You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have been rewriting my scripts that automate building and testing Elvish on my primary server and VMs from Bash to Elvish. When I went to rewrite my ebuild-nodaemon script I noticed it no longer worked due to changes to the Makefile and Elvish source with regard to how build info (e.g., the Git commit ID) is embedded in the Elvish binary. Which caused me to notice that the existing CI configurations do not validate that the ./cmd/nodaemon/elvish variant builds, let alone passes relevant unit tests.
This obviously also applies to targets like ./cmd/withpprof/elvish. However, that specific build target is less likely to suffer bit-rot and therefore probably does not need to be regularly verified by the CI environments.
I can't help but point out that this is yet another argument for reimplementing the interactive history mechanism to not need a daemon. See issue #1222. So rather than adding verification of the ./cmd/nodaemon/elvish variant to the CI configurations it makes more sense to invest the time to replace the dependency on BoltDB, which necessitates a daemon, with a simple flat-file history file.
The text was updated successfully, but these errors were encountered:
I have been rewriting my scripts that automate building and testing Elvish on my primary server and VMs from Bash to Elvish. When I went to rewrite my
ebuild-nodaemon
script I noticed it no longer worked due to changes to the Makefile and Elvish source with regard to how build info (e.g., the Git commit ID) is embedded in the Elvish binary. Which caused me to notice that the existing CI configurations do not validate that the ./cmd/nodaemon/elvish variant builds, let alone passes relevant unit tests.This obviously also applies to targets like ./cmd/withpprof/elvish. However, that specific build target is less likely to suffer bit-rot and therefore probably does not need to be regularly verified by the CI environments.
I can't help but point out that this is yet another argument for reimplementing the interactive history mechanism to not need a daemon. See issue #1222. So rather than adding verification of the ./cmd/nodaemon/elvish variant to the CI configurations it makes more sense to invest the time to replace the dependency on BoltDB, which necessitates a daemon, with a simple flat-file history file.
The text was updated successfully, but these errors were encountered: