Stream: shorthand
Topic: FSH fail
Grahame Grieve (Feb 03 2021 at 20:28):
@Chris Moesel here's my FSH output:
Sushi: ╔════════════════════════ SUSHI RESULTS ══════════════════════════╗ (00:05.0223)
Sushi: ║ ╭──────────┬────────────┬───────────┬─────────────┬───────────╮ ║ (00:05.0224)
Sushi: ║ │ Profiles │ Extensions │ ValueSets │ CodeSystems │ Instances │ ║ (00:05.0225)
Sushi: ║ ├──────────┼────────────┼───────────┼─────────────┼───────────┤ ║ (00:05.0225)
Sushi: ║ │ 18 │ 4 │ 11 │ 2 │ 17 │ ║ (00:05.0226)
Sushi: ║ ╰──────────┴────────────┴───────────┴─────────────┴───────────╯ ║ (00:05.0226)
Sushi: ║ ║ (00:05.0227)
Sushi: ╠═════════════════════════════════════════════════════════════════╣ (00:05.0227)
Sushi: ║ Just keep swimming, Dory. 3 Errors 0 Warnings ║ (00:05.0228)
Sushi: ╚═════════════════════════════════════════════════════════════════╝ (00:05.0228)
Grahame Grieve (Feb 03 2021 at 20:28):
then, FSH returns error code 1 to the IG publisher
Grahame Grieve (Feb 03 2021 at 20:28):
Process exited with an error: 1 (Exit value: 1)
Grahame Grieve (Feb 03 2021 at 20:29):
this is running the main branch of https://github.com/hl7/fhir-spl
Chris Moesel (Feb 03 2021 at 20:58):
I think you need to update your version of SUSHI. When I run this using SUSHI 1.0.2, I get those errors. But if I run it using SUSHI 1.1.0, it succeeds. To update SUSHI locally: npm install -g fsh-sushi
Grahame Grieve (Feb 03 2021 at 21:00):
yes that was it, though it's still a surprising outcome to fail in that fashion
Chris Moesel (Feb 03 2021 at 21:07):
I'm not sure what you mean. What is surprising about the way that it failed?
Grahame Grieve (Feb 03 2021 at 21:23):
it said it succeeded, but didn't actually succeed. There were no apparent errors in the output. If the version was wrong, one would expect some obvious error where it failed, rather than claiming success and then actually failing
John Moehrke (Feb 03 2021 at 21:28):
another one hurt by old sushi
Chris Moesel (Feb 03 2021 at 21:39):
Oh, I think I see? It didn't really "succeed" (as evidenced by the exit code), but SUSHI always does a best effort to complete as much of the processing as it can. So, it produced a bunch of profiles and stuff, but it also is telling you that there were errors so you don't incorrectly assume the outputs are 100% correct.
Isn't that similar to how the publisher works? While there are some fatal errors that can occur, often the publisher reports errors but still completes the overall process. I see that as very similar to what we're doing w/ SUSHI. If the publisher completes, but completes with errors, do you still return exit code 0 despite the errors?
Last updated: Apr 12 2022 at 19:14 UTC
