- Mar 22, 2020
- 4,254
- 10,359
- Mar 7, 2019
- 1,408
- 1,205
i have already explained it on your github issue myself, FaceCrap offered you more insight, and it is well explained in the FAQ in OP as i already mentioned. i can go on repeating the same explanation but that would be counter productive for both of us
- Mar 22, 2020
- 4,254
- 10,359
Well the behavior that I'm seeing is fucking weird, even with the FAQ. I'll pay closer attention and if i can figure out what's happening and write something coherent up I'll come back.
- Mar 7, 2019
- 1,408
- 1,205
the tool is made for users managing games theyre playing and keep track of what they have installed and played and so on. i understand that this isnt exactly what you use it for, but youre the minority here then and i cant change the behavior yet again just for you. if you explain better what exactly is your workflow i can suggest a way to use the checkboxes in a way that makes sense for you.
from what i understand you do game ports and modding right? i would imagine you need to download games to do these things. here is what i imagine could work:
- download an update, then mark as installed (checkbox fully set), this will make the finished checkbox half set
- port and mod the version you just downloaded, then mark as finished (checkbox fully set), now youre done with this game until next update
- when an update happens, the installed checkbox will be half set, as you do not have the latest one installed
- again from the start: update it, then mark as installed
- now finished checkbox is half set, because you havent made ports or mods or whatever for this new version you just downloaded
- port and mod this update, then mark as finished, done until next update
you know which games have been updated by the half-set installed checkbox, and you know what you still need to port/mod by the half-set finished checkbox. you cant port/mod the games without downloading, so first step is downloading the update.
if you dont usually keep games installed after youre done modding them, then just delete the files after marking as finished. the tool will remember you marked it as installed and finished for that version, regardless of what files you have downloaded, thus it will still tell you about an update with the installed checkbox being half-set.
if this is not the kind of workflow you need then ill need more information to understand how to help you use it effectively
- May 20, 2024
- 168
- 193
11.0.1
The game Hot Sand of Antarctica was marked as abandoned back on wednesday, but the tool still hasn’t updated its status.
The game Hot Sand of Antarctica was marked as abandoned back on wednesday, but the tool still hasn’t updated its status.
- Mar 7, 2019
- 1,408
- 1,205
new issue: I went to start a Unity game in linux with this and got this message:
Error - KIO Client:
For security reasons, launching executables is not allowed in this context.
I can run it manually. I get this with all native unity games launched form F95checker 11.0.1
Error - KIO Client:
For security reasons, launching executables is not allowed in this context.
I can run it manually. I get this with all native unity games launched form F95checker 11.0.1
Last edited:
- Oct 1, 2020
- 1,523
- 1,050
And of course, you went straight to google or your fav search engine to find out that you probably need to right-click the Unity executable and probably need to enable "Allow executable" and maybe even "Allow launching"?
At least that's what I got when I did that. Not being a *Nix user I know squat about this, but maybe it means something to you...
Mind you, it also seems to make a difference if it is an "appimage file" or "appimage executable"... One is a native *Nix executable, the other must be ran with wine, proton or whatever.
Any *Nix'ers here that could elaborate on this and tell me if I hit the mark with this?
Last edited:
- Mar 22, 2020
- 4,254
- 10,359
This is what I'm trying to do but updates are coming in and the state of the checkboxes is not what I expect it to be from the FAQ. Again, I'll see if I can capture anything concrete to report back.
- Mar 7, 2019
- 1,408
- 1,205
that should be a notification by kde daemon, not exactly sure why but from what i recall its some kind of permission where an executable is not allowed to launch other executables (as malware often does things like this). the executable of the game should already be marked as executable (the "allow to run as app"), as this is a simple permission flag and the checker will enable this flag on executables before launching *if it detects that they are indeed executables*. it does this by checking the shebang (#!/path/to/interpreter, its used in scripts) or for known exe and msi magic numbers. what executable did you select exactly?
- Mar 7, 2019
- 1,408
- 1,205
simple_human rentalunshipped982 M0narh BrockLanders ascsd blackop Nerro MaxTheEro
thanks to FaceCrap finding
try build 1406 please
basically there is a bug where imgui can return corrupted sort specs if you check them when the specs arent dirty (havent been changed by user). i was doing this because changing tabs with "independent views" option would cause to have different sorting, but would not trigger the dirty flag as its technically just a different table, for which the specs did not change. now i only get new sort specs when dirty flag is set, and save them. meaning when changing tab it just uses the specs it had saved, not ones reported by imgui. and when you change sorting for a given tab, it saved the new specs for it for later use.
the imgui bug has been fixed, but pyimgui is on an older version. the usage i had before was supported by imgui, but not ideal. regardless this new system should be more efficient, and also avoid the bug.
thanks to FaceCrap finding
You must be registered to see the links
, sorting bug might be fixedtry build 1406 please
basically there is a bug where imgui can return corrupted sort specs if you check them when the specs arent dirty (havent been changed by user). i was doing this because changing tabs with "independent views" option would cause to have different sorting, but would not trigger the dirty flag as its technically just a different table, for which the specs did not change. now i only get new sort specs when dirty flag is set, and save them. meaning when changing tab it just uses the specs it had saved, not ones reported by imgui. and when you change sorting for a given tab, it saved the new specs for it for later use.
the imgui bug has been fixed, but pyimgui is on an older version. the usage i had before was supported by imgui, but not ideal. regardless this new system should be more efficient, and also avoid the bug.
