Hi folks. I’m finally back online after about 10 days! No surprise there, when dictators feel threatened, they seem to have no tools other than killing people and shutting down the internet. Let's see if they do it again.
Wow, welcome back! So sorry about what is happening to the people of Iran and so much routing for better times. The current regime has literally sucked too much blood out of Iranian society.
@vick21 Thank you so much. I really appreciate your kindness and solidarity. It means a lot to know that people outside Iran are paying attention and hoping for better days. We’re all holding on and hoping for change, even in times like these.
@serrebi I just updated BlindRSS! Thanks for so many fixes and features. I still have those aforementioned issues with Bluesky and Youtube feeds though.
@serrebi Sure. 1. BlindRSS has issues with Youtube feeds. We can add them and articles get refreshed, but upon pressing Enter on their articles we get the following error: Error dialog Could not resolve media URL via yt-dlp: ERROR: [youtube] Fub7pSvlrlc: Sign in to confirm you’re not a bot. Use --cookies-from-browser or --cookies for the authentication. Check the following Youtube RSS URL as an example: youtube.com/feeds/videos.xml?c… 2. BlindRSS has serious issues with Bluesky RSS feeds. Articles do appear, but the article name/ title/ text all are reported as being unknown. Try the following Bluesky RSS feed for an example: bsky.app/profile/did:plc:mtxyc…
Hey man, just wanted to say thanks for putting back up this awesome client! It kicked ass as a twitter client, and it'll for sure kick ass as a Bsky client!
In the house where I'm staying now, there are three wireless access points in strategic places in the house. Unfortunately, the linking isn't as optimal as I'd like it to be.
I've been primarily using a remote computer from my iPhone for nearly three weeks, so when there is any lag, it becomes pretty obvious.
Pretty much every night here, the lag increases to an annoying degree, so much so that I get better results by turning off WiFi and using mobile data (currently AT&T is working the best in the house). Because I'm using Tailscale, I still get access to all the same resources whether I'm on WiFi or not, and it only takes a second or two to reestablish the connection when the provider is switched, even if I change between my primary and secondary eSIM for data.
This place is so congested, though. If I do an environment scan from my UniFi controller, I see about 85 wireless access points in the area. It's not that bad in my New York City neighborhood.
I've confirmed that latency and jitter are perfectly fine on wired devices here in the house. Even with three wireless access points spread out as much as possible as far as spectrum goes, things still get stupid, especially in the upstream direction.
Anyway, I hate WiFi. The end.
In case anyone cares, here's the latency and jitter I get when pinging my iPhone on AT&T from one of my home computers in New York, about 500 miles away.
36 packets transmitted, 36 received, 0% packet loss, time 73ms rtt min/avg/max/mdev = 40.558/171.173/370.040/104.540 ms
On WiFi, it was more like min 39ms, max 550 ms, average 100 something ms. I've seen better while on AT&T.
New NV Speechplayer V7: Corrects words like "world", and the number 30 to sound less like a New Yorker for US accent. Corrects "broadcast" and "stop" to have nuanced sounding for US and UK. Adds Romanian language (it largely borrows phonemes from existing data.) Link: eurpod.com/nvSpeechPlayer-2025…
@jmd2000 well, it technically last got some code changes in 2021, picked it up again a few days ago because even though ESpeak did eventually integrate Speech Player it just never quite sounded as good as the standalone version that just borrowed it for phonemes. So I'm super happy to have it back, for me it's closest so far to an Eloquence replacement.
A federal judge bars DHS and ICE agents in Minneapolis from using “pepper-spray or similar nonlethal munitions and crowd dispersal tools” against peaceful protesters. Federal agents there now also cannot stop vehicles that are following them at a safe distance. storage.courtlistener.com/reca…
Sometimes, I just like to type (insert random thought) type beat to see what youtube throws up at me. I hope AI makes this game even more fun now. Genre benders/blenders are my fave! maybe i'll find a drake+polka "type beat" lol
@Bri first result? R&B Tracks| "Blue Sky"| No, not That bluesky 2 minutes, 7 seconds lol! welp maybe you will have to be the first to sneak that "type beat" in to the mouth of the all mighty algorithm! :D
I'm listening to a podcast, and the host is saying things like, "first question, numero uno," and I'm mildly annoyed, since he seems not to speak Spanish.
#FastSM 0.3.0 is now available! This is a pretty big release, and contains all of the changes between 0.2.2 and now. Full changelog below, use your open URL hotkey if you wish to skip this.
Improved image viewer The app now includes a much better, and actually working, image viewer, which replaces the image viewer which would appear when you press view image in the view post dialog with an image in the post. The new image viewer provides the ability to see all images in a post, as well as get AI image descriptions (see below).
