Problem Matchers are a way to scan the output of actions for a specified regex pattern and surface that information prominently in the UI. Both GitHub Annotations and log file decorations are created when a match is detected.
Currently, GitHub Actions limit the annotation count in a workflow run.
- 10 warning annotations and 10 error annotations per step
- 50 annotations per job (sum of annotations from all the steps)
- 50 annotations per run (separate from the job annotations, these annotations aren’t created by users)
If your workflow may exceed these annotation counts, consider filtering of the log messages which the Problem Matcher is exposed to (e.g. by PR touched files, lines, or other).
Let's consider the ESLint compact output:
badFile.js: line 50, col 11, Error - 'myVar' is defined but never used. (no-unused-vars)
We can define a problem matcher in json that detects input in that format:
{
"problemMatcher": [
{
"owner": "eslint-compact",
"pattern": [
{
"regexp": "^(.+):\\sline\\s(\\d+),\\scol\\s(\\d+),\\s(Error|Warning|Info)\\s-\\s(.+)\\s\\((.+)\\)$",
"file": 1,
"line": 2,
"column": 3,
"severity": 4,
"message": 5,
"code": 6
}
]
}
]
}
The following fields are available for problem matchers:
{
owner: an ID field that can be used to remove or replace the problem matcher. **required**
severity: indicates the default severity, either 'warning' or 'error' case-insensitive. Defaults to 'error'
pattern: [
{
regexp: the regex pattern that provides the groups to match against **required**
file: a group number containing the file name
fromPath: a group number containing a filepath used to root the file (e.g. a project file)
line: a group number containing the line number
column: a group number containing the column information
severity: a group number containing either 'warning' or 'error' case-insensitive. Defaults to `error`
code: a group number containing the error code
message: a group number containing the error message. **required** at least one pattern must set the message
loop: whether to loop until a match is not found, only valid on the last pattern of a multipattern matcher
}
]
}
Consider the following output:
test.js
1:0 error Missing "use strict" statement strict
5:10 error 'addOne' is defined but never used no-unused-vars
✖ 2 problems (2 errors, 0 warnings)
The file name is printed once, yet multiple error lines are printed. The loop
keyword provides a way to discover multiple errors in outputs.
The eslint-stylish problem matcher defined below catches that output, and creates two annotations from it.
{
"problemMatcher": [
{
"owner": "eslint-stylish",
"pattern": [
{
// Matches the 1st line in the output
"regexp": "^([^\\s].*)$",
"file": 1
},
{
// Matches the 2nd and 3rd line in the output
"regexp": "^\\s+(\\d+):(\\d+)\\s+(error|warning|info)\\s+(.*)\\s\\s+(.*)$",
// File is carried through from above, so we define the rest of the groups
"line": 1,
"column": 2,
"severity": 3,
"message": 4,
"code": 5,
"loop": true
}
]
}
]
}
The first pattern matches the test.js
line and records the file information. This line is not decorated in the UI.
The second pattern loops through the remaining lines with loop: true
until it fails to find a match, and surfaces these lines prominently in the UI.
Note that the pattern matches must be on consecutive lines. The following would not result in any match findings.
test.js
extraneous log line of no interest
1:0 error Missing "use strict" statement strict
5:10 error 'addOne' is defined but never used no-unused-vars
✖ 2 problems (2 errors, 0 warnings)
Problem Matchers are enabled and removed via the toolkit commands.
Registering two problem-matchers with the same owner will result in only the problem matcher registered last running.
Some of the starter actions are already using problem matchers, for example:
Use ECMAScript regular expression syntax when testing patterns.
Enable debug logging to determine why the file is getting dropped.
This usually happens when the file does not exist or is not under the workflow repo.