/hydrus/ - Hydrus Network

Bug reports, feature requests, and other discussion for the hydrus network.


New Thread
X
Max 20 files0 B total
[New Thread]

Page: Prev [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] Next | [Index] [Catalog] [Banners] [Logs]


https://youtube.com/watch?v=BvfHlZ8QRaI
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v580/Hydrus.Network.580.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v580/Hydrus.Network.580.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v580/Hydrus.Network.580.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v580/Hydrus.Network.580.-.Linux.-.Executable.tar.zst

I had a good week working on a mix of stuff.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

I may have fixed a program freeze when minimising to tray via the close button (the settings for this are under options->system tray). If you have had trouble with this before, please, when you are at a convenient point to risk a hang, try it again and see if you have trouble. If you do, what happens if you minimise to system tray from the file menu--still have problems, or is that reliably fine?

I added a new maintenance command to tag right-click, 'regenerate tag display'. This is a catch-all job to fix bad autocomplete counts and sibling/parent display. Previously, the way to fix this was to hit a 'regen the whole cache' job under the database menu, but on large clients this could take ages--we now have a way to debug and simply, fingers crossed, just fix a bad tag, all in, typically, a few seconds. If you have any jank siblings or 'the autocomplete said (23-28), but it gave me 8 files', try it out!

I fixed an issue with temp-file cleanup after importing read-only files. If you do a lot of unusual/misc hard drive imports, you might like to shut your client down, check your temp folder (hit _help->about_ to find it), and delete anything called 'hydrusXXXXXXXX'--it is all useless cruft after the client is shut down. It shouldn't come back!

The new 'eye' icon in the media viewer's top hover window now has the 'apply ICC Profile stuff' option. It updates the image live, as you click it! If you are interested in ICC Profiles--and perhaps comparing this in the duplicate filter--check it out. I also added a shortcut for it, to the 'media viewer' set.

The 'file log' has a new 'search for URLs' menu command, which lets you explore what has the selected URLs mapped. It is basically a copy of what I added to the media right-click 'urls' menu recently, and it'll help figure out weird URL collisions and mis-maps.

The Client API '/add_tags/add_tags' command has a couple of clever new parameters to govern how deleted records are navigated. If you are doing clever migrations or other big mappings operations with the API, check it out.

next week

I want to get some repository janitor workflow stuff done.

https://youtube.com/watch?v=qzTwBQniLSc
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v579/Hydrus.Network.579.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v579/Hydrus.Network.579.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v579/Hydrus.Network.579.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v579/Hydrus.Network.579.-.Linux.-.Executable.tar.zst

I had a great week mostly working on UI quality of life.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

There are several places where you can enter regex (clever text search rules) in the program. I have written a nicer text input widget and spammed it everywhere. It has several improvements: it colours green/red depending on whether the current text will compile; its menu button collects better tutorial links; and the one in the String Converter regex replace now shows how to do (unnamed) or  group replace.

Many multi-column lists across the program have new scroll-to tech. If you import a new parser now, or edit one and change its name, then the list will scroll to the new item once done. Selection preservation through various updates is also improved. The 'manage URL Classes' dialog test input will also select and scroll to the matching URL Class. There's plenty more to do, but I've ironed out a bunch of ugly behaviour.

I've improved some 'hey I think we already saw this URL before' logic, fixing an awkward false positive result. If you have had incorrect 'already in db' on slightly different versions of a file on a booru, and you traced it to a bad source url, they should work in future.

Import Folders now work much more efficiently. Due to some stupid old code, they were wasting a bunch of CPU, particularly when the import folder was very large. If you have big import folders that are normally laggy to load, let me know how your background latency feels this week--it might be that I fixed some weird jitter/hangs here.

I folded in the library updates from the 'future build' test we ran a few weeks ago. There were no problems with the test, so Linux and Windows users get a slightly newer version of Qt, and Windows users get new sqlite and mpv. Let me know if you have any boot problems!

