Stream: committers/git-help
Topic: vscache externals missing
John Moehrke (Nov 01 2018 at 15:38):
I get this build failure and am at a loss
[java] Error: C:\Users\john.moehrke\Git\fhir\vscache\externals\1253998990.xml (The system cannot find the path specified)
[java] java.io.FileNotFoundException: C:\Users\john.moehrke\Git\fhir\vscache\externals\1253998990.xml (The system cannot find the path specified)
Josh Mandel (Nov 01 2018 at 15:42):
Is this a full build or partial? I don't know what this file is, so I have no intuition about how to debug it -- but I'm remembering that you recently deleted your local cache directory and restored from git, and I could imagine that maybe there are some files that only get generated by a full build...
John Moehrke (Nov 01 2018 at 15:51):
I forced a full build, it still happens... Your memory is good, I did delete the vscache due to the conflict problems, with a pull that gave me 'some' values in vscache
Josh Mandel (Nov 01 2018 at 15:51):
Which as far as I'm aware should work perfectly
John Moehrke (Nov 01 2018 at 15:52):
C:\Users\john.moehrke\Git\fhir\vscache>dir /b
all-systems.cache
appointment-cancellation-reason.cache
lang.cache
loinc.cache
null.cache
program.cache
task-performer-type.cache
ucum.cache
urn_oid_2.16.840.1.113883.3.26.1.1.cache
John Moehrke (Nov 01 2018 at 15:52):
no directories within vscache were created from the pull
Josh Mandel (Nov 01 2018 at 15:53):
But what is this externals thing all about?
John Moehrke (Nov 01 2018 at 15:54):
I don't know.. Ill try a clean, and build
John Moehrke (Nov 01 2018 at 16:01):
that didn't make it better
John Moehrke (Nov 01 2018 at 16:02):
however I see this $group thing again
[java] ERROR: Unable to process warnings: An empty sequence is not allowed as the value of variable $groupsError on line 14
[java]
[java] XTTE0570: An empty sequence is not allowed as the value of variable $groups
[java] in variable groups
[java] ; SystemID: ; Line#: 14; Column#: -1
[java] net.sf.saxon.trans.XPathException: An empty sequence is not allowed as the value of variable $groups
[java] at net.sf.saxon.expr.Expression.typeError(Expression.java:1163)
[java] at net.sf.saxon.expr.CardinalityChecker.iterate(CardinalityChecker.java:226)
[java] at net.sf.saxon.expr.ItemChecker.iterate(ItemChecker.java:170)
[java] at net.sf.saxon.expr.parser.ExpressionTool.evaluate(ExpressionTool.java:340)
[java] at net.sf.saxon.expr.instruct.GlobalVariable.getSelectValue(GlobalVariable.java:563)
[java] at net.sf.saxon.expr.instruct.GlobalVariable.actuallyEvaluate(GlobalVariable.java:615)
[java] at net.sf.saxon.expr.instruct.GlobalVariable.evaluateVariable(GlobalVariable.java:587)
[java] at net.sf.saxon.expr.VariableReference.evaluateVariable(VariableReference.java:505)
[java] at net.sf.saxon.expr.VariableReference.iterate(VariableReference.java:453)
[java] at net.sf.saxon.expr.SlashExpression.iterate(SlashExpression.java:813)
[java] at net.sf.saxon.expr.sort.DocumentSorter.iterate(DocumentSorter.java:145)
[java] at net.sf.saxon.expr.Atomizer.iterate(Atomizer.java:294)
[java] at net.sf.saxon.functions.DistinctValues.iterate(DistinctValues.java:76)
[java] at net.sf.saxon.expr.sort.SortExpression.iterate(SortExpression.java:537)
[java] at net.sf.saxon.expr.instruct.ForEach.iterate(ForEach.java:498)
[java] at net.sf.saxon.value.MemoClosure.iterate(MemoClosure.java:112)
[java] This was a Full Build 1.202 31sec 416MB
[java] at net.sf.saxon.expr.VariableReference.iterate(VariableReference.java:454)
