14048 2020-09-18 08:07:52 +0000 devtool: new recipe addition still keeps using the previous recipes variables in the new one 2020-10-29 16:25:17 +0000 1 1 7 Build System, Metadata & Runtime OE-Core devtools / tool chain unspecified x86 Multiple RESOLVED FIXED Medium+ normal 3.2 M4 1 shantanoo_desai shantanoo_desai meta.mr.watcher meta.watcher randy.macleod --- --- No (bug/feature does not impact docs) oldest_to_newest 88173 0 shantanoo_desai 2020-09-18 08:07:52 +0000 ## System Build Host: Debian 9 Stretch Target Hardware: PHYTEC PhyBOARD Mira - iMX6 Quad Board Open-Embedded branch: warrior Bitbake Version: 1.42.0 ## Initial Setup - I ported my python3 app using the following command using `devtool`: devtool add python3-gpsfluxlite <github archive tarball> which create the recipe for me in my $BUILDDIR/workspace/recipes/python3-gpsfluxlite - I only changed the `RDEPENDS_${PN}` from `python-depname` to `${PYTHON_PN}-depname` in the recipe and added `systemd` installation script for the app. - The build was successful and I wrapped up the recipe in my layer using `devtool reset python3-gpsfluxlite` && `devtool finish python3-gpsfluxlite ../to/my/layer` and it worked well ## Problem - I wish to port another python3 app and I followed the same development workflow as above: devtool add python3-bnofluxlite <github archive tarball> This time the Terminal kept mentioning the packages and dependencies still from gpsfluxlite (see logs): INFO: Scanning paths for packages & dependencies: gpsinfluxlite, bin/gpsinfluxlite - upon editing the recipe `devtool edit-recipe python3-bnofluxlite` I still see remnants from the `gpsfluxlite` recipe (Pastebin link: https://pastebin.com/KmyENMRh of generated recipe) e.g. `S = "${WORKDIR}/gpsfluxlite-${PV}"` as well the `RDEPENDS` from the previous recipe. The only workaround is to clean up the workspace and use `--srcrev` with `devtool add` to generate the correct recipe. Seems like `devtool` isn't picking up or changing the downloaded tarball 88212 1 randy.macleod 2020-09-24 07:40:40 +0000 Can you retest with master? Some changes to datastore have been made that may fix the issue. 88515 2 shantanoo_desai 2020-10-29 16:25:17 +0000 Hi Randy, I might be unfortunately late to reply to this issue. I am bound by the Board Provider's (PHYTEC) environment. I might checkout the `master` branch and see if things don't break. If there are changes to the datastore on the master branch, I would like to mark this issue as RESOLVED. thanks, Shan