I also fixed a weird recent problem with the 'requirements.txt' that users who run from source rely on to set up their environment. There's a new version of numpy out that breaks some things, so if you tried to set up a new venv in the past few days and could not boot, please try again with this.

next week

Some more slightly-larger-than-small jobs. I have some weird sibling problems to fix, PTR janitor workflow to figure out, and awful sidecar UI to work on. I'd also like to get back to the duplicate auto-resolution system, but I think I'll be out of time.
 >>/1655/
Hey dev, quick question.
Is there a way to access the import files tag parser system but for files already imported, drawing from filename namespace tags in place of the import filename?
I don't want to reimport everything and mess up existing tag namespaces but I have a lot of files where I downloaded their filenames along with them and the filenames are meaningful, including tags or titles or artist names, so on.
I had a good week working on a mix of stuff. There's a new maintenance job that recalculates the presentation and counts of individual tags, some UI fixes and a couple clever shortcuts for QSS refresh and ICC Profile switching, and some fixes to unusual file import problems.

The release should be as normal tomorrow.

 >>/1656/
Not yet, but I want to figure out a sort of 'internal' metadata migration that uses the sidecar system's tools to do this. You'll be able to set a metadata source and a metadata destination and it'll pipe from one to another, kind of like if you exported files with sidecars and then reimported them.

You would, fingers crossed, be able to say 'suck up my filename: tags from here and then do these conversions to them and then spit them out back here at this service'.

I don't know when I will be able to get this done, but I'm keen on working on sidecar UI improvements in the near future, and this tech will fix a lot of pain in the ass problems like 'I want to parse my URLs into tags' and so on, so it is on my mind.

https://youtube.com/watch?v=ieKs9G1YBl4
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v578/Hydrus.Network.578.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v578/Hydrus.Network.578.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v578/Hydrus.Network.578.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v578/Hydrus.Network.578.-.Linux.-.Executable.tar.zst

I had a good week mostly doing some simple work. In a bonus, animated webps are now fully supported.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

Animated webp decoding is not widely supported, but we discovered a method this week and I plugged it into my old (slightly janky) native animation viewer. I have made it work just like the for gif and (a)png, where, within hydrus, images and animations will count as different filetypes. On update, all your existing webps will be queued for a rescan. If they are actually animated, they will become 'animated webp' and get num_frames and a duration, and they'll play animated in the media viewer. Let me know if you run into any trouble with it!

I added an 'eye' icon menu button to the top hover window in the media viewer. It has those five new checkboxes for 'draw stuff in the background' I added last week, and I expect to add 'always on top' and similar options to it in future.

The 'known urls' media menu has a couple changes. It is now just called 'urls'; it gets a 'manage' command, moved from the 'manage' menu; and you can now open an URL 'in a new page' (i.e. opening a new search page with 'system:known url=blah'), so if you need to find all the files that have a particular URL, this is now just one click.

next week

I am now on vacation for a week. As normal I'm just going to disappear to shitpost fake-E3 and otherwise try diligently to achieve nothing of any importance. I'll be back Saturday 15th, with v579 on Wednesday 19th. Thanks everyone, see you later!
The new incremental tagging is great and the default was what I needed. The move feature saved me too, because "oldest first" has the tenth image in the set imported first for whatever reason, and it looks like they put the 28th image at the end, forgot to post it whil posting the set. Everything seems to be working and nothing to report, I would love a "move forward/backwards x" popup in the future if practical.

