https://gitlab.synchro.net/main/sbbs/-/issues/782#note_5574
Thanks for finding your old commit @echicken. At least two things in common,
which does make it look like it's probably a platform specific optimization or
bug:
- You documented the sysop using "Linux Debian 11 on a raspi 4, 32 bits" and
Tucumcari's sysop said they followed the raspi instructions, although they're
using a banana pi on armbian (I don't know anything about these devices, but I
guess they're similar and if they followed the same instructions, would be using
the same js library)
- You said "Some invalid index was being used to access the array of strings, I
don't know what. This seemed to begin after eight iterations, don't know if that
was consistent.", and based on my testing the unmasking works fine for the first
8 data bytes and then fails from the 9th byte and beyond, very consistently
(tried about 10 times)
Since the sysop is willing to help test, I'm going to go with your first
suggested fix first, which if it works will let us know to avoid using logic
when accessing an array. If it doesn't work, then I'll move on to your second
suggestion and incorporate DM's increment-after suggestion as well.
--- SBBSecho 3.20-Linux
* Origin: Vertrauen - [vert/cvs/bbs].synchro.net (1:103/705)
|
