I'd like to set the merge_when_build_succeeds attribute of a merge request with the Gitlab API. The docs says, that when the optional merged_when_build_succeeds parameter is true, the MR'll be accepted only after the build succeeds.
How should I provide this merged_when_build_succeeds parameter to the API? I tried the following curl commands without any success:
# In the request's body
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" -d merged_when_build_succeeds=true "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge"
{"message":"405 Method Not Allowed"}
# In the URL
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge?merged_when_build_succeeds=true"
{"message":"405 Method Not Allowed"}
# I tried this also, maybe the __merged__ is a mistake in the docs
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" -d merge_when_build_succeeds=true "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge"
{"message":"405 Method Not Allowed"}
# In the URL
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge?merge_when_build_succeeds=true"
{"message":"405 Method Not Allowed"}
# Some other CURL parameters
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" -H "Content-Type: multipart/form-data;" -F "merge_when_build_succeeds=true" "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge"
{"message":"405 Method Not Allowed"}
# Some other CURL parameters
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" -H "Content-Type: multipart/form-data;" -F "merged_when_build_succeeds=true" "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge"
{"message":"405 Method Not Allowed"}
The new feature, 'Only allow merge requests to be merged if the build succeeds' is turned on. When I turn off this feature any call of the previous calls merge the MR before the build finished, see: https://snag.gy/06nOEm.jpg
The private token, and the URL is right. I can query the MR's info:
vilmosnagy#vnagy-dell:~$ curl -X GET --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" "http://localhost:8080/api/v3/projects/1/merge_requests/9"
{"id":9,"iid":9,"project_id":1,"title":"enters","description":"","state":"opened","created_at":"2016-06-29T15:36:15.235Z","updated_at":"2016-06-29T15:36:15.235Z","target_branch":"master","source_branch":"features/long_build_05","upvotes":0,"downvotes":0,"author":{"name":"Vilmos Nagy","username":"vilmos.nagy","id":2,"state":"active","avatar_url":"http://www.gravatar.com/avatar/4f94d9571ec83f42a85651291296f503?s=80\u0026d=identicon","web_url":"http://172.21.0.3/u/vilmos.nagy"},"assignee":null,"source_project_id":2,"target_project_id":1,"labels":[],"work_in_progress":false,"milestone":null,"merge_when_build_succeeds":false,"merge_status":"can_be_merged","subscribed":false,"user_notes_count":0}
And I can accept the MR after the build succeeds:
vilmosnagy#vnagy-dell:~$ curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge"
{"id":9,"iid":9,"project_id":1,"title":"enters","description":"","state":"merged","created_at":"2016-06-29T15:36:15.235Z","updated_at":"2016-06-29T16:13:41.242Z","target_branch":"master","source_branch":"features/long_build_05","upvotes":0,"downvotes":0,"author":{"name":"Vilmos Nagy","username":"vilmos.nagy","id":2,"state":"active","avatar_url":"http://www.gravatar.com/avatar/4f94d9571ec83f42a85651291296f503?s=80\u0026d=identicon","web_url":"http://172.21.0.3/u/vilmos.nagy"},"assignee":null,"source_project_id":2,"target_project_id":1,"labels":[],"work_in_progress":false,"milestone":null,"merge_when_build_succeeds":false,"merge_status":"can_be_merged","subscribed":true,"user_notes_count":0}
Could you help me, and give me the correct call of the given API?
Thanks!
Vilmos
It seems the documentation is wrong, the parameter is actually called merge_when_build_succeeds (without the "d" in "merge").
curl -X PUT --header "PRIVATE-TOKEN: zvzK7CNzx9WviV5iChyg" -d merge_when_build_succeeds=true "http://localhost:8080/api/v3/projects/1/merge_requests/9/merge" should work ("405 Method Not Allowed" was probably because of something else, like the MR was already merged or had conflicts).
I created an issue you can follow if you want to know when the documentation is fixed: https://gitlab.com/gitlab-org/gitlab-ce/issues/19448
Related
i am trying to write a ci.yml to automatically create a merge request and merge it and wait until pipeline succeed. In order to do that, i need the pipeline Id of the merge request. but the only pipeline Id that i recieve from the response of merge call is .head_pipeline.id which is the last pipeline id of the project not the one after merging.
i need the actual pipeline id to poll the status. if the status is "success", job will be ended.
i have tried :
`
PIPELINE_ID=`curl --silent -X PUT "${GITLAB_BASE_URL}/${SERVICE_PROJECT_ID}/merge_requests/${MERGE_REQUEST_IID}/merge" --header "Private-Token: ${PRIVATE_TOKEN}" --header "Content-Type: application/json" | jq .head_pipeline.id`
`
here is the solution, that i found. After merge, we can find "merge_commit_sha" as an attribute in rspBody. then we should get all pipelines and find the one that match with that attribute.