I upgraded my client
464 > 474 > 509 > 529 > 554 > 578
I technically could have skipped 529 since 509 > 554 gave me an error about regenerating missing ratings tables, and I don't use ratings at all (yet), but since it said it was a "serious error" I went ahead and added the extra step anyways. 529 > 554 still gave me a "may or may not be serious" message about some sort of table entries for ratings I think, but since I don't use them, I just went ahead. Other than that, I didn't ignore any warnings aside from the bitrot notice, which led to the path above.
 >>/1651/
I do have one other thing to mention, the page tags are fine but when I open my collection to page through it, it starts oldest-first (so the image correctly tagged page 10). This didn't change when I changed the way I sorted things to something else, it still shows the collection in import time order instead of page order, and the thumbnail is that oldest, page 10, image. Is that just how collections work or do I need to change some setting not available in the files tab?
 >>/1652/
Never mind, it finally occurred to me I can select portions and move them past the single file together instead.
I had a great week back into things. I improved UI quality of life with some better list workflow and regex editing, cut down on import folder inefficiency, and fixed an annoying problem in the known URL checking logic.

The release should be as normal tomorrow.

 >>/1652/
Not sure if I am totally understanding you, but collections have the same internal sort as the outer sort you have set on the page. If you set to sort by import time, collections will sort their internal files by import time and then sort amongst themselves by import time too (usually by something like their minimum held import time, or for metadata that aggregates appropriately, total filesize for filesize). So, if you have paged content collected, I recommend you always sort by 'namespaces->creator-series-volume-chapter-page' kind of thing, and collect-by creator+chapter or similar.

In future I expect I'll add a second sort for collection internals, or an override option, or something, but it is stuck for now.

https://youtube.com/watch?v=9QY0OJ8RHYE
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v577/Hydrus.Network.577.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v577/Hydrus.Network.577.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v577/Hydrus.Network.577.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v577/Hydrus.Network.577.-.Linux.-.Executable.tar.zst

I had an ok week. There's a mix of small improvements, and some neat OS integration.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

Thanks to a user, we have some cool new OS integration this week. For Windows, you can now go open->in another program on a thumbnail to get the Windows open with dialog, and open->properties to get the normal Windows file properties window. Also, the various places that can view a file in your OS file explorer now work better, and are better about selecting the file(s). Users who runs from source will want to rebuild their venvs this week to get this new stuff.

When you have a single 'edit time' panel, like the sub-panels for modified and archived time in the 'manage times' dialog, now accept any date string for the paste button. So, you can now paste something from your web browser or just 'yesterday' or something, and it should all work. This uses the same date parsing tech we have had success with elsewhere.

I broke the 'media' help page into two and added some new checkboxes to set whether to duplicate the text of the different hover windows in the background of the media viewer. If you want a clean background, it is now easy, under the new 'media viewer' page.

I fixed some more URL encoding stuff, this time related to brackets in parameter names, and buffed the 'clear orphan tables' database routine. If I have been working with you on either of these, please try your thing again and let me know how it goes.

next week

This is the last work week before my vacation, so I'll just do some simple cleanup work again.
 >>/1646/
Hi dev, I have a question.
If I already have some images I want to be a collection tagged with their shared title, and I already have them in the proper page order (via time imported or so on), is there an easy way to just select a range of thumbnails and apply page tags to them, starting with 1 for the first?
I'm looking to avoid exporting and reimporting with regex stuff
I had a good week making some simple improvements before my break. There is also full support for animated webps!

The release should be as normal tomorrow.

 >>/1647/
Yeah, you are in luck since I only added this a little while ago--hit up the normal manage tags dialog and then click the new '±' button that is just above the tag text input. This launches a little sub-dialog that can do all sorts of incremental tagging. Note that right-clicking thumbnails has a submenu called 'move' that lets you reorganise things, too.

Let me know if you run into any trouble with it!

https://youtube.com/watch?v=gWmECLnMKGk
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v576/Hydrus.Network.576.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v576/Hydrus.Network.576.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v576/Hydrus.Network.576.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v576/Hydrus.Network.576.-.Linux.-.Executable.tar.zst

I had a great week. The program should be less laggy when busy.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

improved file access latency

I reworked the client file manager's locks to be more fine and sophisticated. When you have several importers working in the background, and particularly importers handling large files, the client will delay access to your files and thumbnails far less, and generally not at all. I have been thinking about the specific change I intended to make here for a while, but this stuff can be tricky and I wanted to think and be careful how I did it. All tests so far are proving good, but let me know if you run into any trouble.

