diff options
| author | Shulhan <ms@kilabit.info> | 2025-05-29 13:15:08 +0700 |
|---|---|---|
| committer | Shulhan <ms@kilabit.info> | 2025-05-29 13:15:08 +0700 |
| commit | cce49486f36bb4ba7702973cc598af9bcb1f4103 (patch) | |
| tree | 2daf641658141834e755f15c526e13ce99f3f98b | |
| parent | fa85558a0d1b20e4d203ddb537c3bde51bc5131f (diff) | |
| download | jarink-cce49486f36bb4ba7702973cc598af9bcb1f4103.tar.xz | |
all: add case for broken HTML
Turn out broken HTML still get parsed by "net/html" package.
| -rw-r--r-- | deadlinks_test.go | 10 | ||||
| -rw-r--r-- | testdata/web/broken.html | 7 | ||||
| -rw-r--r-- | testdata/web/index.html | 1 |
3 files changed, 18 insertions, 0 deletions
diff --git a/deadlinks_test.go b/deadlinks_test.go index a43df3d..9570f50 100644 --- a/deadlinks_test.go +++ b/deadlinks_test.go @@ -65,6 +65,10 @@ func TestDeadLinks_Scan(t *testing.T) { Link: `https://kilabit.info/brokenPage`, Code: http.StatusNotFound, }}, + testUrl + `/broken.html`: []deadlinks.Broken{{ + Link: testUrl + `/brokenPage`, + Code: http.StatusNotFound, + }}, testUrl + `/page2`: []deadlinks.Broken{{ Link: testUrl + `/broken.png`, Code: http.StatusNotFound, @@ -89,6 +93,10 @@ func TestDeadLinks_Scan(t *testing.T) { Link: `https://kilabit.info/brokenPage`, Code: http.StatusNotFound, }}, + testUrl + `/broken.html`: []deadlinks.Broken{{ + Link: testUrl + `/brokenPage`, + Code: http.StatusNotFound, + }}, testUrl + `/page2`: []deadlinks.Broken{{ Link: testUrl + `/broken.png`, Code: http.StatusNotFound, @@ -113,6 +121,8 @@ func TestDeadLinks_Scan(t *testing.T) { tcase.expError, err.Error()) continue } + //got, _ := json.MarshalIndent(result.PageLinks, ``, ` `) + //t.Logf(`got=%s`, got) test.Assert(t, tcase.scanUrl, tcase.exp, result.PageLinks) } } diff --git a/testdata/web/broken.html b/testdata/web/broken.html new file mode 100644 index 0000000..533e542 --- /dev/null +++ b/testdata/web/broken.html @@ -0,0 +1,7 @@ +<html> + <head></head> + <body> + <a href="/brokenPage" + <p> + </body> +</html> diff --git a/testdata/web/index.html b/testdata/web/index.html index 1124813..19a2bae 100644 --- a/testdata/web/index.html +++ b/testdata/web/index.html @@ -8,5 +8,6 @@ SPDX-License-Identifier: GPL-3.0-only <a href="/brokenPage">Broken page</a> <a href="/page2">Page 2</a> <a href="https://kilabit.info/brokenPage">Broken page at kilabit.info</a> + <a href="/broken.html">Broken HTML</a> </body> </html> |