func Cache(u *urls.Url) error {
for _, s := range badUrlStrings {
if strings.Index(u.Url, s) != -1 {
return fmt.Errorf("Url contains bad substring '%s'.", s)
}
}
// Try a HEAD req first to get Content-Length header.
res, err := http.Head(u.Url)
if err != nil {
return err
}
if res.StatusCode != 200 {
return fmt.Errorf("Received non-200 response '%s' from server.",
res.Status)
}
if size := res.Header.Get("Content-Length"); size != "" {
if bytes, err := strconv.Atoi(size); err != nil {
return fmt.Errorf("Received unparseable content length '%s' "+
"from server: %v.", size, err)
} else if bytes > 1<<22 {
return fmt.Errorf("Response too large (%d MB) to cache safely.",
bytes/1024/1024)
}
}
res, err = http.Get(u.Url)
if err != nil {
return err
}
defer res.Body.Close()
// 1 << 22 == 4MB
if res.ContentLength > 1<<22 {
return fmt.Errorf("Response too large (%d MB) to cache safely.",
res.ContentLength/1024/1024)
}
u.CachedAs = Encode(u.Url)
fh, err := os.OpenFile(util.JoinPath(*urlCacheDir, u.CachedAs),
os.O_WRONLY|os.O_CREATE|os.O_TRUNC, os.FileMode(0600))
defer fh.Close()
if err != nil {
return err
}
if _, err := io.Copy(fh, res.Body); err != nil {
return err
}
u.CacheTime = time.Now()
u.MimeType = res.Header.Get("Content-Type")
if _, err := uc.UpsertId(u.Id, u); err != nil {
return err
}
return nil
}
func Cache(u *urls.Url) error {
for _, s := range badUrlStrings {
if strings.Index(u.Url, s) != -1 {
return fmt.Errorf("Url contains bad substring '%s'.", s)
}
}
res, err := http.Get(u.Url)
if err != nil {
return err
}
defer res.Body.Close()
if res.StatusCode != 200 {
return fmt.Errorf("Received non-200 response '%s' from server.",
res.Status)
}
// 1 << 22 == 4MB
if res.ContentLength > 1<<22 {
return fmt.Errorf("Response too large (%d MB) to cache safely.",
res.ContentLength/1024/1024)
}
u.CachedAs = Encode(u.Url)
fh, err := os.OpenFile(util.JoinPath(*urlCacheDir, u.CachedAs),
os.O_WRONLY|os.O_CREATE|os.O_TRUNC, os.FileMode(0600))
defer fh.Close()
if err != nil {
return err
}
if _, err := io.Copy(fh, res.Body); err != nil {
return err
}
u.CacheTime = time.Now()
u.MimeType = res.Header.Get("Content-Type")
if _, err := uc.UpsertId(u.Id, u); err != nil {
return err
}
return nil
}
"github.com/fluffle/sp0rkle/bot"
"github.com/fluffle/sp0rkle/collections/urls"
"github.com/fluffle/sp0rkle/util"
"gopkg.in/mgo.v2/bson"
)
const shortenPath string = "/s/"
const cachePath string = "/c/"
const autoShortenLimit int = 120
var badUrlStrings = []string{
"4chan",
}
var urlCacheDir *string = flag.String("url_cache_dir",
util.JoinPath(os.Getenv("HOME"), ".sp0rkle"),
"Path to store cached content under.")
var uc *urls.Collection
// Remember the last url seen on a per-channel basis
var lastseen = map[string]bson.ObjectId{}
func Init() {
uc = urls.Init()
if err := os.MkdirAll(*urlCacheDir, 0700); err != nil {
logging.Fatal("Couldn't create URL cache dir: %v", err)
}
bot.Handle(urlScan, client.PRIVMSG)