Conversation
Also, only define `VERSION_TAG` if VERSION_TAG exists.
|
Very nice, thank you Peter for looking into this and improving it even further! I tested with physical RadioMaster TX16S mkII and Zorro. First, when flashed with the firmware from this PR/GitHub to TX16S: The same codebase, built locally and flashed to the radio: Bootloader flashed with the binary from this PR/GitHub: Bootloader display output, while the binary from this PR/GitHub is selected: Bootloader, while the self-built binary from this PR is selected: The screenshots in the same vain with RadioMaster Zorro: Should we go and create a test-tag, that we remove afterwards to see the output (not that I doubt your local build-gh.sh tests)? |
|
Sounds good. It's the only way to test if the tag processing in |
|
This is what a tagged version will look like (I deliberately tagged latest commit as I think the only thing left now is to make sure the about screens do the same or similar, and also check Companion. |
|
Today's nightly on TX16S: Presently says Do you think, we should tweak: Lines 61 to 65 in 3493d69 to say nightly?
|
|
|
* fix: Only define SUFFIX if PR or BRANCH build Also, only define `VERSION_TAG` if VERSION_TAG exists. * fix: Omit GIT SHA for colorlcd also if a tagged build















Further builds on #1850, #1887
Summary of changes:
build-gh.shwon't unconditionally setEDGETX_VERSION_SUFFIXon selfbuildsFrom the testing I've done so far (by overriding
EDGETX_VERSION_TAGandEDGETX_VERSION_SUFFIXinbuild-gh.sh), I know the whole pipeline frombuild-gh.shcorrectly showing the tag, or not showing the tag as required, will just need to do another test build with a tag once this is merged to ensure it is being picked up.I was tempted to also have VERSION_SUFFIX only defined on existence, but the code logic atm unconditionally expects it to exist if VERSION_TAG doesn't exist, so I left it.
@rotorman