AI image descriptions You can now get AI descriptions for images in posts. Workflow: first, set up your GPT/Gemini API key, whichever one/ones you want to use, in the AI tab of the settings dialog. Set the service and model you wish to use in the combo box, and tweak the prompt to your liking. Next, on a post with an image, you can view the image/images in the post in a few different ways. The easiest is through the context menu/view image. Other ways include the view post dialog, or the global shortcut, control win alt V by default. Next, you are put in a dialog containing all of the images in the post, which you can move between with the arrows. On each image, there is an edit field with the current description of the image, if one is provided. If you tab, you will eventually find a get AI description button, which when pressed, will kick off the process of getting the AI description. This process may take a bit, depending on your model and service combination. It will then replace the description with the provided AI description.
Mac FastSM has been able to run on Mac from source for a bit, but this is the first version of FastSM to provide a working, runnable MacOS build.
Audio Player FastSM now has an audio player, in addition to just playing audio in audio posts. The audio player can be called up by pressing control shift A in the main window, or control win shift A in the invisible interface. In the audio player, you can press up and down arrows to change media volume, as well as left/right to seek. The media volume persists between audio files. In addition to all of this, the audio player has two settings. One to have the audio player automatically come up when audio starts playing, and another to stop audio from playing when the audio player is closed.
Other audio improvements You can now set a per account soundpack volume. You can either do this by adjusting the volume in the normal way, and it will only now adjust for the focused account, or you can use the new soundpack volume slider in the account options dialog.
Dark Mode Add a very hacky/experimental dark mode support in the advanced tab of the app.
Spell Check It is now possible to check spelling on a post or reply you are writing, with the new spell check button in the new post dialog.
Confirmations It is now possible to enable confirmations for different toggle actions. For example, you may want a confirmation when you unfollow someone, so that you don't do it accidentally. In the program options dialog, there is now a new confirmations tab where you can enable individual confirmations for things such as favorite, unfavorite, boost, unboost, follow, unfollow.
Other additions/bugfixes Add enable/disable notifications buttons in the user viewer dialog. This will make Mastodon send a notification to your notifications buffer every time a user with notifications enabled posts. Add a feature where you can have the app only use a single API call on initial timeline loads. Add a global shortcut for the update profile option, control win shift U. Fix quoting. Quoting now works correctly in all cases. Fix a bug where media descriptions did not show for boosted posts. Fix a bug where you couldn't unschedule a scheduled post from the scheduled posts timeline. Fix a bug which meant you couldn't call up the scheduled posts timeline. Fix the title of conversation buffers when you load a conversation from a boosted post. Fix deleting bookmarks. Fix bugs with user selection when loading timelines/profiles. Fix search timeline errors. Fix pinned sound in user timelines. Fix a 404 error when loading user timelines on app startup. Fix various visual issues with the app.
Oy, WTF? Error: Gemini API error: You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: ai.google.dev/gemini-api/docs/…. To monitor your current usage, head to: ai.dev/rate-limit. * Quota exceeded for metric: generativelanguage.googleapis.com/generate_content_free_tier_requests, limit: 0, model: gemini-2.0-flash * Quota exceeded for metric: generativelanguage.googleapis.com/generate_content_free_tier_requests, limit: 0, model: gemini-2.0-flash * Quota exceeded for metric: generativelanguage.googleapis.com/generate_content_free_tier_input_token_count, limit: 0, model: gemini-2.0-flash Please retry in 53.917787436s.
This unfortunately seems to be broken, I wonder if this is on my end though as this hasn't worked with multiple other things I've tried to get it to work with.
Any idea what might be causing it? I'm on the default model with my API key put in. I do think this is a me problem because I think something else gave me this same trouble.
Hey something I've always wanted for mastodon, and I know there are web versions available, but it would be really cool if post scheduling was available. Not sure if that's gonna be that important to most, but just putting it out there. Very impressed with what i've seen so far.
@Bri Any chance of a way to run completely portably I.E. store user configs in a subfolder of the app's folder, for use on thumb drives, in Dropbox, or other situations where the app and its configuration needs to be able to be moved quickly to another system?
@doctormo I have a branch where I'm redoing clipPath with the new layout system. I'm going to be soooo happy when that one is mergeable; maybe a week more of work.
"“We are not going to shop. We are not going to work. We are not going to school on Friday, January 23. For some people they call that a strike.. For many of us, this is our right to refusal until something changes.”"
Minnesota appears to be in gear for a mass uprising. Unions, community organizations, faith leaders, and small businesses there are calling for a statewide day of “no work (except for emergency services), no school, and no shopping” on January 23.
