msgviewer: open http links from messages

Parse http links from a message and display them as completions in the
:open-link command.

Add the following binds to the [view] section in your binds.conf:
<C-l> = :open-link <space>

Parsing can be disabled in aerc.conf by setting parse-http-links to
false in the viewer section.

Thanks to Moritz for the help with the regular expression.

Signed-off-by: Koni Marti <koni.marti@gmail.com>
Reviewed-by: Moritz Poldrack <git@moritz.sh>
Acked-by: Robin Jarry <robin@jarry.cc>

9 files changed, 192 insertions, 2 deletions

diff --git a/commands/msgview/open.go b/commands/msgview/open.go
index d8a4de0..3d62fe2 100644
--- a/commands/msgview/open.go
+++ b/commands/msgview/open.go
@@ -8,6 +8,7 @@ import (
 	"os"
 	"time"
 
+	"git.sr.ht/~rjarry/aerc/commands"
 	"git.sr.ht/~rjarry/aerc/lib"
 	"git.sr.ht/~rjarry/aerc/logging"
 	"git.sr.ht/~rjarry/aerc/widgets"
@@ -20,10 +21,16 @@ func init() {
 }
 
 func (Open) Aliases() []string {
-	return []string{"open"}
+	return []string{"open", "open-link"}
 }
 
 func (Open) Complete(aerc *widgets.Aerc, args []string) []string {
+	mv := aerc.SelectedTab().(*widgets.MessageViewer)
+	if mv != nil {
+		if p := mv.SelectedMessagePart(); p != nil {
+			return commands.CompletionFromList(aerc, p.Links, args)
+		}
+	}
 	return nil
 }
 
@@ -31,6 +38,17 @@ func (Open) Execute(aerc *widgets.Aerc, args []string) error {
 	mv := aerc.SelectedTab().(*widgets.MessageViewer)
 	p := mv.SelectedMessagePart()
 
+	if args[0] == "open-link" && len(args) > 1 {
+		if link := args[1]; link != "" {
+			go func() {
+				if err := lib.NewXDGOpen(link).Start(); err != nil {
+					aerc.PushError(fmt.Sprintf("%s: %s", args[0], err.Error()))
+				}
+			}()
+		}
+		return nil
+	}
+
 	store := mv.Store()
 	store.FetchBodyPart(p.Msg.Uid, p.Index, func(reader io.Reader) {
 		extension := ""
diff --git a/config/aerc.conf b/config/aerc.conf
index 1a3c1f4..8043412 100644
--- a/config/aerc.conf
+++ b/config/aerc.conf
@@ -209,6 +209,12 @@ header-layout=From|To,Cc|Bcc,Date,Subject
 # Default: false
 always-show-mime=false
 
+# Parses and extracts http links when viewing a message. Links can then be
+# accessed with the open-link command.
+#
+# Default: true
+parse-http-links=true
+
 # How long to wait after the last input before auto-completion is triggered.
 #
 # Default: 250ms
diff --git a/config/binds.conf b/config/binds.conf
index 4f67bbe..589dc55 100644
--- a/config/binds.conf
+++ b/config/binds.conf
@@ -64,6 +64,8 @@ S = :save<space>
 D = :delete<Enter>
 A = :archive flat<Enter>
 
+<C-l> = :open-link <space>
+
 f = :forward<Enter>
 rr = :reply -a<Enter>
 rq = :reply -aq<Enter>
diff --git a/config/config.go b/config/config.go
index 923a1a0..dae32b7 100644
--- a/config/config.go
+++ b/config/config.go
@@ -160,6 +160,7 @@ type ViewerConfig struct {
 	Alternatives   []string
 	ShowHeaders    bool       `ini:"show-headers"`
 	AlwaysShowMime bool       `ini:"always-show-mime"`
+	ParseHttpLinks bool       `ini:"parse-http-links"`
 	HeaderLayout   [][]string `ini:"-"`
 	KeyPassthrough bool       `ini:"-"`
 }
@@ -705,6 +706,7 @@ func LoadConfigFromFile(root *string, logger *log.Logger) (*AercConfig, error) {
 				{"Date"},
 				{"Subject"},
 			},
+			ParseHttpLinks: true,
 		},
 
 		Statusline: StatuslineConfig{
diff --git a/doc/aerc-config.5.scd b/doc/aerc-config.5.scd
index 15c6ca3..47ae752 100644
--- a/doc/aerc-config.5.scd
+++ b/doc/aerc-config.5.scd
@@ -401,6 +401,12 @@ These options are configured in the *[viewer]* section of aerc.conf.
 
