LGTM for graphics/rawtherapee.

It is a lot of preparation fixing all the fallout of switching. There are lots of util-mini tools that break the system one way or another.

All for going UTC.

I am fine with mentioning pitfalls and limitations objectively by pointing to the detail sections. I reject "recommendations", and if we have those in other manual pages, they need to be reviewed and reworded in a similar vein or removed.

In D45073#1038323, @jrm wrote:

I lean towards something stronger.

Relying solely on
.Fl e
to check for failures is discouraged.
Many commands return non-zero values to convey information other than errors, which can cause unexpected program termination.
Thus, it is recommended to explicitly check for errors.

It hedges on outright discouraging set e with "Relying *solely* on set -e..."

Manual pages are reference material and not a place for advocacy. Proper use of the feature, which seems to document its drawbacks (checked function call etc.), is a duty of the one writing a shell script. We don't have Python in base, so we can't recommend that.

Since I haven't had my hands on the script, nor am I volunteering to become its maintainer, I am not going to claim reponsibility for it. I think my stance is clear, "automatically fail everywhere" and handle those few expected errors, and let's remove @jilles's manpage addition (that scares users away from set -e in the sh(1) man page) because it's not substantiated. manpages aren't the place to scare people without giving reasons. We're early enough after a freshly forked quarterly release to clean up fallout over the next ten weeks, so... perhaps that fact makes your decision easier.

add one inline comment

In D42719#1020644, @jrm wrote:

@michaelo, thanks for sticking with this review. My preference is still to go with the || exit \$? with or without the set -e.