FFIUnsupportedUntypedLiteral error in Pharo 12 on MacOS & Windows

[labordep]

The CI does not pass on Pharo 12 due to this error:

FFIUnsupportedUntypedLiteral
FFICallout>>unsupportedUntypedLiteral:
FFIStrictResolutionMode>>resolveUndeclaredTypeForArgument:withResolver:
FFICallout>>resolveUntypedArgument:
FFIUndefinedTypeDeclaration>>resolveUsing:forArgument:
FFIConstantArgument>>resolveUsing:
[ :e | e resolveUsing: aResolver ] in FFIFunctionSpec>>resolveUsing: in Block: [ :e | e resolveUsing: aResolver ]
OrderedCollection>>do:
FFIFunctionSpec>>resolveUsing:
TFCalloutMethodBuilder(FFICalloutMethodBuilder)>>generateMethodFromSpec:
TFCalloutMethodBuilder(FFICalloutMethodBuilder)>>generate
TFCalloutMethodBuilder(FFICalloutMethodBuilder)>>build:
TFCalloutAPI(FFICalloutAPI)>>function:library:
LGitError class(Object)>>ffiCall:library:options:fixedArgumentCount:
LGitError class(Object)>>ffiCall:library:options:
LGitError class(Object)>>ffiCall:
LGitError class(LGitExternalStructure class)>>ffiCallSafely:
LGitError class>>giterr_set_str:
LGitError class>>setError:
[ 
		LGitError setError: 'No ssh-agent suitable credentials found.'.
		^ -1 ] in LGitCredentialsSSHAgent class>>createFrom:username:onto: in Block: [ ...
FullBlockClosure(BlockClosure)>>cull:
Context>>evaluateSignal:
Context>>handleSignal:
LGitNoCredentialsProvided(Exception)>>signal
LGitNoCredentialsProvided(Exception)>>signal:
LGitNoCredentialsProvided class(Exception class)>>signal:
IceTipInteractiveCredentialsProvider(IceCredentialsProvider)>>provideSshAgentCredentialsTo:
[  
		credentials := self new.
		provider provideSshAgentCredentialsTo: credentials.	
		credentials createWith: username onto: lgitCredentials.
		^ 0 ] in LGitCredentialsSSHAgent class>>createFrom:username:onto: in Block: [  ...
FullBlockClosure(BlockClosure)>>on:do:
LGitCredentialsSSHAgent class>>createFrom:username:onto:
LGitRemoteCallbacksV120 class(LGitAbstractRemoteCallbacks class)>>putCredentialsType:username:provider:onto:
3. using HTTPS instead SSH (Just use an url in the form HTTPS://etc.git). I will try to clone the HTTPS variant.

Error with status code 1:
686 run_script /Users/runner/.smalltalkCI/helpers.sh
353 pharo::run_script /Users/runner/.smalltalkCI/pharo/run.sh
378 pharo::load_project /Users/runner/.smalltalkCI/pharo/run.sh
426 run_build /Users/runner/.smalltalkCI/pharo/run.sh
547 run /Users/runner/.smalltalkCI/run.sh
[58](https://github.com/OpenSmock/GeoView/actions/runs/7920273218/job/21644175597#step:4:59)3 main /Users/runner/.smalltalkCI/run.sh
589 main /Users/runner/.smalltalkCI/run.sh
===================================================================
config_clean=false
config_colorful=true
config_debug=false
config_first_arg_or_empty=
config_force_cache=false
config_headless=true
config_image=
config_overwrite_cache=false
config_smalltalk=Pharo[64](https://github.com/OpenSmock/GeoView/actions/runs/7920273218/job/21644175597#step:4:65)-12
config_ston=/Users/runner/work/GeoView/GeoView/.smalltalk.ston
config_tracking=true
config_verbose=false
config_vm=
config_vm_dir=/Users/runner/.smalltalkCI/_cache/vms/Pharo64-12

Error: Process completed with exit code 1.

Complete stack here: https://github.com/OpenSmock/OpenSmock/actions/runs/7907783407/job/21644697228

[labordep]

No problem on GitLab (Linux), this is for GitHub.

[fniephaus]

I have no idea what this means, can someone from the Pharo team take a look at this please? @estebanlm?

[labordep]

I suppose this is ok on Windows (I cannot redo the job for the moment).

[estebanlm]

no, is not. there was a change on FFI for p12, where we no longer assume a number will be an int, you need to cast it explicitly. I will commit a fix, but it will be hard to test, heh.

[labordep]

I confirm: there is the problem on Windows.

[estebanlm]

I fixed this, but now I am stuck trying to push it to the image. I will be back to this soon :)

[labordep]

Hi @estebanlm, I think this is fixed now? :) I haven't re-see this problem.

[estebanlm]

Yes there was some work thereOn 24 Jul 2024, at 10:21, Pierre Laborde @.***> wrote: Hi @estebanlm, I think this is fixed now? :) I haven't re-see this problem.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you were mentioned.Message ID: @.***>