I need to write to a file in a nested directory that may or may not exist.
At first, I checked if the folder existed via os.Stat, doing os.MkdirAll if it doesn't exist, and then opening and writing to a file.
I tried removing the os.Stat and just doing os.MkdirAll, and it appears to work - meaning os.MkdirAll is idempotent.
My question is, is there a benefit of doing the os.Stat check? Is it a much lighter operation than os.MkdirAll?
The first thing MkdirAll does is call os.Stat to check if the path exists and is a directory.
func MkdirAll(path string, perm FileMode) error {
// Fast path: if we can tell whether path is a directory or file, stop with success or error.
dir, err := Stat(path)
if err == nil {
if dir.IsDir() {
return nil
}
return &PathError{"mkdir", path, syscall.ENOTDIR}
}
...
From the docs:
If path is already a directory, MkdirAll does nothing and returns nil.
So no, you don't need to call os.Stat.
Related
Using Go - lang, according to the documentation, the os.Rename should be able to rename either a file or directory on any operating system.
On Linux it works as it should, pass either a file or directory into it and the file or directory are moved.
On windows i recieve an 'Access is denied' Error when trying to pass a folder.
It works 100% for files.
example:
source = c:\sourcefolder
destination = c:\destinationfolder
source contains:
C:\sourcefolder\file1.xml
C:\sourcefolder\file2.xml
C:\sourcefolder\foldername1
C:\sourcefolder\foldername1\file3.xml
C:\sourcefolder\foldername2
C:\sourcefolder\foldername2\file4.xml
both file1.xml and file2.xml will successfully copy to c:\destination.
But the folders and files within the folders crash out with access denied
The script is pretty simple:
source := "C:\\sourcefolder"
destination := "C:\\destinationfolder"
pathSeperator := "\\"
files, err := ioutil.ReadDir(source)
if err != nil {
fmt.Println("Move command execution error: ", err)
}
for _, f := range files {
fmt.Println(f.Name())
fmt.Println(f.Mode())
err := os.Rename(source+pathSeperator+f.Name(), destination+pathSeperator+f.Name())
if err != nil {
fmt.Println("Move command execution error: ", err)
panic(err)
}
}
Having searched stackoverflow and golang's resources, i found the issue listed in 2016 that reported this fault and according to the issue it was fixed, but i am unable to get this to work. Nowhere else that i can find lists this issue go golang.
checking the f.Mode for access, i get drwxrwxrwx and have complete access to all the files and directories.
Any help with this would be great, racking my mind. Thank you.
Quoted from comment. solved my issue.
Found the cause of the fault to be, if a windows explorer window is
open and has ANY visibility of the folders being moved (i.e. in the
tree on the left or right-pane) then access is denied as it can not
move the folders. If i minimize all the tree's so that the
source\destination folders are not visible and select a different sub
folder in windows explorer then the os.Rename works as it should,
moving all content from A to B really quick (as per linux)
I had the same issue with copying files within the same folder. The following solution works just fine (without closing or minimizing windows):
// read original file
origFile, _:= os.ReadFile(filePath)
// create new file with a different name
newFile, _ := os.Create(filePath + ".new")
// print data from original file to new file.
fmt.Fprintf(newFile, "%s", string(origFile))
I'm trying to read from a file in my project's directory.
My problem is, that depending on the caller, the path changes. The caller changes, because I want to unit test this code and the caller is not Main.go anymore.
This is what my project structure looks like:
The code where I try to access specialChars.txt from looks like this:
func RemoveSpecialChars(word string) string {
file, err := ioutil.ReadFile("wordlists/specialChars.txt")
[...]
}
This code works for the start from Main.go but not for the start from CleanupUtil_test.go. To get it working from the test I would need file, err := ioutil.ReadFile("../wordlists/specialChars.txt")
I found answers like this one: https://stackoverflow.com/a/32163888/2837489
_, filename, _, ok := runtime.Caller(0) which is obviously also dependent on the caller.
Is it possible to get the projects root path independent of the calling function?
Or is my code design wrong? Should I pass the file path into the function?
