/hydrus/ - Hydrus Network

Updating from 336 to 339 worked mostly fine, but when attempting to open a random file (not that large either), after running in ~15 minutes, it crashed with an segmentation fault (SIGSEGV). Running source on antergos linux. I'm guessing it has something to do with the wxpython upgrade some updates ago.

Tabs are now failing to close.

AttributeError
'PagesNotebook' object has no attribute 'TestAbleToClose'
  File "include\ClientGUIPages.py", line 1782, in EventMiddleClick
    self._ClosePage( tab_index )
  File "include\ClientGUIPages.py", line 907, in _ClosePage
    page.TestAbleToClose()

Is it a known issue that webm downloads now get ignored with "Unknown filetype!" error?

Here is an example file url that produces the error when I paste it into a url import tab: https://chan.sankakucomplex.com/post/show/7525386

I'm running the Linux binary on Ubuntu 18.04 with KDE Plasma Desktop. (More about the minor issues that causes later, lol.)

>>11556

>>11552

SHIT

Thank you for this report. I will have this fixed for next week. If this is very inconvenient for you, there are no db updates in 338->339, so it is possible to roll back to 338 this week.

>>11554

I am afraid this works on my machine here. Can you post the full error trace (should be right-click->copy notes in the file import status window)? And what version of ffmpeg do you have under help->about?

Do any mp4s work, or just webms are broken? If you have an mkv to test with, is that ok?

If you download the file to your desktop using your browser, open a terminal and go:

ffmpeg -i (that.webm)

What does it say?

>>11557

I got pretty blackpilled about this 'page of pages' bug and the shit from 338. Little bugs keep happening due to normal mistakes. I can't expect to get rid of stupid errors by wishing them gone, so I am going to write a proper ui test suite to make sure this common stuff typo-breaking doesn't slip into releases any more.

Keep the reports coming, thanks lads.

>>11549

Thank you for this report. I am afraid there aren't many super easy answers here. There seems to be some instability remaining in some Linux flavours due to some part of window manager, GTK, wx, or my code not working well together. My best advice is to try running from source, but that isn't working for everyone either.

When you installed wxPython, did you build it yourself or get a packaged wheel, like this?:

pip3 install -f https://extras.wxpython.org/wxPython4/extras/linux/gtk2/(best match for your linux) wxPython

I had bad luck with GTK3, so I build for 2 now. I know some window managers cause more or fewer problems as well. wxPython 4.0.4 came out last month, iirc, so if your wx is older, you might like to try getting the new version.

If you can pin down behaviours (in the past it has often been specific dialogs opening and doing a certain action) that seem to increase the likelihood of certain crashes, happening between 1-60 mins before the crash, that would be helpful to me in narrowing down any objects I am not destroying or handling correctly. I had thought I eventually fixed all these in the long time ago wxPhoenix update, but there could still be a couple more out there.

>>11550

Thanks, I am glad you like it!

The current in-client video player is all mickey-mouse shit I threw together. I don't have any experience with audio, which is why it doesn't do any yet. I would very much like to add it, if it is technically reasonable for my python/wx situation, but the project will be a 'big' one. I am currently working on the Client API as my big job, and I'll add OR searching after that, and then I'll throw up another big list of big jobs to work on next for people to vote on.

Can you talk more about the image viewer bugginess you've had? It might be something just you are getting, or something I haven't noticed yet, and that I can fix quickly. Although some stuff like zooming in to big files is janky and memory laggy, and again will likely have to await another 'big' job so I can rewrite the rendering pipeline.

>>11551

Thanks!

If you end up doing tag select followed by other actions, let me know how the workflow goes. I've been thinking there isn't a good connection after the select step, so you get "which thumbs got selected? oh, right, there, now I'll right-click on it". When the number of files selected is 3/700, it is unreasonable to chase them down for an action. Iirc, there's a default shortcut of ctrl+m for 'put focus on media (thumbnail) panel', so if you want to remove, you might get away with 'select files with x' and then ctrl+m, +r for a quick action.

>>11562

I am not sure about doing them all at once across the program, but I will start on column width memory and selection memory once I am finished with replacing the old list objects still in the program. I have 17 left to do. I am sorry for the annoyance–I find the bad auto-widths frustrating as well. I haven't had a good time with the stock wx ListCtrl.

Advanced tag operation, copy current and pending tags from local tags to PTR doesn't seem to do anything. What gives?

>>11557

"Unknown filetype!" is all it shows when I select copy notes.