[java] at net.sf.saxon.expr.instruct.ForEach.iterate(ForEach.java:498)
[java] at net.sf.saxon.expr.ItemChecker.iterate(ItemChecker.java:170)
[java] at net.sf.saxon.value.MemoClosure.iterate(MemoClosure.java:112)
[java] at net.sf.saxon.expr.VariableReference.iterate(VariableReference.java:454)
[java] at net.sf.saxon.expr.instruct.CopyOf.iterate(CopyOf.java:798)
[java] at net.sf.saxon.expr.instruct.BlockIterator.next(BlockIterator.java:49)
[java] at net.sf.saxon.expr.sort.SortedIterator.buildArray(SortedIterator.java:223)
[java] at net.sf.saxon.expr.sort.SortedIterator.doSort(SortedIterator.java:255)
[java] at net.sf.saxon.expr.sort.SortedIterator.next(SortedIterator.java:140)
[java] at net.sf.saxon.expr.instruct.ForEach.processLeavingTail(ForEach.java:475)
[java] at net.sf.saxon.expr.instruct.Instruction.process(Instruction.java:138)
[java] at net.sf.saxon.expr.LetExpression.process(LetExpression.java:463)
[java] at net.sf.saxon.expr.instruct.ElementCreator.processLeavingTail(ElementCreator.java:431)
[java] at net.sf.saxon.expr.instruct.ElementCreator.processLeavingTail(ElementCreator.java:373)
[java] at net.sf.saxon.expr.instruct.Template.applyLeavingTail(Template.java:239)
[java] at net.sf.saxon.trans.Mode.applyTemplates(Mode.java:1057)
[java] at net.sf.saxon.Controller.transformDocument(Controller.java:2080)
[java] at net.sf.saxon.Controller.transform(Controller.java:1903)
[java] at org.hl7.fhir.utilities.Utilities.saxonTransform(Utilities.java:373)
[java] at org.hl7.fhir.tools.publisher.Publisher.processWarnings(Publisher.java:1903)
[java] at org.hl7.fhir.tools.publisher.Publisher.execute(Publisher.java:667)
[java] at org.hl7.fhir.tools.publisher.Publisher.main(Publisher.java:492)
[java] FHIR build failure @ Thu, Nov 1, 2018 11:00-0500 1.912 33sec 416MBjava.io.FileNotFoundException: C:\Users\john.moehrke\Git\fhir\vscache\externals\1253998990.xml (The system cannot find the path specified)
[java] at java.io.FileOutputStream.open0(Native Method)
[java] at java.io.FileOutputStream.open(FileOutputStream.java:270)
[java] Error: C:\Users\john.moehrke\Git\fhir\vscache\externals\1253998990.xml (The system cannot find the path specified)
[java]
[java] at java.io.FileOutputStream.<init>(FileOutputStream.java:213)
[java] at java.io.FileOutputStream.<init>(FileOutputStream.java:162)
[java] at org.hl7.fhir.utilities.TextFile.bytesToFile(TextFile.java:163)
[java] at org.hl7.fhir.definitions.parsers.SourceParser.loadExternal(SourceParser.java:341)
[java] at org.hl7.fhir.definitions.parsers.SourceParser.loadExternals(SourceParser.java:329)
[java] at org.hl7.fhir.definitions.parsers.SourceParser.parse(SourceParser.java:203)
[java] at org.hl7.fhir.tools.publisher.Publisher.execute(Publisher.java:601)
[java] at org.hl7.fhir.tools.publisher.Publisher.main(Publisher.java:492)
Press any key to continue . . .
John Moehrke (Nov 01 2018 at 16:04):
This is what happened to me before... where the solution was to delete everything and start completely over at step zero
John Moehrke (Nov 01 2018 at 16:06):
frustrating is that turtle is still showing my directory as not green, yet a pull doesn't seem find anything to pull
John Moehrke (Nov 01 2018 at 16:08):
I just did two pull in a row with different results
Lloyd McKenzie (Nov 01 2018 at 16:08):
Are you having a permissions issue? I know that sometimes when I delete stuff, it doesn't truly get deleted. The file is "gone" but Windows thinks it's still locked and prevents access to it, writing to it, etc. Sometimes I need to kill and restart Windows Explorer to address the issue.
