lewisdale.dev/src/blog/posts/2024/5/learning-go-day-11.md
Lewis Dale 406383e3c1
All checks were successful
Build and copy to prod / build-and-copy (push) Successful in 2m4s
Finish tomorrow's WeblogPoMo post
2024-05-12 21:48:27 +01:00

160 lines
4.1 KiB
Markdown

---
title: "Learning Go: Day Eleven"
date: 2024-05-13T08:00:00.0Z
tags:
- learning
- go
excerpt: "Now it's time to actually try and send a \"ping\" to a website"
---
So, next up on my task list is to start populating my database with data. So to do that I want to be able to send a request to a given website, and then store the success or failure status based on the response.
## Sending a request
It turns out that this is actually quite simple to do. The `net/http` module has functions to do this, so all I need to do is:
```go
// ping/ping.go
import (
"net/http"
)
func SendPing(db *sql.DB, site sites.Site) {
response, error := http.Get(site.url)
// Do something with the response or error
}
```
Easy enough!
## Store the result
Now all I need to do is save the output to the database. I do that by using the inline if-statement syntax, and assigning `ping.Status` depending on whether or not there is an error present. I don't actually _care_ about the response, I just care whether or not the call succeeded.
```go
// ping/ping.go
func SendPing(db *sql.DB, site sites.Site) {
p := Ping{
Site: site,
}
if _, err := http.Get(site.Url); err != nil {
p.Status = Failure
} else {
p.Status = Success
}
p.Save(db)
}
```
And as a little refactor, I can default the `Status` to `Success`, and get rid of the `else-` branch.
```go
// ping/ping.go
func SendPing(db *sql.DB, site sites.Site) {
p := Ping{
Site: site,
Status: Success
}
if _, err := http.Get(site.Url); err != nil {
p.Status = Failure
}
p.Save(db)
}
```
Now, if I write a quick test in my `main.go`[^1], I can see that it succeeds:
```go
// main.go
func main() {
site := sites.Site{Name: "Lewisdale.dev", Url: "https://lewisdale.dev"}
site.Save(db)
ping.SendPing(db, site)
}
```
```bash
sqlite> SELECT ping.id, ping.site, ping.timestamp, statuses.name FROM ping LEFT JOIN statuses ON (ping.status = statuses.id);
1|https://lewisdale.dev|2024-05-10 05:57:12|Success
```
And if I then force a failure, it should also work:
```bash
5|https://notreal.tld|2024-05-10 08:00:46|Failure
```
## Adding some output
Right, now to actually output the values in the database. First of all, I've defined the Struct for the response:
```go
// ping/ping.go
type PingResponse struct {
Site sites.Site
Timestamp string
Status string
}
```
And now I've added a `List` function that reads the data I need from the database, and places it into an array slice of `PingResponse` values:
```go
func List(db *sql.DB) []PingResponse {
rows, err := db.Query(`SELECT sites.url as url, sites.name, ping.timestamp as timestamp, statuses.name as status FROM ping
JOIN sites ON ping.site = sites.url
JOIN statuses ON ping.status = statuses.id
ORDER BY timestamp DESC`)
if err != nil {
panic(err)
}
defer rows.Close()
pings := make([]PingResponse, 0)
for rows.Next() {
p := PingResponse{}
rows.Scan(&p.Site.Url, &p.Site.Name, &p.Timestamp, &p.Status)
pings = append(pings, p)
}
return pings
}
```
The interesting parts here are the `defer` statement, and `rows.Scan`. Defer queues that call up until after the function has executed, it's just a way of saying "I will be doing this at the end regardless" as a cleanup operation[^2]. Then `rows.Scan` will automagically insert the values to the variables I pass it, in the order the columns are read from the database[^3].
Then finally, I can update my handler function so that it uses `json.Marshal` to convert the pings to JSON, and output them to the browser:
```go
http.HandleFunc("GET /", func(w http.ResponseWriter, r *http.Request) {
pings := ping.List(db)
if output, err := json.Marshal(pings); err != nil {
w.Write([]byte(err.Error()))
w.WriteHeader(http.StatusInternalServerError)
return
} else {
w.Header().Set("Content-Type", "application/json")
w.Write(output)
}
})
```
And that works! You can see it at https://oopsie.lewisdale.dev, with (hopefully) some actual output.
[^1]: Yes, despite what I said in another post I've not written any _actual_ tests. I'm human, alright?
[^2]: I think
[^3]: This is where it helps to be explicit with what is selected and avoid `SELECT *`.