I tried this mp4 file and got "File did not appear to have a video stream!":

https://chan.sankakucomplex.com/post/show/7530098

Now here seems to be the real problem. The about window says "FFMPEG: unknown".

The version I have installed should be the default from the ubuntu repository. When I run it in the terminal it says "ffmpeg version 3.4.4-0ubuntu0.18.04.1".

When I check the file with it, it says:

"Input #0, matroska,webm, from '1c148da01b4512090e9e7e0a54f27789.webm':

Metadata:

creation_time : 2019-02-07T09:14:03.000000Z

ENCODER : Lavf57.83.100

Duration: 00:00:10.05, start: 0.000000, bitrate: 3031 kb/s

Stream #0:0: Video: vp8, yuv420p(progressive), 1280x720, SAR 1:1 DAR 16:9, 14.93 fps, 14.93 tbr, 1k tbn, 1k tbc (default)

Metadata:

ENCODER : Lavc57.107.100 libvpx

DURATION : 00:00:06.040000000

At least one output file must be specified"

Looks as expected, I assume.

>>11548

Not sure whats the deal but 399 has a major performance drop from 338. The program was 1.2GB Ram for the same thing as 300MB Ram in 338 and it was laggy. I rolledback back to 338 and everything's normal again performance wise.

It did give me a warning about the db being 339 but program as 338 though.

>>11548

Question, does the db backup also backup subscriptions and the like? If not then how would I go about that?

>>11576

Thank you for this report. I will check this today.

>>11579

Thank you. I assume you are using my built release? The way PyInstaller packages/freezes the build means the environment and PATH of the process is a little odd under Linux. In order to call your system's external ffmpeg, I have to move some things around for the child process call.

Did this used to work for you? Maybe changed with the py3 release in 335?

If you hit network->data->manage upnp, do you get a similar problem, or does it (after a few seconds) load your upnp mappings ok? (I am assuming here your router is happy with the mapping call, but even then I think you'll get an error saying like "Router doesn't support upnp" rather than "Weird shit happened with the upnpc executable".

If that works, then it looks like your client can talk to exes with fixed paths, so the temporary solution here is likely to just download a Linux static build here:

https://johnvansickle.com/ffmpeg/

And literally put the 'ffmpeg' executable in your install_dir/bin folder.

Whether that works or not, I have written a new ''help->debug->report modes->subprocess report mode" mode to report more information about your path and subprocess arguments for tomorrow's release. Please attempt to do an ffmpeg import in v340 and see what environment info it gives. This may include private info, so please blank out personal stuff or just email it to me.

Furthermore, do you know where your system ffmpeg is? Is it in /usr/bin, /usr/local/bin, or somewhere else?

>>11579

Oh yeah, and that ffmpeg output is correct. Everything is fine there–hydrus just can't find the exe atm.

>>11583

Thank you for this report. Yeah, mine is running about 800-1GB after running all week, but a restart only needs 273MB. There's nothing that obviously stands out to me that I've changed, memory-wise. Maybe the menubar stuff. I don't think I even updated any libraries.

Is the memory bloat you see from a fresh boot of 339, or does it grow over time? I forgot to try this myself on my above restart, but if you hit help->debug->data actions->clear image rendering cache, does that cut much from it?

Also, can you try hitting help->debug->data actions->print garbage? It'll dump a whole load of shit to your log under your install_dir/db directory–scan what it dumps for privacy and then please post it here (or email it to me). Now I am in py3, I'd like to add a proper memory profiler this year, but until then, the garbage log may highlight something.

>>11596

After a bit more poking around, I think regular garbage collection wasn't running for some reason. I'll do some more work on this today, it may be related to another issue.

>>11587

Thank you for this report. Just to check, this is non-Windows, right? Did you happen to migrate an install from Windows? I have had the worst luck trying to get the floating version of the autocomplete to work on non-Windows for these sorts of focus issues, so Linux and OS X installs start with it always in embedded mode, like how the 'manage tags dialog' works.

Please hit options->gui->always embed autocomplete dropdown results window and let me know if that works better.

>>11599

Oh, and to add to your question: I indeed have migrated from Windows long time ago. I just tried a fresh install and it defaults to that setting, so the problem was just outdated defaults. If I wasn't pantsu-on-head retarded, I would've checked if it was working on a fresh install first. Sorry for wasting time.

>>11603

I reinstalled 339 and havent noticed the memory spike as bad as before but it is locking up and stops responding pretty often so i still end up having to restart anyways for any downloading. the program still works for main archival and search though.