@NicksWorld @Fragglemuppet Actually just noting, Quin is currently not a part of this project, it just uses work previously from her and my Quinter client for Twitter, which is why I mentioned her to give her credit
why does no one make USB-C hubs? i don't mean docks, just a hub with 1 USB-C host port and 8 or so USB-C device ports.
only useful for low-bandwidth devices, obviously, but most USB devices (mouse, keyboard, audio device, wifi/BT radio...) don't need full speed all the time.
i'm going to take a wild guess: USB-C is still new-ish, so perhaps logically you should assume (hah!) a minimum protocol version for the device that isn't the peripheral, and a "low powered hub" that is USB-C only would break this assumption if it was just ... USB 2.0 ?
Honestly WxPython is kinda messy in general, we need a better UI library but good ones like WinUI only work with C++ and C# so we're stuffed. Oh wait there is some WinUI Python thing but it seems basic unless I'm just too newby to grasp it roflfl
Only problem is menu bars don't have that file submenu alt + F reading, and it doesn't do dark mode by default, I've used Accessiweather and I noticed those issues. Also when I tried out the Toga demo thing it had weird areas that said pane.
Feel free to use this yourself of course, though you can't distribute it *with* the client. Licensing. I don't own the sounds, Nintendo does, and you know how they are.
@Bri absolutely not in a billion years. Clang is a horrible nasty C++ codebase with processor-specific stuff and I can almost guarantee that most LLVM/clang develoeprs wouldn't know what SPARC32 was if it beat them over the head
A Mac build of #FastSM can now be downloaded! Note, warning, etc: #FastSM Nightly builds on Mac are unsigned. This means you will have to allow them to run through system settings/privacy and security. github.com/masonasons/FastSM/r…
Hmm. So, To balance clarity and accent reduction, I'll set default stress modifications to (1.25, 1.07), as suggested by Gemini, with some exceptions. English might go to (1.4, 1.1), Hungarian to (1.15, 1.04), and Romance languages to (1.2, 1.05). For Portuguese, I'll consider a new nasalization tweak, removing ̃ and `` globally.
I'm working on these fun things called allophones. Sounds like a musical instrument right? "Now come on my son, go inside to eat and put down those allophones this minute!" Not even close. They are different, non-meaningful pronunciations (variants) of the same phoneme (basic sound unit) that occur in specific, predictable phonetic environments, like the aspirated 'p' in "pin" versus the unaspirated 'p' in "spin". Speakers of a language perceive these variations as the same sound, but they are distinct phonetic realizations that don't change a word's meaning, unlike phonemes which do.
they're all mapped in data.py - right now about 100-130 or so, but the IPA symbol of the sound is what each dictionary key entry is, and then the various voicing properties that create that sound. Many TTS engines do not pronounce IPA notation, oddly enough Espeak and NV Speech player do because Espeak has mapped these well in its own table. Eloquence? Just treats it like a questionmark. Haha, OK, it looks like there are 97 keys in the current data.
@menelion Thanks for taking me back ten years to the times when I was studying English. Fascinating stuff: the realization of some phonemes in Polish as mapped in the Espeak project was always a little odd - "sh" and "r" come to mind but there are more for sure. Good luck!
#FastSM tip: Use the invisible interface? You can call up the context menu, the same one as pressing shift F10/alt M in the window, by pressing control windows alt M! #TheMoreYouKnow
#FastSM tip: Use the invisible interface? You can call up the context menu, the same one as pressing shift F10/alt M in the window, by pressing control windows alt M! #TheMoreYouKnow
Got a question. There's a setting about load all previous items until timeline is full. If applied to the home timeline, is that infinitely? Or does it only apply to when there would be gaps from e.g. hibernation?
@x0 Yes. But this process will only start once you go to load previous. You can stop it in it's tracks by pressing the load previous hotkey again. It's worth noting though that it won't stop until it's finished with it's current pull.
OK. So *never* apply to home timeline unless I want it to try and pull all the way back to account creation from everyone I follow. Which sounds like it would flood something
DO you plan to allow us to reorder buffers, perhaps via move buffer down / move b uffer up actions for a buffer? And should there be a context menu for buffers?
@x0 Probably not any time soon. The buffer system creates buffers systematically in a pretty paricular order at the moment and would be a fun time to try and change that
@x0 Yeah I've noticed that, especially in the invisible interface. Ever since IIRC Qwitter, I've been used to having Mentions to the left of Home, and DM's to the right, so what I'm mostly interested in is on either side of Home.
Alrighty. Suppose I'm really not supposed to be using the invisible interface anyway. Time to make soundpack! Cause I don't like some of these very much. Could I begin with the quinter pack you think? Need to pull up a few new sounds from the source but not too many, pinned and new primarily. Unless I reuse my cake mention sound for new since here can't indicate mentions.
