zeroshade commented on a change in pull request #11566:
URL: https://github.com/apache/arrow/pull/11566#discussion_r738703107
##########
File path: go/parquet/writer_properties.go
##########
@@ -18,7 +18,7 @@ package parquet
import (
"github.com/apache/arrow/go/arrow/memory"
Review comment:
so that's where things get a bit more complex and i'm still trying to
figure out the best way to go about it.
Technically it should get updated to point at the v6, but we can't do that
correctly (and update the go.sum) until the arrow change is done and there is a
tag go/arrow/v6.0.# etc. which points to a branch where go/arrow/go.mod has the
v6. After the tag exists pointing at the v6, we can then update the parquet
dependency on Arrow via `go get`. Basically, we'd need the arrow and parquet
tags to point at different commits so that the arrow tag would already exist
when we update the parquet version tag.
All this is because the arrow and parquet directories specify separate
individual modules and have their own go.mods. The alternative would be to
shift the go.mod up one directory and making the arrow and parquet modules just
two packages of the same module which would have us versioning them together
instead of versioning them separately (which is potentially desireable). Any
drawback would only exist in development in that anyone who wanted to only
depend on arrow would *also* depend on the parquet lib too from a dependency
standpoint, even if the files wouldn't get included when compiling a project
that didn't use the parquet library at all.
so, what're your thoughts?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
