I recently subscribed to a web service that uses various image generators to create online comics and they have been making a move toward making things work better for image to video to make anime. Their services are...inconsistent.
I daily with Waterfox (a Firefox fork) but in attempt to make this stuff work I've installed Vivaldi, Brave, and Google Chrome...they have the same issues. I really don't think it's my browser. And I've told them that.
Now as you can see they're blaming it all on me being on Linux. Now I think I know the answer to this already, but just in case...is there any reason why a web service would have functionality issues because of Linux?
IMO, web application/services endpoint is a web client (i.e. browser) and not an OS itself. That is the real question is whether a multiplatform web browser behaves differently on different OSes.
Well some browsers are/were tightly integrated into their respective OSes and are/were not portable due to proprietary OS-specific extensions to the common web standards. A number of multiplatform browsers are not locked to the particular OS and are intended to provide the same functionality on any platform.
My experience with Firefox tells that web services which list Firefox as a compatible browser work flawlessly on any platform.
Technically, web service can retrieve a platform from browser and behave differently including cutting off Linux clients. But there is no need to suspect a conspiracy of any kind. Actually, it enough that developers find out that their service works unstable on Linux clients. Or it is convenient to blame Linux in the absence of other ideas. Period.
IMHO such a situation is server-side bug or feature and not client-side problem.
What sort of web service is it? If it's using Microsoft technologies it may be that your Linux system is lacking the "hooks" that are needed, but these days even .NET capabilities are available in Linux. You may need to install a component or two, but the developers would have to tell what those are. Then you have to decide if you want to pursue them.
If you have a non-Linux computer (I know, heresy!) try the same web service. It's not likely you can trace it as easily (I think you can launch a browser from the Linux command line and view the output you might not otherwise see), but it would give more credence to Linux being the failure point.
I just tried it on a Mac using Arc, and using Brave on my Linux machine. I don't have an account, so I couldn't generate anything, but it seemed responsive when I clicked around. Both my Brave and Arc servers are at the latest release (Brave just updated).
You mention issues, but not the "nature" or "aspects" of those issues.
Maybe if you shared some specifics or details relating to the issues in question would trigger an "Aha!" moment from the more knowledgeable in out Community.
Sorry I thought it was clear from context of the quote in the image of the first post.
As I said it's an image generator. (They also have some video generation which is also giving me errors but for this discussion I'm sticking to image generation.) I try to use the GPT image generator because it and the Gemni image generator is supposed to be able to support the webapp's character creation tool and churn out images of created characters made with their system. Instead I am getting errors.
The people who manage the webapp are saying it's something related to my configuration that is causing their service to not work, and try demonstrating this by posting images they claim are made with that image generator.
Not perfect but mostly creating an image with created characters.
Does that clarify things now?
Personally I don't understand how my choice of operating system could cause this issue. Or my choice of web browsers which are consistently giving me the same errors. But the people at the AI are shrugging and saying 'Works for me' which is not useful.
Please bear with me (trying to visualize steps/breakdown without the actual software):
Can you load the App with no browser console errors/warnings?
Can you select an image provided by the App Site with no issues (again, any browser console errors/warnings) ?
Can you interact without issues with a Site-offered image (ability to "define task", specify transformation options and have those accepted/recorded/saved) ?
Can you direct the App to perform the specified transformation and see that it recognizes your requested action (visual feedback that it is attempting to perform the action)?
Is the App/Site reporting a failure to perform the requested transformation action? or is it crashing? If reporting failure, is it reporting anything that indicates the nature of the failure during the attempt, or is it simply coming back with no details, just offering to have you define a new request (as if it had amnesia of all that previously happened and is trying to initiate the interaction for a new request, as if nothing had been previously attempted)?
I just want to underline the need to review if there is any feedback in the browser's debugging console for any of the above steps. You should be able to set that up as a detached window and, like me, have the bottom of the console visible below the border of your browser window, for quick click-focus to see the messaging, if there is any.
I hope my above "visualizations" are clear and specific enough for all the steps that might be involved in your failed attempts.
Since I am butting heads with this for my little game project, I have to ask if you have control on the properties, and value assignments, related to the "meta" header similar to the following:
I used my google account, it generated a character after picking up attributes, male, fair skin, beard, etc. Since there is a discord channel (top, right), maybe join discord and ask there.
I am far from a JavaScript expert, but if Pavlos was able to get results from his own UM-based browser ( @pavlos_kairis , which browser was that?), then it must be something having to do with the configuration settings of your browser, as the message "ERR_BLOCKED_BY_CLIENT" is indicating.
For that, I suggest clicking on the link at the end of the line to see where it takes you (either help file, or line in code that is failing). That's the best I can offer to you.
However, armed with these images, you should be able to get more specific assistance from the App developers.
The "http-equiv" statement would be embedded into the application's web page header, and controls what is allowed/disallowed regarding various local-related code/actions.
You cannot edit/modify that, but it informs you as to what the App is allowing or not.
"Transform" is what the software is doing, using your input images as a start-point, then generating a desired output, based on your specifications (as Pavlos apparently did). If it helps you to better understand, it is not exactly the same, but similar, replace "transform" by "morphing" (note that "morf" is the end of "transform" but with the letters in reverse.
I guess I am a little lost, I was able to generate characters with my laptop running UM with no problems. I am using Brave as the browser and I am not having any issues. Where are you having the issues?
It worked on my UM 24.04 with Firefox; that was even with uBlock matrix running and standard Firefox protection settings.
The only javascript it needed was from the firefox.app domain, images came via replicate.delivery.
Console caught ~11 errors on image generation. Some were due to blocking CORS, a few font errors, and some React errors (418 and 423).
I haven't found any means to remove an account yet; so I would not recommend people test this app out without solid privacy protection.
There is no business or contact information provided on the English website. The domain is leased and hence the true owners of the website are hidden. The whole thing reeks of very poor transparency, and I'd trust it about as far as I can throw it.