Secondly, when you do have file access lag, the mpv player now waits more politely. It won't hang the UI any more; it'll just sit on a black screen until the video file is ready. You can roll off to another media and come back, close the media viewer entirely, whatever you want.

misc

Managing 'import options' in gallery downloaders and thread watchers and their sub-downloaders should be a touch easier this week. I have replaced the static 'set options to downloaders' buttons with a dynamic button that appears only when the current selection of downloaders actually differ with what the main page has set. That pages and their downloaders have unsynced options has always been a complicated concept that I have had difficulty highlighting and explaining, so I hope this improves things a little just by osmosis. I also fixed a bunch of holes and bugs in the system.

Thanks to a user, we now have support for .doc, .ppt, and .xls Microsoft Office formats. A whole bunch of old stuff should now work, and it should get word counts in several cases. This requires a new library, so users who run from source will want to rebuild their venvs to get the tech.

If you noticed that colour picker controls would take several seconds to open and close, it shouldn't happen any more! It was a weird hack I injected to get around a Qt bug, and it seems to be fixed in the current version we use. However, if you find you suddenly cannot select any colours in the gradient square of the colour picker, that it feels like your mouse is drag-clicking the whole time, then I guess the bug wasn't fixed for you, so let me know!

future build

I am making another future build this week. This is a special build with libraries that I would like advanced users to test out so I know they are safe to fold into the normal release. More info in the post here:

https://github.com/hydrusnetwork/hydrus/releases/tag/v576-future-1

next week

I have two more weeks of work before my Summer vacation. I'll try not to start anything too big and just do cleanup and little fixes.
I had an ok week. I fixed some bugs and cleaned up some UI. There's also support for 'open with' and 'file properties' in Windows.

The release should be as normal tomorrow.

https://youtube.com/watch?v=Auc5wHXPQaw
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v575/Hydrus.Network.575.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v575/Hydrus.Network.575.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v575/Hydrus.Network.575.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v575/Hydrus.Network.575.-.Linux.-.Executable.tar.zst

I had a great week. I made a bunch of small improvements, and I am gearing up to start duplicate auto-resolution.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

The new tag autocomplete 'children' tab now sorts by count and clips to the top n (default 40) results. You can change the n under options->tags. This takes a bit of extra CPU to figure out, so let me know what performance you see in IRL situations.

When you ask subscriptions to 'check now', if there are a mix of DEAD and ALIVE subs, it now asks you which ones you want to check. Should be easier to resuscitate DEAD subs en masse now.

More twitter URLs are supported by default. You should now be able to drag and drop a vxtwitter or fxtwitter embed URL on the client and it'll figure it all out.

I fixed the (non) alphabetisation of GET URL parameters in URL Class settings, which I broke by accident a few weeks ago with the new URL encoding tech. If you have a super advanced downloader that relies on parameters being in a certain order, please try again and let me know if you still have any issues.

duplicate auto-resolution

This week I also put time into planning out the duplicate auto-resolution system, which will optionally and highly-user-configurably resolve the easiest duplicate pairs without human input. I am happy with my plan, and I have sketched out a skeleton of code. The whole problem looks simpler than I expected, which is good, but as always there will be a lot to do. 

I will keep chipping away at this in the coming weeks and will roll out the features in waves for advanced users to try out. The first version will resolve jpeg/png pixel duplicate pairs.

next week

A bit more auto-duplicate resolution work, and I'd like to see if I can reduce some of the juddery latency when you try to look at media, particularly videos, while the client is busy doing imports in the background.
I had a great week. I cleaned some jank code, fixed some bugs, improved some downloader UI, and massively reduced file load lag when there are many imports going on.

