On Windows with Firefox, the overall start 'Fold'ing button doesn't start Folding
Hou5e opened this issue · comments
Win10-FAH v8.3.0 (Debug), Builder file 113 from 2023-12-30.
In Windows, pressing the overall start 'Fold'ing button doesn't start Folding (It works on Linux). The RG control works though.
{Edited} --> On Windows, it doesn't issue anything for the overall start 'Fold'ing button, just show log to view nothing:
08:59:46:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T08:59:46.666Z"}
On Linux, it issues this, shown in the log:
�[92m12:33:16:D3:msg: {"state": "fold", "cmd": "state", "time": "2023-12-30T12:33:16.467Z"}�[0m
The last few lines were from pressing the Start Folding RG control:
08:59:46:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T08:59:46.666Z"}
09:02:19:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T09:02:19.124Z"}
09:02:40:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T09:02:40.816Z"}
09:03:09:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T09:03:09.047Z"}
09:04:00:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T09:04:00.223Z"}
09:04:20:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T09:04:20.876Z"}
09:04:38:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T09:04:38.374Z"}
09:04:49:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T09:04:49.916Z"}
09:04:55:D3:msg: {"state": "fold", "group": "", "cmd": "state", "time": "2023-12-30T09:04:55.785Z"}
09:04:55:D1:Default:Remaining CPUs: 0, Remaining GPUs: 1, Active WUs: 0
09:04:55:I1:Default:Added new work unit: cpus:0 gpus:gpu:04:00:00
09:04:55:D3:Default:Unit enabled
09:04:55:D1:Default:Remaining CPUs: 0, Remaining GPUs: 0, Active WUs: 1
Same with a RG starting the next WU:
10:47:47:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T10:47:34.228Z"}
10:47:47:D3:msg: {"state": "pause", "group": "/RG1", "cmd": "state", "time": "2023-12-30T10:47:43.027Z"}
10:47:47:I1:Loaded cores/openmm-core-23/windows-10-64bit/release/fahcore-23-windows-10-64bit-release-8.0.3/FahCore_23.exe
10:48:10:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T10:48:10.089Z"}
10:49:11:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T10:49:11.584Z"}
10:49:29:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T10:49:29.751Z"}
10:50:11:D3:msg: {"enable": false, "cmd": "log", "time": "2023-12-30T10:50:11.727Z"}
10:50:31:D3:msg: {"state": "fold", "group": "/RG1", "cmd": "state", "time": "2023-12-30T10:50:31.178Z"}
Note that 08:59:46:D3:msg: {"enable": true, "cmd": "log", "time": "2023-12-30T08:59:46.666Z"} is just web control enabling the log. It will happen every time you open log view and has nothing to do with pause/fold.
I suspect there is an error in the browser console when you click Fold on Windows.
Yes, you're right, I pressed the show log button after pressing the main Fold button many times, only to not have any entries in the log for starting folding. I was logged-in and using a Firefox browser.
I'm guessing: this issue, #186, and #187, are related to the last 2 sentences in this post: #178 (comment) about the web socket not reporting status.
Testing this again with the latest Windows (Debug) builder file 116, using Firefox, I get the same behavior:
If logged out, then Pause/Resume works correctly locally.
When logged-in, then you get nothing pressing either the main Fold or Pause buttons, locally or remote.
I just tested this in Windows with Firefox, Chrome and Edge. All three worked as expected. Is this still a problem?
Yes, it still happens for me (with the latest Windows Release build 57 or Debug build 117). The requirements to make this happen are: Use Firefox (use a private window to avoid cache issues) on Windows, Login to your FAH account. I get this now about the certificate, and I press OK:
Press the Fold or Pause buttons. I get no responses. It also doesn't connect to the remotes.
In this screenshot, the same Windows PC is the 2 left browsers, and Linux is on the right. I'm pressing the fold button many times and it's not doing anything.
It sounds like the setup of 'SubjectAltName' in the certificate is a common issue about self-signed certificates. Are either of these links related or useful about this problem?
https://stackoverflow.com/questions/7580508/getting-chrome-to-accept-self-signed-localhost-certificate/57511038#57511038
https://serverfault.com/questions/1115557/generate-self-signed-certificate-for-firefox-to-accept
The certificate problem should be fixed now. Does it also resolve this issue?
This is working now.