diff --git a/README.md b/README.md index be53e9c..56f8d2d 100644 --- a/README.md +++ b/README.md @@ -141,7 +141,7 @@ go install github.com/lapwat/papeer@latest ```sh # use platform=darwin for MacOS platform=linux -release=0.6.1 +release=0.6.2 # download and extract curl -L https://github.com/lapwat/papeer/releases/download/v$release/papeer-v$release-$platform-amd64.tar.gz > papeer.tar.gz diff --git a/book/link.go b/book/link.go index e776ce4..c23843f 100644 --- a/book/link.go +++ b/book/link.go @@ -1,7 +1,7 @@ package book type link struct { - Href string `json:"href"` + Href string `json:"url"` Text string `json:"name"` } diff --git a/book/scraper.go b/book/scraper.go index 2b8004d..e3808b8 100644 --- a/book/scraper.go +++ b/book/scraper.go @@ -365,7 +365,12 @@ func GetLinks(url *urllib.URL, selector string, limit, offset int, reverse, incl // RSS feed for _, item := range feed.Items { - links = append(links, NewLink(item.Link, item.Title)) + u, err := url.Parse(item.Link) + if err != nil { + log.Fatal(err) + } + + links = append(links, NewLink(u.String(), item.Title)) } pathMax = "RSS" @@ -385,11 +390,16 @@ func GetLinks(url *urllib.URL, selector string, limit, offset int, reverse, incl // visit and count link classes c := colly.NewCollector() c.OnHTML(selector, func(e *colly.HTMLElement) { - href := e.Attr("href") text := strings.TrimSpace(e.Text) path := GetPath(e.DOM) key := path + u, err := url.Parse(e.Attr("href")) + if err != nil { + log.Fatal(err) + } + href := u.String() + if selectorSet { // if selector is set, we use the selector specified by the user diff --git a/cmd/version.go b/cmd/version.go index 4739fa0..61cbcdc 100644 --- a/cmd/version.go +++ b/cmd/version.go @@ -14,6 +14,6 @@ var versionCmd = &cobra.Command{ Use: "version", Short: "Print the version number of papeer", Run: func(cmd *cobra.Command, args []string) { - fmt.Println("papeer v0.6.1") + fmt.Println("papeer v0.6.2") }, }