The release should be as normal tomorrow.

https://youtube.com/watch?v=XBETVhHpcPk
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v574/Hydrus.Network.574.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v574/Hydrus.Network.574.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v574/Hydrus.Network.574.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v574/Hydrus.Network.574.-.Linux.-.Executable.tar.zst

Note: If you are updating from v570 or earlier and you use the Windows or Linux .zip or .tar.zst 'Extract' releases, you have to do a clean install one time to get v571 or later! (https://github.com/hydrusnetwork/hydrus/releases/tag/v571, https://hydrusnetwork.github.io/hydrus/getting_started_installing.html#clean_installs). If you are a Windows installer/macOS App/source user, you do not need to do a clean install to get over the v570->v571 bump; just update as normal.

I had a good week. The database has some new repair tech, and the new default twitter downloader now gets full-size images. The update this week may take a minute due to some database maintenance.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

highlights

For some time, some users who sync with the PTR have had sporadic '404' errors that interrupt downloading/processing. We never pinned down exactly what was causing or fixing this (a bunch of different maintenance routines seemed to fix it eventually) until this past week. Long story short is there was a SQLite issue that was crossing a wire on some program crashes, and a bunch of new repair code now detects and fixes this situation. Several rare 'this file appears to be this file' issues are fixed as a result. Your 'local hashes cache' will do some work on update, but the new code should run pretty fast, so even if you have millions of files, it shouldn't take too long to finish. If you have had PTR syncing problems, please try unpausing after updating this week--while it may fail one more time, the new repair code should be able to fix it now. Let me know how you get on!

I screwed up the new default twitter downloader last week--it was not getting the full-size versions of images! This is fixed today, so if you spammed a bunch of tweet URLs last week, please do this to fix it: do a search for 'imported within the last 7 days/has a twitter  tweet url/height=1200px' and then copy/paste the results' tweet URLs into a new urls downloader page. The full-size images should download. Once done, assuming the file count is the same on both pages, go back to your first search page and delete the 1200px tall files. Then repeat for 'system:width=1200px'! Sorry for the trouble!

If some of your files get an odd colour tint, is it always on files under 'system:file properties->has icc profile'? If so, try unchecking the new 'apply image ICC Profile colours adjustments' checkbox under options->media. Let me know what you see.

next week

I blinked and it is May already. I think I have four more releases before my vacation week, and it feels like I haven't done anything, so I'd like to get a larger job done. I want to try the first framework of an automatic duplicate pair resolver.
I had a great week. I improved quality of life and fixed some bugs.

The release should be as normal tomorrow.

https://youtube.com/watch?v=1glH_p16WB0
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v573/Hydrus.Network.573.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v573/Hydrus.Network.573.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v573/Hydrus.Network.573.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v573/Hydrus.Network.573.-.Linux.-.Executable.tar.zst

I had a good week working on some quality of life and a tagging experiment.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

my experiment

I've written a new tab, 'children', for the tag autocomplete dropdowns. It shows the tag children of anything in the list the autocomplete is managing. The idea here is you have 'evangelion' in your manage tags dialog but you have forgotten the names of the characters--there is now an easy-select list for you.

Since this is a first draft, it is workflow-ugly. It works well when there are only one or two parent tags in the list, but any more than that (or something with hundreds of children like pokemon or azur lane) and it becomes spam. Please have a play with it, especially if you are an advanced user, and let me know what you think. I think I'll filter and sort it according to the top n most-counted tags, and there's probably some simple selection-preservation stuff to add, to better let you run through its list quick-hitting enter over and over (atm it resets the selection position on every change).

other stuff

With the help of a user, I've written a simple twitter tweet downloader. This isn't a full search, just something that can download single tweets so you can drag-and-drop tweet URLs onto the client or paste them into an 'urls downloader'. It can handle video and multi-media posts, and it fetches quote-tweet contents too. I've learned there is a different, more advanced twitter downloader written by another user--if you use this, my new tweet downloader will not be added today, since you have that ability already. I'll see if I can add some of the other downloader's tech to the defaults in the near future.

