Stream: committers
Topic: R4B Build Failure
Khalid Shahin (Mar 05 2021 at 14:25):
I try to do a clean build of R4B with no changes made and I get this error:
This was a Full Build 11.496 20sec 294MB
FHIR build failure @ Fri, Mar 5, 2021 08:56-0500 2.002 22sec 303MB
Error: Error loading \fhir\source\activitydefinition\valueset-action-participant-role.xml: URL mismatch. Expected http://terminology.hl7.org/ValueSet/action-participant-role but found http://hl7.org/fhir/ValueSet/action-participant-role
org.hl7.fhir.exceptions.FHIRException: Error loading \fhir\source\activitydefinition\valueset-action-participant-role.xml: URL mismatch. Expected http://terminology.hl7.org/ValueSet/action-participant-role but found http://hl7.org/fhir/ValueSet/action-participant-role
Lloyd McKenzie (Mar 05 2021 at 15:19):
@Grahame Grieve
Mark Iantorno (Mar 05 2021 at 20:06):
Grahame has done a few releases to the build library for R4B and main that haven't been updated yet because I'm stuck in the middle of something else right now
Mark Iantorno (Mar 05 2021 at 20:06):
I'll make the update right now
Mark Iantorno (Mar 05 2021 at 20:06):
and then later next week I'll update the pipelines to handle this automagically
Mark Iantorno (Mar 05 2021 at 20:06):
gimmy 5 min
Mark Iantorno (Mar 05 2021 at 20:06):
eh 10
Mark Iantorno (Mar 05 2021 at 20:12):
I lied, it's done now, I'm 98% sure this will fix a lot of the issues that people have been reporting this week. Grahame has been putting time into fixes for the build library (kindling)
Mark Iantorno (Mar 05 2021 at 20:12):
and his changes should now be included
Mark Iantorno (Mar 05 2021 at 20:13):
FYI, you will have to pull the R4B branch into your current branch to get those changes
Mark Iantorno (Mar 05 2021 at 20:13):
please do that
Khalid Shahin (Mar 06 2021 at 13:59):
Mark Iantorno said:
FYI, you will have to pull the R4B branch into your current branch to get those changes
I pulled R4B as is, and it still won't build for me.
Mark Iantorno (Mar 06 2021 at 18:28):
this is not a bug I can solve then, I would punt this over to @Grahame Grieve
Grahame Grieve (Mar 07 2021 at 06:04):
well, it's building for me. @Mark Iantorno can you release a new core, kindling, and then figure out why this PR isn't working?
Grahame Grieve (Mar 07 2021 at 06:05):
https://github.com/HL7/fhir/pull/1152
Mark Iantorno (Mar 07 2021 at 19:11):
Build works...but the upload fails due to
mv: cannot stat '/home/***/uploading/www/branches/gg-r4b-2021-03-qa-fixes-2': No such file or directory
find: './br-30808-detectedissue-maturity': No such file or directory
Mark Iantorno (Mar 07 2021 at 19:12):
I'm going to rerun it tomorrow
Grahame Grieve (Mar 07 2021 at 22:43):
I don't know what that means.
Mark Iantorno (Mar 08 2021 at 01:49):
when it tried to upload the file to the server, it tried to move a file that wasn't there? Might jsut be a pipeline error, might be more
Mark Iantorno (Mar 08 2021 at 01:49):
when I rerun we will know for sure
Josh Mandel (Mar 08 2021 at 03:04):
If the mv: cannot stat... output @Mark Iantorno quoted is coming from this line of the publication script:
mv ~/uploading/www/branches/$DEPLOY_TO_BRANCH ~/uploading/branch.$TARGET.old || true
... then it shouldn't be stopping anything from working; the || true should ensure that a missing branch folder isn't an error in the build. (The first time a new branch is deployed, the || true will always be invoked because the mv is designed to get rid of old content, and of course there is no old content on the first deploy.)
Grahame Grieve (Mar 08 2021 at 19:35):
so @Mark Iantorno where are we with this?
Mark Iantorno (Mar 08 2021 at 19:54):
Sorry was working on something else, re-running it now
Mark Iantorno (Mar 08 2021 at 19:54):
(deleted)
Mark Iantorno (Mar 08 2021 at 19:58):
Alright, I figured out what the issue is
Mark Iantorno (Mar 08 2021 at 19:59):
(deleted)
Mark Iantorno (Mar 08 2021 at 20:00):
(deleted)
Mark Iantorno (Mar 08 2021 at 20:01):
This is what is running now
Mark Iantorno (Mar 08 2021 at 20:01):
- task: Bash@3
inputs:
targetType: 'inline'
script: |
cd publish
chmod 400 $(Agent.TempDirectory)/deploy.rsa
TARGET_DIRECTORY=$(echo "$(System.PullRequest.SourceBranch)" | sed -r 's/\//_/g')
echo "Publishing to target $TARGET_DIRECTORY"
tar czf - * | ssh -i $(Agent.TempDirectory)/deploy.rsa -p 2222 $(BUILD_FHIR_ORG_USERNAME)@build.fhir.org ./publish $TARGET_DIRECTORY
Mark Iantorno (Mar 08 2021 at 20:03):
So there is no (1) pruning of old branches, or (2) continuing on failed upload of a directory
Grahame Grieve (Mar 08 2021 at 20:05):
so I'll do the merge.... it's holding me up.