John Moehrke (Nov 01 2018 at 16:08):
I see the same from a commandline
Lloyd McKenzie (Nov 01 2018 at 16:08):
Nothing in the build should depend on a cache file existing - the cache files should be re-created by the build
Lloyd McKenzie (Nov 01 2018 at 16:09):
You see the same what? You see the file is there? Not there?
Josh Mandel (Nov 01 2018 at 16:09):
If they are recreated by the build, then Why do we check them into Source control?
John Moehrke (Nov 01 2018 at 16:09):
for example the above dir from the command line of the vscache directory
John Moehrke (Nov 01 2018 at 16:18):
I seem to be heading backward. How do I tell git to pull to my local master EVERYTHING?
John Moehrke (Nov 01 2018 at 16:19):
GIT seems to be preventing me from getting a complete copy of master
John Moehrke (Nov 01 2018 at 16:19):
yet it tells me 'your branch is up to date with 'original/master'."
John Moehrke (Nov 01 2018 at 16:20):
it can't be... there is lots of stuff missing
Lloyd McKenzie (Nov 01 2018 at 16:20):
Are you currently switched into master or are you sitting in a branch?
John Moehrke (Nov 01 2018 at 16:21):
I think I am at master
Lloyd McKenzie (Nov 01 2018 at 16:22):
If you right-click , does it say "Git commit - "master""?
Lloyd McKenzie (Nov 01 2018 at 16:22):
If so, you are
John Moehrke (Nov 01 2018 at 16:22):
yes
Lloyd McKenzie (Nov 01 2018 at 16:23):
If you go to commit, does it show you deleted files you can revert to undelete?
John Moehrke (Nov 01 2018 at 16:23):
both commandline git, and turtle say I am up to date with master
John Moehrke (Nov 01 2018 at 16:23):
yet, I have almost nothing in my filesystem
John Moehrke (Nov 01 2018 at 16:23):
hmm, let me check
John Moehrke (Nov 01 2018 at 16:24):
yes a bunch of files 'missing'
Lloyd McKenzie (Nov 01 2018 at 16:25):
Try reverting them
John Moehrke (Nov 01 2018 at 16:25):
do what?
Lloyd McKenzie (Nov 01 2018 at 16:25):
Select all the missing files
Lloyd McKenzie (Nov 01 2018 at 16:25):
Then right-click and say "revert"
Lloyd McKenzie (Nov 01 2018 at 16:26):
I suspect you did a git delete instead of a local delete when you removed the files - so git thinks you have an uncommitted change to permanently remove those files.
John Moehrke (Nov 01 2018 at 16:27):
okay... 9000+ files... tortoise is busy...
John Moehrke (Nov 01 2018 at 16:28):
so the advice Josh gave to delete the vscache directory to force a refresh of the content was not good advice...
Lloyd McKenzie (Nov 01 2018 at 16:29):
It was good advice, but perhaps not sufficiently clear. He didn't want you to do a Git delete, just a local delete.
Lloyd McKenzie (Nov 01 2018 at 16:29):
I.e. a Windows delete
John Moehrke (Nov 01 2018 at 16:30):
that is all I did... from the filesystem I deleted the directory
John Moehrke (Nov 01 2018 at 16:30):
tortoise must be watching a bit too closely
Lloyd McKenzie (Nov 01 2018 at 16:30):
I don't know why Git thinks the files are supposed to be deleted then. I've never noticed that behavior in my space...
John Moehrke (Nov 01 2018 at 16:31):
my corporate laptop does have a screwy filesystem addon, I wonder if that is fighting with tortoiseGIT
John Moehrke (Nov 01 2018 at 16:32):
tortoiseGIT is still working on that revert...