Wow thanks
, do you want it to be tested? I will report back if the problem still persist
- Mar 7, 2019
- 1,408
- 1,205
that would be great, thanks. no need to spend hours testing, just install that beta build and use as normal, just keep an eye out to see if it happens again, hopefully it wontWow thanks
Any unity, but here is an example: world of sisters:
World Of Sisters 0.26.3 (Linux).x86_64
- Aug 21, 2020
- 1,272
- 1,799
app need auto memory cleaning or something
started app (GPU memory around 10% used) -> hover over list of video collection threads (their banners almost always animated) -> GPU memory skyrockets with every hovered thread (with 100+ threads i have in list it reaches 92% used out of 8gigs) -> it stays like this till app is closed, holding all this memory for itself, even in BG mode
started app (GPU memory around 10% used) -> hover over list of video collection threads (their banners almost always animated) -> GPU memory skyrockets with every hovered thread (with 100+ threads i have in list it reaches 92% used out of 8gigs) -> it stays like this till app is closed, holding all this memory for itself, even in BG mode
- Dec 19, 2020
- 443
- 1,559
I suppose it may be counter intuitive in the sense of reducing load on the forums but is this something that would we ever be able to run locally rather than relying on another party for?
Secondly, if we're receieving cached data from your cache server and comparing it with ours, why does the program require signing in? This seems redundant, I personally don't care about the notifications from this program, telling me someone replied to a thread, etc. That could be done away with, it has no impact on the purpose of this software for me which is keeping track of games and that's it.
Worst case scenario would be we're passing our account info onto your cached servers and you have a copy of it... obviously this could be checked by going through the source code. It wouldn't be the first time someone has done something malicious despite it being available in open source code, after all someone has to go looking for it in order to find it. I'm not saying you're doing that, I'm just highlighting an example for others.
So I guess my question is if we're receiving cached data from your server, is there any plans to be able to remove our account from this program and keep it focused on it's main task which from what I can see, requires no account at all.
- Aug 21, 2020
- 1,272
- 1,799
- Mar 7, 2019
- 1,408
- 1,205
This is precisely how it was until last month. This program singlehandedly caused multiple downtimes for f95zone. It's basically a DDOS. Now it isn't anymore.
If you mean self hosting your own cache api, technically nothing is stopping you, it's open source and even in the same repository. Only dependency is redis and the xf_user cookie in the env file. But of course this defeats the whole purpose, as we go back to multiple people fetching data from f95zone with the same objective and in the same way for no reason. Also, Sam helped me out by disabling some rate limits on latest updates and the version check api, so you wouldn't be able to run the indexer api as it is (mostly just changing the version check api limit from 1000 to 100), and would encounter some rate limits when getting precise data from latest updates if more than a certain amount of requests in an hour.
The idea is that Sam has some more lenient rate limits for my cache api and that it's the only one, so on its own it's more stressful on the server than 1 single f95checker user in the past, but means that it's only this one thing, not thousands of users all doing it at the same time.
As stated in the changelog and readme, this is no longer the case. It can be used fully without an account, and it is so by default as notification checking is disabled for new users from now on.
The limitation is mostly download links. To prevent links being taken down, I obviously cannot include naked download links in the cache api. It only includes an indication of where to find the download link in the thread, which then the checker on your machine will retrieve from the f95zone thread, which of course requires an account. This is true for regular, masked, and donor ddl links. Rpdl torrents require their own account, but not an f95zone account.
You can see in modules/api.py that both fast_check() and full_check(), which are the only methods connecting to my api, use cookies=False.
Last edited:
- Mar 7, 2019
- 1,408
- 1,205
Unfortunately that's how it is, banner images are very high quality and the program allows you to zoom a lot into them. With the imgui paradigm this means sending the full resolution texture to the gpu. You can compress the images on disk for yourself to lower both disk and gpu vram usage, and I plan to implement something like this in the tool eventually but it's low priority. Also I hope it's just vram you mean, it used to handle images a bit weirdly in the past, it would keep the texture data in ram if images were shown for a moment then hidden, now it properly passes to gpu.
but yeah keeping all images in vram is a bit of a problem. Not sure what is the correct way to handle it. Loading them is slow, and just having them unload when they go off screen seems like it will just make it horrible to use as it keeps loading images all the time.
As always cache invalidation is the unsolved computer science problem