Repository navigation
Tail Recursive Package Resolution - #2066
Conversation
6314fc8 to
65b29e7
Compare
2625e15 to
0560eab
Compare
0560eab to
abaebbf
Compare
|
The verbose output for the failing test is I spent all day yesterday working on this, but I have no idea what the issue is 😞 |
|
I assume one of the tree cuts is gone.
Am 04.12.2016 6:41 nachm. schrieb "Jared Hester" <notifications@github.com>:
… The verbose output for the failing test is
Test Name: #1579 update allows unpinned
Test Outcome: Failed
Result Message:
System.Exception : Paket failed with:
There was a version conflict during package resolution.
Resolved packages:
- MvvmLight 4.2.32.7
- Paket.Test.A 1.0.0-prerelease
Could not resolve package Mvvmlight >= 4.0 < 5.0:
- Paket.Test.A 1.0.0-prerelease requested package Mvvmlight: 4.2.30
Please try to relax some conditions.
StackTrace:
at ***@***.***(String message) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\PackageResolver.fs:line 212
at ***@***.***(GroupName groupName, DependenciesGroup dependenciesGroup) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\UpdateProcess.fs:line 149
at Microsoft.FSharp.Collections.MapTreeModule.mapiOpt[a,b,c](FSharpFunc`3 f, MapTree`2 m)
at Microsoft.FSharp.Collections.FSharpMap`2.Map[b](FSharpFunc`2 f)
at Paket.UpdateProcess.selectiveUpdate(Boolean force, FSharpFunc`2 getSha1, FSharpFunc`2 getSortedVersionsF, FSharpFunc`2 getPackageDetailsF, LockFile lockFile, DependenciesFile dependenciesFile, UpdateMode updateMode, SemVerUpdateMode semVerUpdateMode) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\UpdateProcess.fs:line 139
at Paket.UpdateProcess.SelectiveUpdate(DependenciesFile dependenciesFile, UpdateMode updateMode, SemVerUpdateMode semVerUpdateMode, Boolean force) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\UpdateProcess.fs:line 200
at Paket.UpdateProcess.SmartInstall(DependenciesFile dependenciesFile, UpdateMode updateMode, UpdaterOptions options) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\UpdateProcess.fs:line 215
at ***@***.***(Unit unitVar0) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\PublicAPI.fs:line 206
at Paket.Utils.RunInLockedAccessMode[a](String rootFolder, FSharpFunc`2 action) in C:\Users\jared\Github\Forks\Paket\src\Paket.Core\Utils.fs:line 733
at ***@***.***(ParseResults`1 results) in C:\Users\jared\Github\Forks\Paket\src\Paket\Program.fs:line 432
at Paket.Program.processWithValidation[T](Boolean silent, FSharpFunc`2 validateF, FSharpFunc`2 commandF, ParseResults`1 result) in C:\Users\jared\Github\Forks\Paket\src\Paket\Program.fs:line 36
at ***@***.***(ParseResults`1 results, Boolean silent, Boolean fromBootstrapper, Unit unitVar) in C:\Users\jared\Github\Forks\Paket\src\Paket\Program.fs:line 432
at Paket.Program.main() in C:\Users\jared\Github\Forks\Paket\src\Paket\Program.fs:line 414
Result StandardOutput:
C:\Users\jared\Github\Forks\Paket\integrationtests\scenarios\i001579-unlisted\temp already exists.
C:\Users\jared\Github\Forks\Paket\bin\paket.exe pack templatefile paket.A.template version 1.0.0-prerelease output bin
C:\Users\jared\Github\Forks\Paket\bin\paket.exe update -v
Paket version 4.0.0-alpha030
found: C:\Users\jared\Github\Forks\Paket\integrationtests\scenarios\i001579-unlisted\temp\paket.dependencies
Parsing C:\Users\jared\Github\Forks\Paket\integrationtests\scenarios\i001579-unlisted\temp\paket.dependencies
Resolving packages for group Main:
0 packages in resolution.
2 requirements left
- Mvvmlight, >= 4.0 < 5.0
- Paket.Test.A,
Trying to resolve Mvvmlight >= 4.0 < 5.0 (from C:\Users\jared\Github\Forks\Paket\integrationtests\scenarios\i001579-unlisted\temp\paket.dependencies)
- fetching versions for Mvvmlight
- Mvvmlight 4.2.32.7
1 packages in resolution.
- MvvmLight, 4.2.32.7
2 requirements left
- Paket.Test.A,
- MvvmLightLibs, >= 4.4.32.7
Trying to resolve Paket.Test.A (from C:\Users\jared\Github\Forks\Paket\integrationtests\scenarios\i001579-unlisted\temp\paket.dependencies)
- fetching versions for Paket.Test.A
- Paket.Test.A 1.0.0-prerelease
2 packages in resolution.
- MvvmLight, 4.2.32.7
- Paket.Test.A, 1.0.0-prerelease
2 requirements left
- Mvvmlight, 4.2.30
- MvvmLightLibs, >= 4.4.32.7
Trying to resolve Mvvmlight 4.2.30 (from Paket.Test.A 1.0.0-prerelease)
Could not resolve package Mvvmlight >= 4.0 < 5.0:
- Paket.Test.A 1.0.0-prerelease requested package Mvvmlight: 4.2.30
==> Trying different resolution.
- Mvvmlight 4.2.30.0
unlisted
- Mvvmlight 4.1.27.1
unlisted
- Mvvmlight 4.1.27.0
unlisted
- Mvvmlight 4.1.26.1
unlisted
- Mvvmlight 4.1.26.0
unlisted
- Mvvmlight 4.1.23.0
unlisted
I spent all day yesterday working on this, but I have no idea what the
issue is 😞
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2066 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AADgNMOnytsKb3vwMg6hU2WZph3ZGUoKks5rEvtKgaJpZM4LDQYw>
.
|
eb4c1f2 to
41a2d2f
Compare
2b784cb to
b14dcc3
Compare
b14dcc3 to
804165c
Compare
|
you disabled magic mode? |
|
I can't remember where I first saw it, but magic mode was causing travis builds to fail so I disabled it here. |
|
Please bring back magic mode. Pack did never work. I think @en is working
on it
Am 12.12.2016 8:57 nachm. schrieb "Jared Hester" <notifications@github.com>:
… I can't remember where I first saw it, but magic mode was causing travis
builds to fail so I disabled it here.
I have no idea why dotnet pack is failing now
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2066 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AADgNEnzgEaAuLb7BrYhvVgL81tn2pJjks5rHac4gaJpZM4LDQYw>
.
|
ea07d8f to
9292f52
Compare
9292f52 to
b5b1491
Compare
|
@forki it's back to appveyor passing |
|
I'm still a bit concerned about how the packages are resolving, even though it works properly
It's going through a lot more packages than it should be. Even after I got the unlisted package traversal stage working it's still going through a larger package list than it should. I think something is wrong with the list of compatible packages being provided for a specific version constraint. |
more than before? |
|
I think so, but it's hard to tell since the tests aren't giving full output when they ran on appveyor in previous PRs. This is another reason #2070 would be useful. |


I checked the IL to make sure that it's tail recursive
I tested it against the repos that were causing stackoverflows in #2030, it successfully resolved for both
Although the output for
show that something is still clearly wrong