/hydrus/ - Hydrus Network

>>9775

Yes in options > files and trash

Once you start a dupe search at a selected distance it's not possible to stop it despite there being a stop button because the entire program locks up and stops responding. :(

Got this error then the whole client froze

wxAssertionError
C++ assertion "IsRunning()" failed at ..\..\src\common\evtloopcmn.cpp(83) in wxEventLoopBase::Exit(): Use ScheduleExit() on not running loop
  File "include\ClientGUITopLevelWindows.py", line 479, in EventOK
    self.DoOK()
  File "include\ClientGUITopLevelWindows.py", line 667, in DoOK
    self.EndModal( wx.ID_OK )

>>9762

>>9761

I am sorry for the awkwardness of the new listctrl. I'd like to make these sorts of improvements, but it will have to wait for the rewrite from the old listctrl (which crashed on sort, so we are getting better) to be finished first.

>>9775

>>9779

Thanks, yeah–those media viewer hover window links don't use that option yet, but I will be fixing them in the coming weeks. The right-click->known urls->open menu does work with the option for now.

>>9781

Thank you for this report. I will see if I can insert a better stop check in that.

>>9782

And thank you for this as well. I have seen this a couple of times before but have not been able to pin it down. Can you remember what you were doing at the time? Had you recently closed a dialog, or were attempting to? Do you have a 'busy' client with say dozens of pages and downloaders all working at once?

What are the chances of ever getting an RSS-feed-like thing for certain imageboards that shows all threads containing certain user-set tags/strings which you can then quick-open a watcher for?

Hi, I'm having some issues with webms on this version. When trying to import:

>Unparsable file

when closing the import dialog after trying to import a webm the client crashes with

>double free or corruption (out)

>Aborted (core dumped)

when trying to view a webm in the client

>Exception

>Unable to render that video! Please send it to hydrus dev so he can look at it!

>Traceback (most recent call last):

> File "include/ClientRendering.py", line 370, in THREADRender

> File "include/HydrusVideoHandling.py", line 767, in read_frame

>Exception: Unable to render that video! Please send it to hydrus dev so he can look at it!

when trying to open externally:

>mpv: relocation error: /usr/lib/libssh.so.4: symbol gcry_mpi_ec_decode_point version GCRYPT_1.6 not defined in file libgcrypt.so.20 with link time reference

I posted before about issues with liblzma so maybe this is another similar library issue.

>>9789

Thanks. The layout of the ratings bubbles here is actually a tricky one–the ones on the hover window are actual legit OS GDI objects that get laid out like any other button or text widget, but the ones on the media canvas are just shapes I draw on a bitmap using my own +2px pseudo-layout shit (which is also why they are un-clickable, if you ever manage to get your mouse on them before the hover window pops into place). Depending on some OS layout and dpi/zoom preferences and all that, these layout calculations can differ a bit, and I guess that is extreme in your case due to the number of rows you have.

Which OS version are you on, and do you run 100% dpi zoom or something else?

>>9791

Perhaps in a future iteration on the downloader engine, but not this pass. I've completed most of this rewrite's objectives, which was to convert the legacy system to something user-editable and saner behind the scenes, and don't have time to write any new custom ui. If you don't have a workflow already, I recommend you set up a system where you check the catalogs of the boards you like every x time units and do some manual text searches or pick the top y threads you like in the first three pages and just drag and drop on the client.

I am not familiar with the various imageboard add-ons for browsers–and I think there are other thread-watcher apps, although I don't remember the names–so maybe one of those can provide you somewhere with a subject/url summary list of the boards you like? Again, if you can just isolate the text and either copy/select, it can just be one more click to get the thread into hydrus.

Or, looking at it, it look like you can just plug this rss into a reader and you might be off:

https://8ch.net/hydrus/index.rss

Or if your browser can render json nice:

https://8ch.net/hydrus/0.json

>>9795

Something about PyInstaller–the program I use to freeze an executable python environment–is funky on some flavours of linux, and the cwd or shell environment or something just gets messed up and the program can't run/find other exes well. I am still trying to pin it down, but I am not a linux expert by any means.

Since you are also getting crashes, I recommend you try running from source. There is more info here:

>>9768

https://hydrusnetwork.github.io/hydrus/help/running_from_source.html

>>9799

Yeah, no hurry, I'm already using 4chan X to track individual threads and also pin/hide threads based on tag lists and blacklists, I was just wondering if it was in the cards down the road.