The archive/delete and duplicate filters now yes/no confirm with you when you select to 'forget' your work.

All the builds should now be compatible with very new (>=0.38.x) versions of mpv/libmpv. If you run from source, you might like to rebuild your venv this week.

I continued to deconstruct the local booru (an old way of sharing files from the client) this week. It now no longer starts its server. If you used this, please note that this Client API project replaces it very well: https://github.com/floogulinc/hyshare

next week

I am pretty exhausted, I'm sorry to say, so I'll keep it simple and just do some more quality of life and little bug fixing work.
I had a good week. I wrote a new database maintenance routine (which should fix the PTR '404' bug) and some small UI improvements and code cleanup.

The release should be as normal tomorrow.

https://youtube.com/watch?v=ZMbBsR6Hc6o
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v572/Hydrus.Network.572.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v572/Hydrus.Network.572.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v572/Hydrus.Network.572.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v572/Hydrus.Network.572.-.Linux.-.Executable.tar.zst

I had a good simple week mostly doing just one thing: overhauling the 'share' menu.

Note: If you are updating from v570 or earlier and you use the Windows or Linux .zip or .tar.zst 'Extract' releases, you have to do a clean install one time to get v571 or later! (https://github.com/hydrusnetwork/hydrus/releases/tag/v571, https://hydrusnetwork.github.io/hydrus/getting_started_installing.html#clean_installs). If you are a Windows installer/macOS App/source user, you do not need to do a clean install to get over the v570->v571 bump; just update as normal.

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

share menu

Like the 'open' menu a couple of weeks ago, the 'share' submenu you see off of the thumbnail or media viewer right-click menu has been given a full pass. The layout is clearer, and exactly the same in all locations, and everything it does is now mappable in the shortcuts system.

Also, you can now copy files' thumbnails as bitmaps from this menu. Should let you do saucenao style lookups on unusual filetypes that have a thumb but no normal image.

The shortcuts here are a bit smarter, too--like with the 'open similar files in a new page', I collapsed the four 'copy xxx hashes' commands down to one command with a dropdown (which has options for blurhash/pixel_hash, if you need them); and the three 'copy xxx bitmap' commands are now one with a dropdown that also has the new 'copy thumbnail bitmap' option. Most of the shortcuts also have a choice between 'do the focused file' vs 'do all selected files', if that is important (existing shortcuts will default to the latter, which was previous behaviour for stuff like 'copy sha256 hashes').

Note: I have removed the 'share on the local booru' command. The Local Booru (an optional way to share files with friends using hydrus) was a fun experiment, but I have decided to finally retire it. I never found enough time to dev it properly (I don't think I've touched it in at least five years), and there are many better ways to share files online, be that a third-party host you simply drag and drop to or a clever Client API solution. So, it is gone from the menu today, and I'll slowly dismantle it over the next few weeks until it is completely removed.

couple other things

There's a new checkbox under options->files and trash that adjusts the 'remove processed files after an archive/delete filter' to also remove the stuff you skipped.

Also, thanks to a user, there's a new 'e621' themed stylesheet under options->style. Users who run from source who want to try this probably have an extra step to make sure this works--check the changelog/options panel help text.

next week

I've been thinking of adding a new tab to the autocomplete dropdown that will show children of the current context (e.g. you are looking at something with 'evangelion', in that tab you'll see 'asuka', 'rei', etc.. for quick selection). I think I will play with that idea and do some more misc small jobs.
I had a good week. I improved some quality of life, added an experimental new tagging tab, and have a simple tweet downloader working again.