 	Default: false
 
+*parse-http-links*
+	Parses and extracts http links when viewing a message. Links can then be
+	accessed with the open-link command.
+
+	Default: true
+
 ## COMPOSE
 
 These options are configured in the *[compose]* section of aerc.conf.
diff --git a/lib/parse/hyperlinks.go b/lib/parse/hyperlinks.go
new file mode 100644
index 0000000..7a00538
--- /dev/null
+++ b/lib/parse/hyperlinks.go
@@ -0,0 +1,44 @@
+package parse
+
+import (
+	"bufio"
+	"bytes"
+	"io"
+	"regexp"
+	"strings"
+)
+
+var submatch = `(https?:\/\/[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,10}\b(?:[-a-zA-Z0-9()@:%_\+.~#?&\/=]*))`
+var httpRe = regexp.MustCompile("\"" + submatch + "\"" + "|" + "\\(" + submatch + "\\)" + "|" + "<" + submatch + ">" + "|" + submatch)
+
+// HttpLinks searches a reader for a http link and returns a copy of the
+// reader and a slice with links.
+func HttpLinks(r io.Reader) (io.Reader, []string) {
+	var buf bytes.Buffer
+	tr := io.TeeReader(r, &buf)
+
+	scanner := bufio.NewScanner(tr)
+	linkMap := make(map[string]struct{})
+	for scanner.Scan() {
+		line := scanner.Text()
+		if !strings.Contains(line, "http") {
+			continue
+		}
+		for _, word := range strings.Fields(line) {
+			if links := httpRe.FindStringSubmatch(word); len(links) > 0 {
+				for _, l := range links[1:] {
+					if l != "" {
+						linkMap[strings.TrimSpace(l)] = struct{}{}
+					}
+				}
+			}
+		}
+	}
+
+	results := []string{}
+	for link, _ := range linkMap {
+		results = append(results, link)
+	}
+
+	return &buf, results
+}
diff --git a/lib/parse/hyperlinks_test.go b/lib/parse/hyperlinks_test.go
new file mode 100644
index 0000000..f16392a
--- /dev/null
+++ b/lib/parse/hyperlinks_test.go
@@ -0,0 +1,99 @@
+package parse_test
+
+import (
+	"io/ioutil"
+	"strings"
+	"testing"
+
+	"git.sr.ht/~rjarry/aerc/lib/parse"
+)
+
+func TestHyperlinks(t *testing.T) {
+	tests := []struct {
+		text  string
+		links []string
+	}{
+		{
+			text:  "http://aerc-mail.org",
+			links: []string{"http://aerc-mail.org"},
+		},
+		{
+			text:  "https://aerc-mail.org",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text https://aerc-mail.org more text",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text (https://aerc-mail.org) more text",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text \"https://aerc-mail.org\" more text",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text <https://aerc-mail.org> more text",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "<a href=\"https://aerc-mail.org\">",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text https://aerc-mail.org more text https://aerc-mail.org more text",
+			links: []string{"https://aerc-mail.org"},
+		},
+		{
+			text:  "text https://aerc-mail.org more text http://git.sr.ht/~rjarry/aerc more text",
+			links: []string{"https://aerc-mail.org", "http://git.sr.ht/~rjarry/aerc"},
+		},
+		{
+			text:  "text http://www.ietf.org/rfc/rfc2396.txt more text",
+			links: []string{"http://www.ietf.org/rfc/rfc2396.txt"},
+		},
+		{
+			text:  "text <http://example.com:8042/over/there?name=ferret#nose> more text",
+			links: []string{"http://example.com:8042/over/there?name=ferret#nose"},
+		},
+		{
+			text:  "text http://cnn.example.com&story=breaking_news@10.0.0.1/top_story.htm more text",
+			links: []string{"http://cnn.example.com&story=breaking_news@10.0.0.1/top_story.htm"},
+		},
+		{
+			text:  "text https://www.ics.uci.edu/pub/ietf/uri/#Related more text",
+			links: []string{"https://www.ics.uci.edu/pub/ietf/uri/#Related"},
+		},
+		{
+			text:  "text https://www.example.com/index.php?id_sezione=360&sid=3a5ebc944f41daa6f849f730f1 more text",
+			links: []string{"https://www.example.com/index.php?id_sezione=360&sid=3a5ebc944f41daa6f849f730f1"},
+		},
+	}
+
+	for _, test := range tests {
+
+		// make sure reader is exact copy of input reader
+		reader, links := parse.HttpLinks(strings.NewReader(test.text))
+		if data, err := ioutil.ReadAll(reader); err != nil {
+			t.Errorf("could not read text: %v", err)
+		} else if string(data) != test.text {
+			t.Errorf("did not copy input reader correctly")
+		}
+
+		// check correct parsed links
+		if len(links) != len(test.links) {
+			t.Errorf("different number of links: got %d but expected %d", len(links), len(test.links))
+		}
+		linkMap := make(map[string]struct{})
+		for _, got := range links {
+			linkMap[got] = struct{}{}
+		}
+		for _, expected := range test.links {
+			if _, ok := linkMap[expected]; !ok {
+				t.Errorf("link not parsed: %s", expected)
+			}
+		}
+
+	}
+}
diff --git a/widgets/msgviewer.go b/widgets/msgviewer.go
index 9876467..b41b57c 100644
--- a/widgets/msgviewer.go
+++ b/widgets/msgviewer.go
@@ -19,6 +19,7 @@ import (
 	"git.sr.ht/~rjarry/aerc/lib"
 	"git.sr.ht/~rjarry/aerc/lib/auth"
 	"git.sr.ht/~rjarry/aerc/lib/format"
+	"git.sr.ht/~rjarry/aerc/lib/parse"
 	"git.sr.ht/~rjarry/aerc/lib/ui"
 	"git.sr.ht/~rjarry/aerc/logging"
 	"git.sr.ht/~rjarry/aerc/models"
@@ -325,6 +326,7 @@ func (mv *MessageViewer) SelectedMessagePart() *PartInfo {
 		Index: part.index,
 		Msg:   part.msg.MessageInfo(),
 		Part:  part.part,
+		Links: part.links,
 	}
 }
 
