option to "keep going" when entry connection fails

[khimaros]

Problem to solve

i'd like to test a number of endpoints from a single hurl file and continue testing other entries when one or more of them have a connection timeout.

Proposal

add a --keep-going flag to the command line or a configuration option which can be specified on a per-entry basis. i see benefits to either, with the former being less work when there are many entries which should run through to the end regardless, and with the latter being a bit more flexible.

Additional context and resources

when a single endpoint times out, i'd still like to see the success report eg. 7/8 entries passed

[jcamiel]

Hi, we have --fail-at-end option that continue the execution of an Hurl file when there are asserts errors. I don't know what we do in case of timeout but it's worth trying.

[khimaros]

@jcamiel that did the trick, thanks!

i interpreted that option to mean "always fail at the end, even if all tests pass" :)

perhaps the flag description in --help could be improved?

[jcamiel]

Agreed, I had to double checked myself

[fabricereix]

we could maybe have --ignore-assert-error

[jcamiel]

=> agreed on continue-on-error (like GitHub Action), we deprecate --fail-at-end => #1739