The release should be as normal tomorrow.

https://youtube.com/watch?v=2umMp1VHWiA
windows
zip: https://github.com/hydrusnetwork/hydrus/releases/download/v571/Hydrus.Network.571.-.Windows.-.Extract.only.zip
exe: https://github.com/hydrusnetwork/hydrus/releases/download/v571/Hydrus.Network.571.-.Windows.-.Installer.exe
macOS
app: https://github.com/hydrusnetwork/hydrus/releases/download/v571/Hydrus.Network.571.-.macOS.-.App.dmg
linux
tar.zst: https://github.com/hydrusnetwork/hydrus/releases/download/v571/Hydrus.Network.571.-.Linux.-.Executable.tar.zst

I had a simple week working on some quality of life and background stuff. There are special install instructions this week!

Full changelog: https://hydrusnetwork.github.io/hydrus/changelog.html

new build

tl;dr: If you use the Windows or Linux .zip or .tar.zst 'Extract' releases, you have to do a clean install! (https://hydrusnetwork.github.io/hydrus/getting_started_installing.html#clean_installs). If you are a Windows installer/macOS App/source user, you do not need to do a clean install; just update as normal.

The future build test went well, so I am folding the updates into the main build. The above releases are updated from Python 3.10 to 3.11, and Qt (UI) and OpenCV (image processing) are moved to newer versions. There aren't any super important changes, but a bunch of little things should work better or be a bit faster.

Unfortunately, the new libraries cause a dll conflict with v570 and earlier (basically the executable sees the py310 dlls beside the py311 ones and gets confused), so we need to clear the install directory of the old files. Just do the clean install and you should be fine!

You don't have to do it for the other modes because: the Windows installer basically does a clean install every time; the macOS App is always enclosed in its own thing doesn't have to worry about old files; and running from source doesn't care about dlls in the same way, although you might like to rebuild your venv today, just to catch up your own library versions.

If you do have trouble booting v571, then please revert to your v570 backup and let me know what happened! There were no problems in the test that people tried a couple weeks ago, so I'm not expecting anything much, but I'll jump on any reports. Also, if you have been struggling with some annoying menu or drag and drop bug, let me know if the new version of Qt fixed you.
the rest

The archive/delete filter gets a couple of workflow changes: first, if you finish a filter and there is more than one possible local file service to delete from, those 'commit' buttons are now disabled for 1.2 seconds. This is to catch you from spamming 'enter' through this dialog when it is suddenly different (I've done this myself more than once). Second, if you hate the idea of these buttons being disabled, and you always want to delete from all local file services anyway, please hit the new 'when finishing filtering, always delete from all possible domains' checkbox under options->files and trash, which lets you always have a simple 'commit' dialog that only shows 'delete from all local file services'.

The client now tries to load truncated images by default. The damaged images it now allows might be missing one pixel in the bottom right, or have a few lines of grey at the bottom, or might appear fine but just have some crazy metadata, but they won't, fingers crossed, fail with a 'malformed image' error any more. We had some stability problems with this mode some years ago, so I turned it off and only allowed it on in a debug menu on a per-session basis, but the situation seems to have cleared up, so it is now back on. If you need to turn it off, hit options->media.

Any time you have a normal single column list in the program, e.g. the list of URLs in 'manage urls', you can hit Ctrl+C or Ctrl+Insert and now it'll copy better strings (e.g. without '(1)' decorator cruft), and it'll copy every row you have selected.

I wrote a new emergency help document, 'help my db disappeared.txt', for the install_dir/db folder. If you ever boot up and get the 'this looks like the first time you have run the program' popup, there's now a guide to figure out what the hell happened.

next week

I didn't find the time to get to the 'share' menu rewrite, so I'll try again.
I had a good if simple week. The 'share' menu (off of thumbnails or the media viewer) is completely overhauled, and everything in it added to the shortcuts system.

The release should be as normal tomorrow.

Post(s) action:


Moderation Help
Scope:
Duration: Days

Ban Type:


New Thread
Max 20 files0 B total
Refresh

Page: Prev [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] Next | [Index] [Catalog] [Banners] [Logs]