mwinkel-dev

Hi @sflanagan -- I am very embarrassed by yet another failure. This time, it was a failure of communication on my part to explain what is causing the database proxy...

Hi @sflanagan -- And unfortunately, I am also able to reproduce the `mdsvalue` error. This is likely a latent bug that has been present in the IDL API for years....

There are two different failure modes: one is likely easy to fix, one is more complicated. **Simple** The `mdsvalue` error can be triggered immediately by using the following sequence. ```...

Surprisingly, the `mdsconnect` routine was also doing a `mdsdisconnect` in one common scenario. (Apparently that was being done to create the `!MDS_SOCKET` system variable?). Commented out that `mdsdisconnect` and used...

Fix v3 (PR 2632) was not robust. It failed to preserve the database proxy established by `set_database`. The root issue again was simply that IDL's `keyword_set()` has slightly different behavior...

Hi @sflanagan and @ModestMC -- Fix v4 is now available as PR 2635. Josh, Stephen and I all tested the fix. However, our testing was only on Ubuntu. (We do...

Hi @smithsp -- Thanks for suggesting that I obtain access to GA's RHEL8 server, in case it is needed for testing future PRs. My understanding (perhaps incorrect) is that the...

Hi @sflanagan and @ModestMC -- I just noticed that Josh reviewed and approved PR #2635 (aka Fix v4). So I have just merged it into the `alpha` branch. Although today...

In the process of fixing Issue 2625, the investigation / testing uncovered these additional issues: - The data integrity check of the connection (socket) linked list was not being executed...

This issue has also generated the following . . . PR #2643, Issue #2650 and PR #2656.