*<217 new commits> pushed to <https://github.com/e...
# openlane
g
217 new commits pushed to `master` `03f1f06d` - config: If PDK_ROOT exists in the enviornment, don't overwrite it `d1f6b8cd` - flow.tcl: Use whatever default TCL version instead of fixed 8.5 `7ce6c412` - scripts/libtrim.pl: Don't warn on experimental `f454e819` - scripts: Use proper PDK_ROOT when generating PDN `16018c13` - cts skipping fix `370ebc53` - initial ms library configs `b5c5bee7` - initial hs library configs `62ae8fbc` - initial ls library configs `e182b470` - initial hdll library configs `ddc3819c` - removed csv2html options until the new script is added `6ab59a3c` - minor README.md update `805ccc62` - Merge remote-tracking branch 'origin/master' into develop `44294c86` - update --defaultTestSet `98ff2240` - missing design configurations added `96bc5458` - Merge branch 'develop' into minor-fixes `212ba615` - Merge pull request #7 from smunaut/minor-fixes `9f9903a6` - libtrim.pl fix `45ee1c40` - ROUTING_CORES base commit `166f5b31` - README updates `f823d444` - Merge pull request #14 from efabless/develop-ROUTING_CORES `43f50a9c` - readme updates + ROUTING_CORES=4 `59b5abf8` - Added HTML preview links for the regression sheets `ad0075dd` - remoeEnvVar.sh doesn't change global configs `72c3b7a6` - HDLL html added `a2712953` - minor readme updates `6d0ea591` - csv2html util under run_designs.py `76b787f0` - added --excluded_designs option for run_designs.py `40048ed2` - Added openlane Makefile `c345c3e5` - HTML sheets rename `64f6eb36` - compare_regression_reports.py base `091b1227` - regression_test not integrated `46de0e39` - order preservation after designs exclusion `d29e33fc` - benchmark compare integrated into run_designs.py `01026732` - regression_test target added to the main Makefile `b8e27adc` - HDLL benchmark sheet update `fa73b2cd` - digital_pll name fix `28c0d015` - Magic updated to 8.3.43 `d4f670ad` - PDK_VARIANT to a more proper name STD_CELL_LIBRARY `174db1f6` - Some diagnostic messages & error checking `65a40b07` - LIB_MIN and LIB_MAX deprecated `e094af91` - Merge branch 'develop_new_magic' into develop `3baa773f` - Updated to the latest version of netgen `2205555f` - Merge branch 'develop_fixed_netgen' into develop `05d2675e` - Using TritonRoute14 by default `4cf4ec77` - Arguments passed to prep in interactive mode `81430384` - Merge remote-tracking branch 'origin/master' into develop `61aa11a0` - Updated magic and netgen `3c0b3547` - designs sky130_fd_sc_hd configs updatedt `2d0ef1d9` - designs sky130_fd_hdll configs updated `0d94718c` - handle_deprecated_command in the utils package `a418e29f` - Some enhancements/fixes for synth-related utils `94bd71aa` - Updated synth and LEC scripts `f8bb9be5` - Added run_lvs command and LVS_INSERT_POWER_PINS `c480afca` - Develop latest pdk (#23) `c5ca35ca` - minor README updates `2e6c9118` - default Makefile THREADS=8 & BM64 hd config update `aedf9c0a` - fixed yosys.log being overwritten … efabless/openlane
r
"217 commits pushed" I hope you understand that nobody else is going to review this for mistakes until it's already breaking their project.
a
We're hoping that the CI tests/benchmarks passing on around 30-40 designs means it's not breaking for a wide range of cases, but, of course, the coverage is never perfect. We will try not leave master that far behind again. Do you have other specific suggestions? Subscribing to commits on the develop branch here for example?
a
I'd say the safest way for constant reviews is subscribing
staging
since it's considered "the latest stable version of develop". It's more tested and it's not frequently updated like develop (we merge develop into it every couple of days based on the results of the CI tests). So, it will be a review every ~10 commits or so, instead of ~220. What do you think about that?
r
Yeah, telling people to look at staging for review sounds good.
Maybe set up the GitHub bot to announce pushes to staging?
a
@Riking28: We looked it up and found we could either subscribe to all commits on all branches or subscribe to the default branch (master) only. So, we went with all commits on all branches, at least until someone complains about being bombarded with slack messages. If you're aware of a way/a workaround to subscribe to master and staging only, please let us know!
r
Let's see how that goes! Good luck?