TashTalk USB
- Thread starter Bolacore
- Start date
I just put an order in today, so should be here a week on Monday most likely. I also had PCBWay reach out to sponsor a run, so I might be able to do some reduced price ones for forum members. Still trying to figure it out with them of course.
I messaged Joe as well and heard back, and I think it was just a misunderstanding thankfully, so no hard feelings.
I messaged Joe as well and heard back, and I think it was just a misunderstanding thankfully, so no hard feelings.
@Bolacore Trying out the new version of TailTalk, 0.4.2, and indeed, bullet characters are now appearing correctly! However, some other change(s) you've made appear to be causing numerous files/folders, though not all, to appear as individual, unnamed files on the vintage Mac side, hundreds of megs in size. If I try to open Get Info on one of the files, the window quickly closes again and a number of duplicates appear in the window.
(Also, it seems I may have been using a faulty cable initially, as a new one appears to be behaving perfectly!)
(Also, it seems I may have been using a faulty cable initially, as a new one appears to be behaving perfectly!)
Attachments
Last edited:
Ooo that doesn't look good, sorry about that! Is this only happening in 0.4.3 or what is the first version that you see this in? Could you possibly do a file list of the directory you are serving up as an attachment (Perhaps with this as the command)
Bash:
find /path/to/your/share -exec stat -f "%HT %N" {} \;If you serve up a brand new empty folder does it also do the same?
Last edited:
Ah, the joys of dealing with pre-Unicode systems and protocols.Figured out the bullet issue - It helps if I convert back to MacRoman on the reverse directionFix will be in the next release.
I eventually have to get one of these USB adapters. Now that Netatalk 4.5 is out, I'd like to add support to A2SERVER and finally get it out the door.
As the Netatalk maintainer you can have one on the house I have relied so much on Netatalk code for understanding how to make AFP work properly and it is the least I can do to say thanks. When the new set comes in I'll put one aside for you, just DM me your shipping info if you're interested.
I have relied so much on Netatalk code for understanding how to make AFP work properly and it is the least I can do to say thanks. When the new set comes in I'll put one aside for you, just DM me your shipping info if you're interested.I'll get it over later today. I'm assuming this is detected as an off-the-shelf USB-to-serial bridge. The user is likely to have to select a device, but thats for me to figure out. Kinda wish it was easier, but its annoying to get custom USB VID/PID identifiers. Not to mention it costs money.
Yup yup, it is a CP2102N and just shows up as a virtual serial port. The device itself lets you set the manufacturer and product names which I have been setting (product name is "TashTalk USB" when querying) and then auto-select and display in the drop down box for device selection in the GUI. It's not perfect of course, but seems to work pretty OK for a plug-n-go setup - a lot easier though than any custom identifiers as then we'd need matching drivers, signing, and thats bleh.
Thanks to @Iesca I think we have the regression in 0.4.3 fixed in 0.4.4 here: https://github.com/FeralFirmware/TailTalk/releases/tag/0.4.4. The TL;DR is - I had wrong offsets in the FPEnumerate response after MacRoman conversion, and also character encodings are hard and we should just all go back to using MacRoman everywhere instead.
