rahul_anand
07/22/2020, 10:27 PMdouwe_maan
07/22/2020, 10:45 PMWas not the idea to benefit from the Singer Taps and Targets?Definitely! We'd rather not fork any taps or targets at all and use the existing versions supported by the community (or in the case of those in the
singer-io GitHub org: the Singer team/people at Stitch).
In this case it looks like we forked it because it was incompatible with tap-salesforce (https://gitlab.com/meltano/target-csv/-/issues/1), and while we submitted that bug upstream as well (https://github.com/singer-io/target-csv/issues/14), we presumably wanted to fix it in Meltano right away without having to wait for Singer to accept our fix (https://gitlab.com/meltano/target-csv/-/merge_requests/1) into their repo and release a new version.
About a month after we fixed it in our fork, the fix in the original repo was released as well (https://github.com/singer-io/target-csv/pull/18, https://github.com/singer-io/target-csv/releases/tag/v0.3.0) so there's no longer a need for our fork of target-csv to exist at all. In general, these forks should be short-lived and be cleaned up when the upstream catches up, but it looks like we neglected to do so in this case.
I've created an issue to deprecate our fork and change the pip_url for target-csv back to git+<https://github.com/singer-io/target-csv.git> so users benefit from the latest changes, including the new destination_path setting you mention: https://gitlab.com/meltano/meltano/-/issues/2197douwe_maan
07/22/2020, 10:46 PMtarget-csv . Let me know if you'd like some help getting that to work 🙂