@@ -518,6 +520,8 @@ type PartViewer struct {
 	term        *Terminal
 	grid        *ui.Grid
 	uiConfig    config.UIConfig
+
+	links []string
 }
 
 func NewPartViewer(acct *AccountView, conf *config.AercConfig,
@@ -670,6 +674,14 @@ func (pv *PartViewer) writeMailHeaders() {
 	}
 }
 
+func (pv *PartViewer) hyperlinks(r io.Reader) (reader io.Reader) {
+	if !pv.conf.Viewer.ParseHttpLinks {
+		return r
+	}
+	reader, pv.links = parse.HttpLinks(r)
+	return reader
+}
+
 func (pv *PartViewer) copyFilterOutToPager() {
 	stdout, _ := pv.filter.StdoutPipe()
 	stderr, _ := pv.filter.StderrPipe()
@@ -708,7 +720,7 @@ func (pv *PartViewer) copyFilterOutToPager() {
 }
 
 func (pv *PartViewer) copySourceToSinkStripAnsi() {
-	scanner := bufio.NewScanner(pv.source)
+	scanner := bufio.NewScanner(pv.hyperlinks(pv.source))
 	// some people send around huge html without any newline in between
 	// this did overflow the default 64KB buffer of bufio.Scanner.
 	// If something can't fit in a GB there's no hope left
diff --git a/widgets/providesmessage.go b/widgets/providesmessage.go
index c1d821c..b0f261d 100644
--- a/widgets/providesmessage.go
+++ b/widgets/providesmessage.go
@@ -10,6 +10,7 @@ type PartInfo struct {
 	Index []int
 	Msg   *models.MessageInfo
 	Part  *models.BodyStructure
+	Links []string
 }
 
 type ProvidesMessage interface {