Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/quasilyte/go-parsefix
Fixes simple parse errors automatically. Works great in combination with goimports.
https://github.com/quasilyte/go-parsefix
autofix go golang parsing tools utility
Last synced: about 2 months ago
JSON representation
Fixes simple parse errors automatically. Works great in combination with goimports.
- Host: GitHub
- URL: https://github.com/quasilyte/go-parsefix
- Owner: quasilyte
- License: bsd-2-clause
- Created: 2018-05-26T10:03:51.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-12-05T19:09:47.000Z (almost 5 years ago)
- Last Synced: 2024-04-30T07:19:34.697Z (5 months ago)
- Topics: autofix, go, golang, parsing, tools, utility
- Language: Go
- Size: 30.3 KB
- Stars: 82
- Watchers: 6
- Forks: 4
- Open Issues: 10
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# go-parsefix
[](https://goreportcard.com/report/github.com/quasilyte/go-parsefix)
[](https://godoc.org/github.com/quasilyte/go-parsefix/parsefix)
[](https://travis-ci.org/quasilyte/go-parsefix.svg?branch=master)
Fixes simple parse errors automatically. Works great in combination with [goimports](https://godoc.org/golang.org/x/tools/cmd/goimports).
## Installation
```bash
go get -v github.com/quasilyte/go-parsefix/cmd/parsefix
```
Notes:
1. You need `go` command to be installed.
2. Executable is saved at your `$(go env GOPATH)/bin` by default.
Do `parsefix -help` to see usage and flags documentation.
## Motivation
Sometimes you miss a trailing comma.
The other time it's a missing `;` or `}`.
If you're a beginner, you'll probably put `{` on a wrong line several times,
breaking the `gofmt` due to the parsing errors.
Stop interrupting yourself with such nuisances!
Let `parsefix` perform it's magic.
You do `ctrl+s` in your favourite IDE/editor, it tries to do `gofmt` (or `goimports`), which fails due
to parsing errors, then plugin invokes `parsefix`, which could fix all those issues so `gofmt`
can be executed again successfully. In the end, you don't even notice that there were a minor parsing
error at all. It's just re-formatted and cleaned up.
**Note**: in bright future we could fix **more** errors, not less, as parsing errors
could be improved in some cases to avoid too vague descriptions that are not
precise enough to perform action with high confidence.
## What can be fixed
`parsefix` does not do anything too smart. It only follows safe suggestions from
error messages that usually lead to fixed source code.
Note that it fixes *parsing* errors, not semantic or type errors.
Sometimes it performs not quite right actions, for example, it could insert a `,` where `:`
would make more sense, but you will notice that in the *typecheck* phase.
The best part is that *typecheck* could actually run over your previously unparsable code.
Type checker usually gives far more precise and concise error messages.
### Fix misplaced opening brace
```go
func f()
{
}
// =>
func f() {
}
```
### Fix missing range keyword
```go
for x := xs {
}
=>
for x := range xs {
}
```
### Fix missing comma
```go
xs := []string{
"a"
"b"
}
// =>
xs := []string{
"a",
"b",
}
xs := []int{1 2}
// =>
xs := []int{1, 2}
foo(1 2)
// =>
foo(1, 2)
func f(a int b int32) {}
// =>
func f(a int, b int32) {}
```
### Fix missing colon
```go
switch v {
case 1
return a
case 2
return b
}
// =>
switch v {
case 1:
return a
case 2:
return b
}
```
### Fix missing semicolon
```go
x++ y++
// =>
x++; y++
if x := 1 x != y {
}
// =>
if x := 1; x != y {
}
```
### Fix misplaced tokens
```go
func f() {
:=
g()
}
// =>
func f() {
g()
}
```
### Fix illegal characters
```go
func f() {
$ g()
🔥 g()
# g()
â„– g()
}
// =>
func f() {
g()
g()
g()
g()
}
```
## Problems
Some parsing errors drive Go parser mad.
A single typo causes multiple parts of the source file to issue parsing errors.
This could lead to false-positives, unfortunately.
It would also be great to fix things like `var x := y` to `var x = y`, but
there is no way to do so currently, as messages for this kinds of errors are ambiguous and
do not mention `:=` at all.
Maybe `parsefix` will learn to avoid those in future.
We need more user feedback and bug reports.
## Integration
For ease of integration there are two modes:
1. Accept (full) filename, parse file, try to fix errors that are found during parsing.
2. Accept (full) filename + list of parsing errors, try to fix all provided errors. This is useful if you already have parsing errors and want to speedup things a little bit (avoids re-parsing). Filename is used to filter errors. Caller may provide errors for multiple files, but only those that match filename will be addressed.
Fixed file contents are printed to `stdout` by default.
Flag `-i` causes `parsefix` to overwrite `-f` contents.
Exit code:
* 0 if at least one issue was fixed.
* 1 and no output if no issues were fixed. With `-i` no re-write is done.
* 2 and no output if there were no parsing issues at all. With `-i` no re-write is done.
* 3 some error occured, message is printed to stderr.
Examples:
```bash
# Uses 1st mode. parsefix does parsing itself and prints fixed code to stdout.
parsefix -f=foo/main.go
# Uses 2nd mode. No parsing is done by the parsefix.
parsefix -f=foo/main.go "foo/main.go:9:3: expected '{', found 'EOF'"
```