 Likes

<pedant>“web-based” means that it uses HTTP (the protocol that the World Wide Web uses). While it’s designed to be used by humans using web browsers, there is also the possibility of using other software than can talk HTTP to interface with a web-based system.</pedant>

So I’m assuming it’s not really “web-based”. IMHO “web-based” means that it uses the world wide web (the internet, cloud based, SaaS).
I’m assuming it’s running local on my computer, not on the web. It uses HTTP like the Duplicati GUI is doing.

Yes, it uses HTTP, which means that, so long as you have a connection, you can access the GUI from anywhere in the world, using the WWW.

Looking at it another way, the GUI is built as an interactive web site. You can use any web browser you like to view it. It’s based on web technology, not a specific web browser.

Anyway, we’re splitting hairs. What GChen means is that the control GUI will be a web site. Which is nice, as web technology is available on almost any type of computer.

The webui looks great!
What is the updated ETA?

Update: I think I found my answer here:

Excellent, I have been using the CLI version on 3 headless linux VMs for over a year. These have been setup and basically left, whenever I need to change anything or do anything I have to reacquaint myself with the Duplicacy commands because there was no frontend available to manage these and don’t need to type the commands often enough to remember them. I was just the other day thinking about bodging a frontend of my own to solve this problem, now don’t need to. This does lead me to a couple of questions though.

  1. Can this read my existing config, or will it shortly if not yet?
  2. What will the pricing be, presume its not free for personal use like CLI version. Happy to pay, but how will it be billed. You mentioned one web install can eventually manage multiple Duplicacy installs, would it be per machine backed up or one off fee for Web version?
1 Like

It doesn’t read existing configs and this is unlikely to change.

I plan to stick with the current policy so the new web GUI will be completely free if you only run check, copy, prune and restore with it. That means, if you only use it for checking the storage that the CLI backs up to from other machines then you don’t need to pay. When you do use the GUI to run backups, it will accept your current per-machine GUI licenses (both personal and commercial should work the same way).

4 Likes

Assuming that this also includes existing CLI configs, could you elaborate on why you chose to keep those separate? I thought the new GUI is built as a wrapper around the CLI version…

3 posts were split to a new topic: Run web-ui in a docker container?