Nikolai Kondrashov

Hey, thanks, @lskillen, I'll take note of Cloudsmith :)

I think we'll implement it as the default, without requiring the config option. This was in #12 for a while.

Thank you, Nathan! Yes, I agree, this is a good list of conditions to handle. This brings me into thinking that we should test how failures to log to Journal...

This concerns modifying the `tlog_rec_get_id` function to produce a more compact ID.

This looks like something set the tlog's shell-to-start to an empty string.

The first part of this was done with the introduction of error stacks, only the assertions need to be added now.

This means asserting that `tlog_*_conf_validate` functions return `TLOG_RC_OK` on entering functions accepting corresponding configurations, and on exiting functions modifying them.

Functions extracting data from loaded configuration expect it to be correct. Adding an `assert` to the start of the function verifying that, will help us ensure that. Similarly, functions which...

Yeah, we don't need to assert things in `tlog_*_conf_validate` functions for this issue, but instead use those functions to validate the configuration in functions which create, access or modify it....

Thank you for the report @jeannegreulich! I think the documents might be confusing. Because tlog-rec-session works in the opposite way: it actively avoids recording a session, if anoter tlog-rec-session is...