It seem there may be a bug in the processing of profiles. I have nf-core/rnasplice installed in

$pipelinedir

(via ), with

nf-core download ...

, with the custom institutional configs installed as well, and I have added my own profile to those. • When I run

nextflow run "$pipelinedir" -profile test,singularity --outdir "$pipeoutdir"

, it works as expected (but of course uses the

local

executor). • When I try to use my own profile together with

test

(

nextflow run "$pipelinedir" -profile test,my_cluster ...

, the order of profiles makes no difference), however, it seems to read

my_cluster.config

OK, but crashes with an error saying that it cannot find a file

test.config

in the directory of the custom profiles. I.e., it ignores the

test

profile in

nextflow.log

but looks for it where my custom profile is located. Is this a known issue?

Never heard that one before. Can you detail a bit more what you did, or even better make a minimal reproducible example?

OK, I figured out what was wrong -- I didn't have my own profile properly installed in the config directory. But there is still a bug, because the error message didn't tell me that my profile was missing, it complained about the built-in

test

profile ... This happens when a profile is listed in the

nfcore_custom.config

, but the actual file it points to does not exist (in that location). What exactly happens in that case depends: • If the missing profile is the only one or there is another profile from

nfcore_custom.config

(which can be properly installed), then the error message is about something not applicable at all:

N E X T F L O W  ~  version 23.04.1
ERROR ~ Plugins definition is only allowed in config top-most scope

 -- Check '.nextflow.log' file for details

• If there is another profile specified, and that profile is a built-in one from

nextflow.config

of the pipeline itself, e.g.

test

, then the error is

N E X T F L O W  ~  version 23.04.1
ERROR ~ No such file or directory: Config file does not exist: /path/to/nf-core-rnasplice-1.0.3/1_0_3/../c
onfigs/conf/test.config

 -- Check '.nextflow.log' file for details

As you can see, neither scenario results in a reasonable error message. I was using Nextflow 23.04.1, but I got the same error with 22.10.6 (this is what I have installed, and I don't have time to test more at the moment). Is this a bug in Nextflow itself or in how the institutional custom configs are handled by the pipeline?