John Moehrke (Nov 01 2018 at 16:50):
wow, it finished with that revert. 30 minutes... and I have an ssd
John Moehrke (Nov 01 2018 at 16:51):
doing clean build.... will update you in 80 minutes...
Josh Mandel (Nov 01 2018 at 17:07):
I'm confused -- a revert for me usually takes on the order of a minute :/
John Moehrke (Nov 01 2018 at 17:13):
per file... sure.. .but I had 9000+ files to revert. still not reasonable, but not much else I can do
Lloyd McKenzie (Nov 01 2018 at 17:39):
We might have been able to do the revert at the root folder level, which would have been faster - though that would also have undone any changes you wanted to keep.
John Moehrke (Nov 01 2018 at 17:40):
I had already manually copied out the tiny changes I had
Josh Mandel (Nov 01 2018 at 17:40):
(I meant for the entire repository.)
Grahame Grieve (Nov 01 2018 at 17:58):
If they are recreated by the build, then Why do we check them into Source control?
Because it takes several hours to build them
Josh Mandel (Nov 01 2018 at 17:59):
So running the build with these files missing would be a 3+ hour mistake?
Josh Mandel (Nov 01 2018 at 18:00):
Can we keep a cache in a cloud bucket outside of git and make the files immutable?
Grahame Grieve (Nov 01 2018 at 18:00):
they're not immutable
Josh Mandel (Nov 01 2018 at 18:00):
Let's discuss this. Any given version can be immutable...
Grahame Grieve (Nov 01 2018 at 18:01):
discuss what?
Josh Mandel (Nov 01 2018 at 18:01):
And the build could reference the version it required, fetching from elsewhere
Josh Mandel (Nov 01 2018 at 18:01):
Discuss how to avoid the confusion that's costing John and others.
Grahame Grieve (Nov 01 2018 at 18:01):
"fetching it from elsewhere" is exactly what takes so long
Josh Mandel (Nov 01 2018 at 18:01):
Brainstorm about the problem and possible solutions
Josh Mandel (Nov 01 2018 at 18:01):
The files aren't large, so fetching can't be slow
Grahame Grieve (Nov 01 2018 at 18:04):
They are supposed to be stable; they are supposed to only change when something changes about how the code system is used. Trouble is for me debugging the reason they are not stable take a very long time - build after build after build, and git makes it impossible for me to do this in the background
Rob Hausam (Nov 01 2018 at 18:06):
Agree with Josh about brainstorming regarding a solution. This is a recurring source of at least annoyance and sometimes a significant problem, so it would be good if we could come up with a solution that would make it easier (and hopefully seamless) to manage.
Lloyd McKenzie (Nov 01 2018 at 18:10):
With SVN, they'd just merge and all would be happy. The question is why in Git they're showing up as a conflict instead of merging cleanly.
Grahame Grieve (Nov 01 2018 at 18:11):
there's a couple of places where git won't merge.
Lloyd McKenzie (Nov 01 2018 at 18:14):
On occasion you'd hit a conflict with SVN - generally on SNOMED. The solution then was to revert your local copy and then commit. The fact the conflicts show up when trying to merge the pull request and that reverting your local copy doesn't seem to be an option also causes grief
Grahame Grieve (Nov 01 2018 at 18:19):
that's a piece of work flow in git that seems exceptionally obnoxious - here: these files have a problem; you must do one or the other. But I won't help you by letting you chose and do it; instead, you have to retype them and do it your self.
Josh Mandel (Nov 01 2018 at 19:01):
I'm losing internet for a bit
Grahame Grieve (Nov 01 2018 at 19:06):
on investigation, it seems that most of the issues are unicode related. I suspect git is altering the unicode encoding
Grahame Grieve (Nov 01 2018 at 21:44):
no it's at least partly my fault; I read the files as UTF-8 but write them using default encoding.
Rob Hausam (Nov 01 2018 at 22:10):
Ah, that makes sense with the behavior that I've been seeing. Hopefully after fixing that it will work a lot smoother.
Last updated: Apr 12 2022 at 19:14 UTC
