Latest firmware version?

Posted on
  • Hi. I feel I must be missing something. I went to update the Bangle.js firmware in case it improved my Gadgetbridge connection issues. When I connect to the watch with the web IDE it reports the version as 2v04.294. When I look at http://www.espruino.com/binaries/travis/­master/ the latest build (dated today) is 2v04.84 which sounds like it's older. Am I misunderstanding something?

    Andrew

  • Fri 2020.01.10

    Hi Andrew @Andrewmk maybe a computer ordering vs human visible ordering issue?

    It's likely this was the starting page:

    http://www.espruino.com/Bangle.js#firmwa­re-updates

    from which page may be found:

    https://www.espruino.com/Download#bangle­js

    Beneath 'Find a Binary' DDL the specific firmware update for the desired board type may be selected.

    Note: A week ago, this page indicated a visible link of 2v04.284 and the current 'Cutting Edge' build of 2v04.84
    As Gordon etal slave away at improvements, this page will constanlty update

    The change log may be found at:

    https://www.espruino.com/ChangeLog

    Further complicating things is that the order of the firmware changes is by a unique GUID assigned index and not the release date. This means to find a build from December 2019, one will need to hunt down the page for the release date:

    http://www.espruino.com/binaries/travis/­



    When one aligns decimal numbers, one may view the difference between how computers would sequence, in comparrison to how humans logically sequence. Remember the trailing zero (invisible) place holders to the right of the decimal in any base numbering. A trailing zero is not considered a significant figure:

    Too many digits: the presentation of numerical data

    Tons of articles on version numbering conventions:

    Naming and version numbering conventions



    2v04.294
    2v04.840

    So, using the firmware updates you provided, shows that 840 is a larger value than 294 further proving the cutting edge build is more recent than the firmware installed on your device.

    Searching down that page, one may see that there were two releases of the latest cutting edge build today:

    2v04.81
    7001b3adf8e1e2882d99df5cbdb8a08d3bead589­/ 2020-01-10 16:14

    http://www.espruino.com/binaries/travis/­7001b3adf8e1e2882d99df5cbdb8a08d3bead589­/

    2v04.84
    63587f2b6c01456461401ba503e1c936de2573cc­/ 2020-01-10 16:30

    http://www.espruino.com/binaries/travis/­63587f2b6c01456461401ba503e1c936de2573cc­/

    It would take some trial and error Hi-Lo date estimating to locate the actual date for the 2v04.294 version.

  • I don't think so. "espruino_2v04.84_banglejs.zip" is a filename on the firmware page so if it meant "espruino_2v04.840_banglejs.zip" it could say "espruino_2v04.840_banglejs.zip". Hopefully @Gordon or someone else involved can comment.

    Andrew

  • I think .84 is older than .294. It is number of commits after latest release, see https://github.com/espruino/Espruino/blo­b/master/Makefile#L98
    Current changelog is here https://github.com/espruino/Espruino/blo­b/master/ChangeLog

  • Sat 2020.01.11

    That method #2 post seemed consistent with revisions of Pico I followed two years ago. Made the assumption was still the means today.

    While I'll agree that L98 is coded the way you describe, the only line that makes sense is L94-L95

    'This may fail and give inconsistent results'

    proven by the 8th date stamping for 2v04.83
    383dfcc910ef082f4dfbf718277dbab2fcd9d924­/ 2020-01-08 11:03

    http://www.espruino.com/binaries/travis/­383dfcc910ef082f4dfbf718277dbab2fcd9d924­/

    which occurs on th 7th before 2v04.84
    4da6614bdda20162af5ea845b0d8f73804da6c31­/ 2020-01-07 16:47

    http://www.espruino.com/binaries/travis/­4da6614bdda20162af5ea845b0d8f73804da6c31­/

    Very unlikely the incremental commit decreases as date/time moves forward.

    L94 reigns and does fail as documented.

    Only spent a few minutes attempting to locate 2v04.294 with no luck.

  • In the binaries/travis the directory names are github commit hashes.
    You can check out the commits at https://github.com/espruino/Espruino/com­mits/master The latest commit in master is 3641941. If you search for that, you find .82 in that directory. But that's indeed not the absolute latest. There are newer commits in the add_quadBezier branch.
    Altho' it's not completely valid, since the latest in the travis binaries is 795475, and I don't see that commit...

  • Ah, I see (some of) my confusion now. I was looking in http://www.espruino.com/binaries/travis/­master which of course only contains builds from the master branch, and if you're following 'normal' Git workflow you only commit to master for a release build (at least that's what we do at work) not for nightlies.

  • I believe 2v04.84 is the latest (right now) even though the numbers are different.

    For Espruino master is basically the development branch. It's intended to always stay building and working, but things get added all the time and that's what the travis/cutting edge builds come from. Releases are tagged with RELEASE_...

    The number after 2v04 is calculated from the number of commits since the latest tag beginning with RELEASE_.

    The problem comes because I'm using Travis to do the builds, and I think it must do something like only partially downloading the GitHub repo, because git log --oneline RELEASE_2V04..HEAD doesn't seem to be returning all the commits that are actually there!

  • Post a reply
    • Bold
    • Italics
    • Link
    • Image
    • List
    • Quote
    • code
    • Preview
About

Latest firmware version?

Posted by Avatar for Andrewmk @Andrewmk

Actions