Conversation
Very good finds, bud!! I can only assume that this scenario with a Gnuton F/W build being published and then, hours later, being "recalled" (i.e. unpublished or hidden somehow) is an extremely rare event, and that's why we've never seen a user reporting that MerlinAU "downgraded" the installed F/W version. I don't recall reading about this same scenario happening with RMerlin's F/W releases at all. In any case, it's always good to add code that double-checks & verifies that the correct F/W release version is in fact being installed. Excellent job, bud! |
Thank you! Yes i think it's fairly rare, but we should probably build in a "safety" against it. I won't lie, it really confused me to wake up to my router DOWNGRADED by MerlinAU hahaha
Agreed! |
|
Btw check SNB, spdMerlin reports. I'm troubleshooting as we speak |
Minor Fixes
Seems to me what happened at some point for my RT-AX92U router is @gnuton built the firmwares; (lets say at midnight) and his Github action automatically published it. Shortly after it was automatically published, the Webs-Update.sh script ran and found the update.
MerlinAU then ran a few hours later by cron job, by that point the builds were hidden. (Gnuton had unpublished or hidden the builds by 3AM) But because they were published and the webs_update.sh script detected it, and started to flash.
So instead MerlinAU downloaded the latest 388.8_2 and downgraded me from the latest alpha lol.
This additional check will cancel the script if this happens again and the versions don't match. Along with additional Gnuton specific changes to make sure we grab the right version by parsing for matching version and matching Product ID.
New Checkmarks on WebGUI #395 (comment)
In short; we shouldn't be using:
_GetLatestFWUpdateVersionFromRouter_ 1because that does not split on the dash. That was a me mistake when I implemented this I think?While
_GetLatestFWUpdateVersionFromRouter_does split the version information at the dash for things like rog, or -gnuton.