Stream: committers
Topic: IGPublisher source code
Grahame Grieve (Apr 18 2019 at 21:47):
I'm in the process of moving the IGpublisher source code to Maven (with help from @James Agnew - thanks). The IG publisher code in the main build is now orphaned - do not make any changes to it . But note that the maven code is not up and go yet
Grahame Grieve (Apr 18 2019 at 21:47):
the new repo is https://github.com/HL7/fhir-ig-publisher
Rob Hausam (Apr 18 2019 at 21:48):
when do you expect the transition to be complete?
Grahame Grieve (Apr 18 2019 at 21:48):
in the next few days
Grahame Grieve (Apr 25 2019 at 01:52):
great. it's up and going, much thanks to @James Agnew. The new location for the latest copy of the IG Publisher is now
Grahame Grieve (Apr 25 2019 at 01:52):
Grahame Grieve (Apr 25 2019 at 01:52):
@Josh Mandel what do we need to do to switch the autobuild over to this location?
Josh Mandel (Apr 25 2019 at 02:55):
We'll need to update the docker image at https://github.com/hl7-fhir/auto-ig-builder/blob/master/images/ig-build/builder/builder.py#L36 and then push to the registry.
Josh Mandel (Apr 25 2019 at 02:57):
(we might want to take the source URL for the IG Publisher in as an environment variable rather than hard coding it, too ;-))
Josh Mandel (Apr 25 2019 at 04:20):
What's the level of urgency on this? I'm planning to sleep first :-)
Grahame Grieve (Apr 25 2019 at 04:52):
sleep first is ok.
Grahame Grieve (Apr 25 2019 at 04:52):
don't know what it would be an environment variable of? the setting has to be somewhere.... but the one place we don't want it to be is on the github hook
Josh Mandel (Apr 25 2019 at 12:32):
Oh, I just meant an environment variable for the docker container -- it gets passed in via https://github.com/hl7-fhir/auto-ig-builder/blob/master/triggers/ig-commit-trigger/job.json#L31 . Certainly don't want to create more work (or reactivate changes) for IG authors :-)
Grahame Grieve (Apr 25 2019 at 12:43):
I don't mind where it goes into the system
Grahame Grieve (Apr 25 2019 at 12:43):
as long as we control it
Josh Mandel (Apr 25 2019 at 13:10):
OK, just updated the docker image. Let's see what happens.
Josh Mandel (Apr 25 2019 at 13:10):
Hitting rebuild on "cqframework / draft-measures".
Josh Mandel (Apr 25 2019 at 13:13):
So, a couple of issues here :-)
Josh Mandel (Apr 25 2019 at 13:13):
https://build.fhir.org/ig/cqframework/draft-measures/build.log
Josh Mandel (Apr 25 2019 at 13:14):
Does
Caused by: java.lang.ClassNotFoundException: org.commonmark.ext.gfm.tables.TablesExtension
ring a bell?
Grahame Grieve (Apr 25 2019 at 13:17):
well loosely, but I do not undestand why commonmakr is found, but not gfm. It looks like their definition pathway is the same. @James Agnew ?
Josh Mandel (Apr 25 2019 at 13:20):
Similarly:
Caused by: java.lang.NoClassDefFoundError: com/github/rjeschke/txtmark/Processor
just occurred in https://build.fhir.org/ig/ahdis/ihe-mhd-ig/build.log
Grahame Grieve (Apr 25 2019 at 13:20):
hmm. when the autobuilder runs the ig, does it point at {root} or {root}/ig.json as the parameter? it looks like the second?
Josh Mandel (Apr 25 2019 at 13:21):
let's look...
Josh Mandel (Apr 25 2019 at 13:21):
https://github.com/hl7-fhir/auto-ig-builder/blob/master/images/ig-build/builder/builder.py#L51
built_exit = do(['java', '-Xms%s'%java_memory, '-Xmx%s'%java_memory, '-jar', '../publisher.jar', '-ig', 'ig.json', '-auto-ig-build', '-target', 'https://build.fhir.org/ig/%s/%s/'%(details['org'], details['repo']), '-out', clone_dir], clone_dir)
Josh Mandel (Apr 25 2019 at 13:22):
It's just passing -ig ig.json. Is that not what's expected?
Grahame Grieve (Apr 25 2019 at 13:22):
hmm. I will have to do a new release for some other reason then
Grahame Grieve (Apr 25 2019 at 13:35):
k sorted that issue. I will do a new release with a crude attempt to fix the markdown problem.
Grahame Grieve (Apr 25 2019 at 14:14):
got a bit further in this time
Grahame Grieve (Apr 25 2019 at 14:17):
I'll do another release, but I'll be asleep before this one is complete
Last updated: Apr 12 2022 at 19:14 UTC