MERGE_COMMIT_SHA=`curl --silent -X GET "${GITLAB_BASE_URL}/${SERVICE_PROJECT_ID}/merge_requests/${MERGE_REQUEST_IID}" --header "Private-Token: ${PRIVATE_TOKEN}" --header "Content-Type: application/json" | jq -r .merge_commit_sha`
PIPELINE_ID=$(curl --silent -X GET "${GITLAB_BASE_URL}/${SERVICE_PROJECT_ID}/pipelines" --header "Private-Token: ${PRIVATE_TOKEN}" | jq '.[] | select(.ref=='\"$TARGET_BRANCH\"') | select (.sha=='\"$MERGE_COMMIT_SHA\"') | .id')
I am referring to the thread: Passing a large text to bash curl url
I have a shell script that looks like:
#!bin/bash
curl --location --request POST 'example.com/abcxyz' \
--header 'Content-Type: application/json' \
--header 'Cookie: my_secret_key' \
--data-raw '{_my_very_long_json_content_here}'
Approach 1: I put everything after the curl, i.e. from --location ... to the end to a test.txt file then
tempfile = test.txt
curl "#$tempfile"
and I get the error curl: (6) Could not resolve host: test.txt
Approach 2: I put everything in a test.sh file:
#!bin/bash
curl --location --request POST 'example.com/abcxyz' --header 'Content-Type: application/json' --header 'Cookie: my_secret_key' --data-raw '#-' <<< 'my_very_long_json'
and I have an error {"error":"invalid character '#' looking for beginning of value","code":3}(base)
What should I do to pass a very long JSON content to curl POST?
Note that, if the JSON content is not too long, I can put everything in a SH file then run
bash test.sh
and it is okay.
--data-raw specifically disables # interpretation. Use --data #- or --data-binary #- to read from stdin.
curl --location --request POST 'example.com/abcxyz' \
--header 'Content-Type: application/json' \
--header 'Cookie: my_secret_key' \
--data #- <<< '{_my_very_long_json_content_here}'
You might find a heredoc more ergonomic than a long string:
curl --location --request POST 'example.com/abcxyz' \
--header 'Content-Type: application/json' \
--header 'Cookie: my_secret_key' \
--data #- <<'DATA'
{
'json': 'here'
}
DATA
Could someone tell me how to pass the output of one CURL GET to another CURL POST? I mean something like this:
curl --header "Content-Type: application/json" --request POST --data "{\"specification\": curl --header "Content-Type: application/json" --request GET "https://user:pass#anotherhost"}" "https://user:pass#localhost"
Use jq to craft JSON in the shell.
response="$(curl --header "Content-Type: application/json" --request GET "https://user:pass#anotherhost")"
data="$(jq "{specification: .}" <<< "$response")"
curl --header "Content-Type: application/json" --request POST --data "$data" "https://user:pass#localhost"
Keep in mind, that passwords in command line arguments are public to the host.
It looks like you need to use command substitution. The result would be something like this.
curl --header "Content-Type: application/json" --request POST --data "{\"specification\": $(curl --header "Content-Type: application/json" --request GET "https://user:pass#anotherhost")}" "https://user:pass#localhost"
$() is used for command substitution and it invokes a subshell. The command in the parentheses (a.k.a. round brackets) of $() is executed in a subshell and the output is then placed in the original command.
So the curl GET will be executed in a subshell and the result will be passed to the curl POST
You can read more about it at the link below:
https://www.gnu.org/software/bash/manual/html_node/Command-Substitution.html
We saw https://developer.shopware.com/docs/guides/integrations-api but couldn't find a detailed documentation of the APIs.
EDIT: I found https://my-shop-host.example/api/v2/_info/swagger.html which describes all available APIs and import-export-file endpoints, but it is not clearly described how to use them (and which ones to use).
I believe we need to call (extracted from the admin panel work flow)
/api/v2/_action/import-export/prepare
and then
/api/v2/_action/import-export/process
to trigger an import.
But how are the files uploaded?
Is there an easier way, for example in one call?
I start with the authentication
curl 'http://example.com/api/oauth/token'\
-H 'Accept: pplication/vnd.api+json'\
-H 'Content-Type: application/json' \
--data '{
"grant_type": "client_credentials",
"client_id": "<my client ID from the integration>",
"client_secret": "<my secret>"
}' | jq .access_token
This returns and auth token which is valid for 10 minutes.
I store it to the variable $B
Next I upload the file and set the expiry date to 10 days in the future:
curl 'http://example.com/api/v2/_action/import-export/prepare' \
-H 'Accept: application/vnd.api+json' \
-H "Authorization: Bearer $B"
--form "file=#\"products.csv\"" --form "profileId=d4ec3999a33242a690ca5c213a7145cd" --form "expireDate=+10 days" | jq .log.id
This returns a log.id field
This field must be passed to the process call
curl 'http://example.com/api/v2/_action/import-export/process' \
-H 'Accept: application/vnd.api+json' \
-H 'Content-type: application/json' \
-H "Authorization: Bearer $B" --data '{"logId":"1410eaa1ca434744a8c211b60f65a3c5","offset":0}'
EDIT: Each chunk has to be triggered separately by using the last returned offset and pass it again to the process call.
curl -X POST https://api.commerce.coinbase.com/charges/ \
-H "Content-Type: application/json" \
-H "X-CC-Api-Key: YOUR_API_KEY" \
-H "X-CC-Version: 2018-03-22" \
-d "#data.json"
how do i make this api work?
-H attaches a header to your request. See manpage of curl.
The -H is adding a Header to the request so everything between the speech marks after the -H is a header.
Your CURL request has 3 headers.
See https://developer.mozilla.org/en-US/docs/Glossary/Request_header