If you do try 4chan X, do some searching around and testing, there are a variety of forks, some are dead or broken, and I haven't upgraded in a while for fear of breaking everything.

I knew that 8ch had json, but I didn't know it had RSS feeds, so thanks for that.

Now if people will just post on 8ch again…

>>9796

Thank you for this report. Every time I think I pin down one of these miscounts, another appears. Please try help->debug->data actions->clear db service info cache. This will force a clear-and-regen of those counts and should fix you up.

This is probably a one-in-a-thousand random event, but if it keeps coming back for you, please let me know and we'll go deeper and try to figure out where the miscount is coming from.

Thank you for petitioning these bad tags. I hope to roll out a tag filter for the PTR in the nearish future that will ban 'invalid tag' and 'artist request' and the other booru artifacts so we won't have to deal with them again.

>>9785

>Can you remember what you were doing at the time? Had you recently closed a dialog, or were attempting to? Do you have a 'busy' client with say dozens of pages and downloaders all working at once?

I might be mistaken but I think I was closing a manage tags window. It's not a busy client. Less than 10 pages with few items in them, nothing downloading.

>>9809

I happened to get this error myself this week and think I figured it out. Please let me know if you have any more trouble in v320 on!

I found a cool bug in v318, but not sure what it is or if you already know about it. Here goes:

>I had several multiwatcher tabs open, most idle but one downloading 12 or so threads at once

>I had one simple downloader tab open, downloading from around fifteen threads from an off-site archive

>also syncing tumblr subscriptions at same time

>switch to the dupes tab and run the rebalancing, hadn't done it since before the new tumblr fuckery so I had I think 84 branches to rebalance

>goes for a while, I'm tabbed out reading a thread and messing with jdown and stuff

>tab back in, it's just finished, hit OK, instead of letting it sit a minute and then running dupe discovery, I immediately switch back to the working multiwatcher, but hydrus hangs as it tends to because my laptop sucks, especially after finishing big operations, a message box has popped up under the subscriptions box, but it's blank because the client froze there

>after ten or twenty seconds it resolves itself and refreshes the screen, seems normal, except the subscriptions downloader box and the message box below it are just gone

>figure it was just a message saying it thought it was done syncing for the day, not sure why they closed themselves though

>elapse ten or twenty seconds of normal behavior

>open up tags to enter in new tag for thread I'm about to paste into multiwatcher, enter them and remove last thread's unique tags, hit OK or Apply or whatever, box closes

>client hangs again, this time doesn't recover, but rather than going (Not Responding) it just makes that annoying system bell whenever I try to click on any tab or anything at all

>downloader pops back up and starts, figure it's fine and I can just wait a bit since it's actually downloading

>stays frozen, can't select anything, but downloader still going and there's a rare C++ level error spat out at me instead of Python

>running it in debug, so go to console window and look at it, shows basically same thing (forgot to screenshot the full error, sorry)

>was able to select the program itself before from other front windows, but after selecting the debug console now can't select the GUI and the console stays up in front, still dinging

>Ctrl+C from the console so it doesn't crash my computer when I try to close the GUI

>restart, everything is completely normal and fine except the new tags weren't saved as entered in

The error that popped up under the subscriptions box when it first froze up was:

C++ assertion "IsRunning()" failed at ..\..\src\common\eventloopcmn.cpp(83) in wxEventLoopBase::Exit();
Use ScheduleExit() on not running loop

Doesn't really hurt anything and may not be duplicable on a non-potato, but I thought it was interesting.

>>9819

Thank you for this report. This looks the same as >>9782 . The 'modal' popup that comes up during idle maintenance and a couple of other places was being rude to any pre-existing dialogs. The dialog gets hidden but then can't restore when the maintenance job is over, and since the dialog is still modal, ui functionality can't return to the main ui frame. In your case, it was probably the tags window for your new multi-watcher. Maybe some idle job got delayed while the dupe search was going on, arrived late once your db-commit hang finished, and then blatted itself into the ui event loop right when you had that dialog open.

I believe I have this fixed for today, but if you experience it again, particularly with the dupe search stuff–which works on slightly different logic–please let me know. I now know what the error is actually about!