Mark Iantorno (Mar 08 2021 at 20:05):
yeah if it builds it should be okay
Grahame Grieve (Mar 08 2021 at 20:06):
no, it won't let me
Mark Iantorno (Mar 08 2021 at 20:06):
oh yeah if the checks don't pass it will not
Mark Iantorno (Mar 08 2021 at 20:06):
one sec i will fix
Mark Iantorno (Mar 08 2021 at 20:08):
it's merged
Grahame Grieve (Mar 08 2021 at 20:08):
thx
Mark Iantorno (Mar 08 2021 at 20:11):
@Grahame Grieve are you getting that issue on any of your other PRs
Mark Iantorno (Mar 08 2021 at 20:11):
or is this the first time?
Mark Iantorno (Mar 08 2021 at 20:13):
@Josh Mandel I am unsure why this would be an issue with the current script?
Grahame Grieve (Mar 08 2021 at 20:13):
I hadn't had that problem before
Mark Iantorno (Mar 08 2021 at 20:13):
if I run the command
tar czf - * | ssh -i $(Agent.TempDirectory)/deploy.rsa -p 2222 $(BUILD_FHIR_ORG_USERNAME)@build.fhir.org ./publish $TARGET_DIRECTORY
Mark Iantorno (Mar 08 2021 at 20:14):
that should just upload it to the directory regardless of if that directory exists or not...no? Or am I totally incorrect here?
Mark Iantorno (Mar 08 2021 at 20:15):
I have all sorts of PR runs succeeding?
Grahame Grieve (Mar 08 2021 at 20:24):
Hey @Mark Iantorno is there any way for us to get the version of the kindling release in the code
Mark Iantorno (Mar 08 2021 at 20:24):
I'm not clear on what you mean
Mark Iantorno (Mar 08 2021 at 20:25):
you want to know the current version being used?
Mark Iantorno (Mar 08 2021 at 20:25):
As of right now, the project uses 1.0.0-SNAPSHOT
Mark Iantorno (Mar 08 2021 at 20:26):
and any updates committed to the master branch of kindling are on 1.0.0-SNAPSHOT
Mark Iantorno (Mar 08 2021 at 20:26):
sometime this week, I'm going to set up a pipeline to actually publish non-snapshot builds
Mark Iantorno (Mar 08 2021 at 20:26):
I just wanted to finish doing session based validation on the front end of the validator wrapper first
Grahame Grieve (Mar 08 2021 at 20:26):
the first think that the validator and java do is dump their own version to the log, so I know what version is being used.
Mark Iantorno (Mar 08 2021 at 20:27):
oh, yeah I can do that
Grahame Grieve (Mar 08 2021 at 20:27):
when people ask me to troubleshoot
Josh Mandel (Mar 08 2021 at 20:27):
that should just upload it to the directory regardless of if that directory exists or not...no?
that's certainly the intention (and if it doesn't exist, you'll see a line of loggin output that mv failed to move out the old stuff)
Mark Iantorno (Mar 08 2021 at 20:27):
for right now, you can just assume it's using whatever is on master
Mark Iantorno (Mar 08 2021 at 20:27):
because... it should be
Mark Iantorno (Mar 08 2021 at 20:27):
but I'll add that line in right now to assist
Grahame Grieve (Mar 08 2021 at 20:27):
I'm looking for an equivalent to IGVersionUtil.getVersionString()
Mark Iantorno (Mar 08 2021 at 20:27):
That is not difficult to do in gradle, I'll will do that now
Mark Iantorno (Mar 08 2021 at 20:28):
won't take long
Mark Iantorno (Mar 08 2021 at 20:28):
@Josh Mandel any idea why it might be failing...sometimes?
Mark Iantorno (Mar 08 2021 at 20:28):
really this is the first time it's been brought up to me
Josh Mandel (Mar 08 2021 at 20:28):
I don't; is there any logging output aside from the mv command?
