Stijn van Dongen

Additional comment: the IRODS errors I am tracking above may be transient; if I rerun the pipeline hours or a day later this stage may then succeed. However, in this...

That looks good. In `onError`, would the code still have access to files produced by the `script` section? I don't have a scenario, but perhaps worth considering?

Just looking at our rnaseq pipeline, I have this tracking in two places, but need it in at least two more places, and generally I would probably use this type...

This is a fairly basic question I fear; with the above code, so ``` onError: if( task.exitStatus==64 ) ch_lostcause_irods

I'm a bit slow ... I want to collect all the values from the channel, stick it into a file, and ideally do 1) publish it 2) send it also...

Just to remark that in Simone Coughlan's talk at Nextflow 'Reproducible In silico Genomics' it seemed that she is doing similar accounting tasks in Nextflow code: https://github.com/coughls/nextflow-example-scripts . Edit: tagging...

Allright, it sounds like a principled approach. I'll stay tuned to see what happens. This could enable me to create a report as sketched above; but if it is something...

This is working at the moment, as custom content. This is example data that I create using the lostcause channels: ``` # plot_type: 'table' # section_name: 'Lost samples' Sample Process...

A short update: After the above I've concluded that what I'd like most is to be able to _catch unforeseen process failures_, e.g. a program crashing or not producing an...

A further thought (experiment). A very general solution in this problem space could be the option to specify that a channel can still be activated when a process fails. This...