Starting from Go 1.16, you can use the embed package. This allows you to embed the files in the running go program. It comes with the caveat that the referenced directory needs to exist at or below the embedding file. In your case, the structure would look as follows:
-- main.go
-- cleanup
-- wordlist
\- specialChars.txt
CleanupUtil.go
CleanupUtil_test.go
You can reference the file using a go directive
// CleanupUtil.go
package cleanup
import (
"embed"
)
//go:embed wordlists/specialChars.txt
var content embed.FS
func RemoveSpecialChars(word string) string {
file, err := content.ReadFile("wordlists/specialChars.txt")
[...]
}
This program will run successfully regardless of where the program is executed. You should be able to reference this code in both your main.go file and your CleanupUtil_test.go file.
Pass in the filepath as a parameter to the function (as indicated in your last question).
More details:
The relative path "wordlists/specialChars.txt" is in fact not dependent on where the source file is located (such as Main.go or CleanupUtil_test.go), but where you execute it from. So you could run your tests from your root directory and then it would actually work. In short, the current working directory is relevant.
Still, I'd recommend specifying the path, because that makes your function more reusable and universal.
Maybe you don't even need to put this information into a file, but can simply have a string containing those chars. In this case you could also check if https://golang.org/pkg/regexp/#Regexp.ReplaceAll already covers your use case.
I'm trying to package my go application binary which is accessible by a web interface running on localhost:8080 so that when downloaded it's able to find the JS(front-end) files in the folder where the file is ran but i can't seem to make it work.
I've been doing something like this :
pwd, err := os.Getwd()
if err != nil {
fmt.Println(err)
os.Exit(1)
}
Then trying to use the working directory of the binary to access the files inside it but that doesn't seem to work.
The binary is located at :
/Users/admin/Desktop/testappfolder
but when i run the program with just :
pwd, err := os.Getwd()
if err != nil {
fmt.Println(err)
os.Exit(1)
}
fmt.Println(pwd)
I end up getting /Users/admin as the working directory instead.
I wondering :
Where i'm going wrong ?
Is this has something to do with the Gopath ?
Am i going at it the right way regarding distributing the app as a "zip" and having file path setup directly inside my program relative to the working directory ? or is it that logic that's wrong ?
os.Getwd is going to correspond to where you start your binary from not where the binary is located.
To make for a more robust solution I would pass in the location of the files directory using a flag or using a config value.
I have a project with next structure:
|_main.go
|_config
|_config.go
|_config_test.go
|_config.json
I'm having next code line in config.go:
file, _ := os.Open("config/config.json")
When I'm executing method contained this code line from main.go all is working. But when I'm trying to execute this method from config_test.go it produces error:
open config/config.json: no such file or directory
As I understood it is a working directory issue because I'm launching same code with relative path from different directories. How can I fix this problem without using full path in config.go?
Relative paths are always resolved basis your current directory. Hence it's better to avoid relative paths.
Use command line flags or a configuration management tool (better approach) like Viper
Also according to The Twelve-Factor App your config files should be outside your project.
Eg usage with Viper:
import "github.com/spf13/viper"
func init() {
viper.SetConfigName("config")
// Config files are stored here; multiple locations can be added
viper.AddConfigPath("$HOME/configs")
errViper := viper.ReadInConfig()
if errViper != nil {
panic(errViper)
}
// Get values from config.json
val := viper.GetString("some_key")
// Use the value
}
Now I am using Walk to go through my folder, I want to filter some folder when scan.
err := filepath.Walk("/home", func(path string, f os.FileInfo, err error) error {
...
})
The folder structure is:
home
/ | \
a b c
Can I make some exception list that filepath.Walk not scan folder a? That is, I do not want any files in folder a add into my scan result.
From the documentation for WalkFuc:
If an error is returned, processing stops. The sole exception is that if path is a directory and the function returns the special value SkipDir, the contents of the directory are skipped and processing continues as usual on the next file.
So simply make the function you pass to filepath.Walk return filepath.SkipDir when path is the directory you want skipped.