Mark Iantorno (Mar 08 2021 at 20:31):
the only logging I do in the bash script is that one line
Mark Iantorno (Mar 08 2021 at 20:31):
echo "Publishing to target $TARGET_DIRECTORY"
Mark Iantorno (Mar 08 2021 at 20:32):
this is the whole upload log:
Mark Iantorno (Mar 08 2021 at 20:32):
Starting: Bash
==============================================================================
Task : Bash
Description : Run a Bash script on macOS, Linux, or Windows
Version : 3.182.0
Author : Microsoft Corporation
Help : https://docs.microsoft.com/azure/devops/pipelines/tasks/utility/bash
==============================================================================
Generating script.
========================== Starting Command Output ===========================
/bin/bash --noprofile --norc /home/markiantorno/azure-pipelines/_work/_temp/ba036623-f848-4617-87a4-9d9e698fecb4.sh
Publishing to target gg-r4b-2021-03-qa-fixes-2
491M .
932M .
1.5G .
mv: cannot stat '/home/***/uploading/www/branches/gg-r4b-2021-03-qa-fixes-2': No such file or directory
find: './br-30808-detectedissue-maturity': No such file or directory
##[error]Bash exited with code '1'.
Finishing: Bash
Mark Iantorno (Mar 08 2021 at 20:32):
the upload starts and almost finishes
Mark Iantorno (Mar 08 2021 at 20:33):
this is all server side, the only thing I really control is the command that pushes the ./publish directory up
Josh Mandel (Mar 08 2021 at 20:52):
I'm seeing:
find: './br-30808-detectedissue-maturity': No such file or directory
... and I'm not sure what's producing this line of output. That might be worth tracing down.
Josh Mandel (Mar 08 2021 at 20:54):
What's "br-30808-detectedissu-maturity"? https://github.com/FHIR/auto-ig-builder/blob/master/images/ci-build/publish#L35 is the only reference to find I see.
Mark Iantorno (Mar 08 2021 at 22:01):
@Grahame Grieve I just merged your requested log change to master and R4B
Mark Iantorno (Mar 08 2021 at 22:01):
people will have to pull R4B into their current branch to get those changes
Khalid Shahin (Mar 09 2021 at 14:40):
I'm still getting the same error message on a fresh pull with no changes. (R5 builds just fine)
\fhir\source\activitydefinition\valueset-action-participant-role.xml: URL mismatch. Expected http://terminology.hl7.org/ValueSet/action-participant-role but found http://hl7.org/fhir/ValueSet/action-participant-role
Mark Iantorno (Mar 09 2021 at 18:38):
which pull request?
Khalid Shahin (Mar 09 2021 at 20:06):
Mark Iantorno said:
which pull request?
Sorry, I meant clone. I cloned this as is and I'm unable to publish locally. https://github.com/HL7/fhir/tree/R4B
Bryn Rhodes (Mar 09 2021 at 20:38):
I tried this same thing and was able to get a successful build.
Bryn Rhodes (Mar 09 2021 at 20:38):
I do see that the URL is "hl7.org/fhir/ValueSet/action-participant-role" just like it's complaining about, but for some reason mine doesn't complain about that. :(
Khalid Shahin (Mar 09 2021 at 20:39):
Bryn Rhodes said:
I do see that the URL is "hl7.org/fhir/ValueSet/action-participant-role" just like it's complaining about, but for some reason mine doesn't complain about that. :(
Maybe it has to do with a dependency issue?
Grahame Grieve (Mar 10 2021 at 07:06):
@Rik Smithies what is the status of this? https://github.com/HL7/fhir/pull/1123
Rik Smithies (Mar 10 2021 at 07:54):
It's not needed, I've closed it. It's from when image files were accidentally excluded and is no longer an issue. All my stuff is in, except for some issues Lloyd raised yesterday about documentation of the new datatypes e.g. CodeableReference. Should be in soon - but I don't know how to add to the list of special datatypes that is specified using [%wildcard-type-list%]
Rik Smithies (Mar 10 2021 at 07:59):
By the way the build seems to always make broken XSDs, for a long time, but they seem worse in R4B. I was unable to fix them manually, which I can normally manage by commenting out some duplicates. Not sure if it is feasible to fix that in the time available.
Grahame Grieve (Mar 10 2021 at 10:43):
probably not. No one had let me know till now
Grahame Grieve (Mar 10 2021 at 10:44):
what duplicates do you have to comment out?
Rik Smithies (Mar 10 2021 at 11:37):
there is a long standing Jira about it
Rik Smithies (Mar 10 2021 at 11:39):
an example is ActivityDefinition <xs:simpleType name="RequestResourceType-list">
When you load the full schema set that is in there twice, so you have to comment one of them out. I have never worked out a good reason for why that is in twice
Rik Smithies (Mar 10 2021 at 11:40):
also ObservationDefinition <!--xs:simpleType name="PublicationStatus-list">
Rik Smithies (Mar 10 2021 at 11:42):
those are normally the only two I think. But that didnt fix the R4B set. it has other issues: Circular definitions detected for complex type ':Element'. This means that ':Element' is contained in its own type hierarchy, which is an error.