Oh cause this buffer order is clunky for it. So I wondered if it should really be considered something akin to an afterthought or something that gets less attention, given it can't work on mac and the context menu requires visible.
Alrighty. Perhaps you should write to Jonathan or Kev about adding a FastSM category to soundpack central, and I'll gladly stick my super mario maker pack as a variant there.
Yeah, the issue is due to the way the UI works, you can't just, pop a context menu out of nowhere without being tied to a window. It'll end up in focus limbo. I supose you could do a fake menu modal like what cake does for polls, but that isn't a real menu. Also, why do the checkboxes in settings say button checked and button? And when I unchecked the one for new item sounds in buffers it didn't seem to ever say button without the checked until I hit OK and opened it again.
@jaybird110127 I'm currently running from source. And whenever I invisible over to my mentions buffer it lags for some reason. I guess that's that UI sync setting?
@jaybird110127 Huh. I just figured sync with UI while UI is hidden becomes a no-op, but checked it for the case that I use invisible keystrokes in the UI.
@jaybird110127 Although I guess the window doesn't actually *vanish*, it's just obscured. So some UI update operations are actually happening in a window that nothing can see?
@jaybird110127 I actually wonder something. If you had a systray icon, would you be able to pop context menus out from that that are properly focusable without having to unearth the entire visible UI if it's hidden?
When I do this with the window hidden, it unhides the window, but then when I hit escape without doing anything and the menu goes away, the window doesn't auto-hide itself again. Is this a bug, or does it auto-hide if I actually do something, or…
is there a way of getting rid of the gui and just using invisible like with tweesecake and similar, I have looked through and if there is a way of disabling it I cannot find it lol
Hey @Bri, there seems to be a further bug with the template system in #FastSM. When I look at my scheduled posts, it say $account.display_name$, rather than my actual display name. Not a big deal since i don't imagine I'll be checking those very often, but I figured I'd bring it up.
Gotcha. Is there no way to resolve that on my end? Or is there something I can do about it? Maybe add a $scheduled_for$ tag unless that exists, and it's just not implemented yet.
In any event this is like, super fuckin' slick and I'm totally here for it. I did notice that there's no way to record an audio file to attach to a post, but that's not a big deal by any means. For on-the-fly posting of something though it'd be cool to see that. That being said you'd have to add a few more libs like opusinc or lame_enc and i'm not sure how big that'd make the repo at the end of the day.
Nodnods. I gotcha. I love what's already here though for sure. TBH I kinda miss clients like this. I know Tweesecake is still being worked on but until the new version's released, I'll probably use this. It's got features I was looking for and it's really light on resources.
Yeah. Me and Josh already talked about it and he's burnt out on TC which is totally valid. So I unintentionally kinda picked up the slack a little bit here in that regard.
You legit did. I love the actual shit out of it. I might end up switching to this full time. About the only thing I kept tc around for was for easily checking my hard drive space from the file explorer thing and getting github releases without fucking around on the website. Other than that...meh.
!["[Old PostgreSQL logo from 1996, featuring "PostgreSQL" text above what looks like a broken wall made out of brick, and with the space in the background] This was the first logo PostgreSQL used when released in July 1996. Frankly, it was a mistake to ever stop using it."](https://fedi.ml/photo/preview/640/749434 "\"[Old PostgreSQL logo from 1996, featuring \"PostgreSQL\" text above what looks like a broken wall made out of brick, and with the space in the background] This was the first logo PostgreSQL used when released in July 1996. Frankly, it was a mistake to ever stop using it.\"")
victor tsaran
in reply to Amir • • •Amir
in reply to victor tsaran • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •1. BlindRSS has issues with Youtube feeds. We can add them and articles get refreshed, but upon pressing Enter on their articles we get the following error:
Error dialog Could not resolve media URL via yt-dlp: ERROR: [youtube] Fub7pSvlrlc: Sign in to confirm you’re not a bot. Use --cookies-from-browser or --cookies for the authentication.
Check the following Youtube RSS URL as an example:
youtube.com/feeds/videos.xml?c…
2. BlindRSS has serious issues with Bluesky RSS feeds. Articles do appear, but the article name/ title/ text all are reported as being unknown. Try the following Bluesky RSS feed for an example:
bsky.app/profile/did:plc:mtxyc…
Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •And can we do something for Bluesky feeds?
Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Brandon
in reply to Amir • • •Amir
in reply to Brandon • • •Bri🥰
in reply to Amir • • •Brandon
in reply to Bri🥰 • • •Cleverson
in reply to Amir • • •