And The content type of a derived type and that of its base must both be mixed or both be element-only. Type 'BackboneElement' is element only, but its base type is not.
Khalid Shahin (Mar 10 2021 at 13:52):
Even though I made a clone of the R4B branch, it still wouldn't publish. But just now I made one change to: valueset-action-participant-role.xml
Which was changing:
to
And now it builds.
Khalid Shahin (Mar 10 2021 at 14:11):
Now that I got the R4B build to work locally do I still have time to submit my changes? (I won't be committing my changes to valueset-action-participant-role.xml)
Rob Hausam (Mar 10 2021 at 14:22):
I have some minor text-only updates to Observation and DiagnosticReport to make to get that in alignment with the structural updates to subjectthat have already been applied, if that's still possible.
Rob Hausam (Mar 10 2021 at 15:54):
Lloyd suggested doing it as a ballot comment rather than updating it now.
Lloyd McKenzie (Mar 10 2021 at 16:29):
My final clean-up of the ballot-intro page and the home page is now done
Rik Smithies (Mar 10 2021 at 17:28):
I've made my additions to the datatypes page. I've added new types to the diagrams.
https://build.fhir.org/branches/R4B/datatypes.html
Even before these additions, the diagrams are lacking some inheritance arrows that are there in the main build.
compare to http://build.fhir.org/datatypes.html
Khalid Shahin (Mar 10 2021 at 17:40):
What's the time deadline for R4B changes tonight? Also, can I make some minor changes to the Statistic datatype structure for the R4B build?
Lloyd McKenzie (Mar 10 2021 at 18:16):
The expectation is that all R4B commits are already made. FMG is going to be reviewing the ballot to go forward at 4 Eastern. What changes are you wanting to make @Khalid Shahin?
Khalid Shahin (Mar 10 2021 at 19:10):
Lloyd McKenzie said:
The expectation is that all R4B commits are already made. FMG is going to be reviewing the ballot to go forward at 4 Eastern. What changes are you wanting to make Khalid Shahin?
For Statistic, adding two elements just approved today: Statistic.category and Statistic.numberAffected
The other changes to the build is adding an example for Evidence and EvidenceVariable, and adding three or four meta-data elements that would normally be inherited from the MetaDataResource in R5 (name, experimental, copyright)
Lloyd McKenzie (Mar 10 2021 at 20:47):
@Grahame Grieve has the final say. You can submit your pull request, but don't merge it. If the FMG says 'no', you can make the changes as part of R4B ballot reconciliation
Grahame Grieve (Mar 10 2021 at 20:49):
FMG will decide, but I'll be voting no.
Khalid Shahin (Mar 10 2021 at 20:55):
Lloyd McKenzie said:
Grahame Grieve has the final say. You can submit your pull request, but don't merge it. If the FMG says 'no', you can make the changes as part of R4B ballot reconciliation
Thanks. I created the pull request and it's currently doing the checks: https://github.com/HL7/fhir/pull/1161
Grahame Grieve (Mar 10 2021 at 20:57):
This is my R4B to do list:
Schema issues
Arrows in diagrams
Polish front page
I have to build and publish after FMG finishes. @Mark Iantorno is working on getting the ci-build to work, but it builds for me
Mark Iantorno (Mar 10 2021 at 20:57):
@Khalid Shahin do you need these changes in for the meeting today
Grahame Grieve (Mar 10 2021 at 20:58):
no. just focus on getting my stuff to build on the ci-build.
Grahame Grieve (Mar 10 2021 at 20:59):
FMG will decide about Khalid's PR
Mark Iantorno (Mar 10 2021 at 20:59):
@Khalid Shahin I am cancelling your build
Lloyd McKenzie (Mar 10 2021 at 22:52):
Decision from FMG was to wait on this and deal with it post-ballot
Grahame Grieve (Mar 11 2021 at 19:08):
ok, the R4B ballot is posted here: http://hl7.org/fhir/2021Mar
Gino Canessa (Mar 22 2021 at 21:18):
Is the package/NPM for the ballot available anywhere? I cannot find it at the expected URL of http://hl7.org/fhir/2021Mar/hl7.fhir.r4.core.tgz . @Grahame Grieve
Gino Canessa (Mar 23 2021 at 14:53):
For anyone looking in the future, someone pointed me to the current Publication Directory, which has different URLs from the ballot Downloads page.
Last updated: Apr 12 2022 at 19:14 UTC
