Improve action log display with control chars (#23820)

Close #23680

Some CLI programs use "\r" and control chars to print new content in
current line.

So, the strings in one line are actually from
`\rReading...1%\rReading...5%\rReading...100%`

This PR tries to make the output better.

2 files changed, 75 insertions, 3 deletions

diff --git a/web_src/js/components/RepoActionView.test.js b/web_src/js/components/RepoActionView.test.js
new file mode 100644
index 0000000000..011c20d8d1
--- /dev/null
+++ b/web_src/js/components/RepoActionView.test.js
@@ -0,0 +1,30 @@
+import {expect, test} from 'vitest';
+
+import {ansiLogToHTML} from './RepoActionView.vue';
+import AnsiToHTML from 'ansi-to-html';
+
+test('processConsoleLine', () => {
+  expect(ansiLogToHTML('abc')).toEqual('abc');
+  expect(ansiLogToHTML('abc\n')).toEqual('abc');
+  expect(ansiLogToHTML('abc\r\n')).toEqual('abc');
+  expect(ansiLogToHTML('\r')).toEqual('');
+  expect(ansiLogToHTML('\rx\rabc')).toEqual('x\nabc');
+  expect(ansiLogToHTML('\rabc\rx\r')).toEqual('abc\nx');
+
+  expect(ansiLogToHTML('\x1b[30mblack\x1b[37mwhite')).toEqual('<span style="color:#000">black<span style="color:#AAA">white</span></span>');
+  expect(ansiLogToHTML('<script>')).toEqual('&lt;script&gt;');
+
+
+  // upstream AnsiToHTML has bugs when processing "\033[1A" and "\033[1B", we fixed these control sequences in our code
+  // if upstream could fix these bugs, we can remove these tests and remove our patch code
+  const ath = new AnsiToHTML({escapeXML: true});
+  expect(ath.toHtml('\x1b[1A\x1b[2Ktest\x1b[1B\x1b[1A\x1b[2K')).toEqual('AtestBA'); // AnsiToHTML bug
+  expect(ath.toHtml('\x1b[1A\x1b[2K\rtest\r\x1b[1B\x1b[1A\x1b[2K')).toEqual('A\rtest\rBA'); // AnsiToHTML bug
+
+  // test our patched behavior
+  expect(ansiLogToHTML('\x1b[1A\x1b[2Ktest\x1b[1B\x1b[1A\x1b[2K')).toEqual('test');
+  expect(ansiLogToHTML('\x1b[1A\x1b[2K\rtest\r\x1b[1B\x1b[1A\x1b[2K')).toEqual('test');
+
+  // treat "\033[0K" and "\033[0J" (Erase display/line) as "\r", then it will be covered to "\n" finally.
+  expect(ansiLogToHTML('a\x1b[Kb\x1b[2Jc')).toEqual('a\nb\nc');
+});
diff --git a/web_src/js/components/RepoActionView.vue b/web_src/js/components/RepoActionView.vue
index ebc42829b4..fa09d2fb95 100644
--- a/web_src/js/components/RepoActionView.vue
+++ b/web_src/js/components/RepoActionView.vue
@@ -77,6 +77,8 @@ import AnsiToHTML from 'ansi-to-html';
 
 const {csrfToken} = window.config;
 
+const ansiLogRender = new AnsiToHTML({escapeXML: true});
+
 const sfc = {
   name: 'RepoActionView',
   components: {
@@ -91,8 +93,6 @@ const sfc = {
 
   data() {
     return {
-      ansiToHTML: new AnsiToHTML({escapeXML: true}),
-
       // internal state
       loading: false,
       intervalID: null,
@@ -214,7 +214,7 @@ const sfc = {
 
       const logMessage = document.createElement('div');
       logMessage.className = 'log-msg';
-      logMessage.innerHTML = this.ansiToHTML.toHtml(line.message);
+      logMessage.innerHTML = ansiLogToHTML(line.message);
       div.appendChild(logMessage);
 
       return div;
@@ -307,6 +307,48 @@ export function initRepositoryActionView() {
   view.mount(el);
 }
 
+// some unhandled control sequences by AnsiToHTML
+// https://man7.org/linux/man-pages/man4/console_codes.4.html
+const ansiRegexpRemove = /\x1b\[\d+[A-H]/g; // Move cursor, treat them as no-op.
+const ansiRegexpNewLine = /\x1b\[\d?[JK]/g; // Erase display/line, treat them as a Carriage Return
+
+function ansiCleanControlSequences(line) {
+  if (line.includes('\x1b')) {
+    line = line.replace(ansiRegexpRemove, '');
+    line = line.replace(ansiRegexpNewLine, '\r');
+  }
+  return line;
+}
+
+export function ansiLogToHTML(line) {
+  if (line.endsWith('\r\n')) {
+    line = line.substring(0, line.length - 2);
+  } else if (line.endsWith('\n')) {
+    line = line.substring(0, line.length - 1);
+  }
+
+  // usually we do not need to process control chars like "\033[", let AnsiToHTML do it
+  // but AnsiToHTML has bugs, so we need to clean some control sequences first
+  line = ansiCleanControlSequences(line);
+
+  if (!line.includes('\r')) {
+    return ansiLogRender.toHtml(line);
+  }
+
+  // handle "\rReading...1%\rReading...5%\rReading...100%",
+  // convert it into a multiple-line string: "Reading...1%\nReading...5%\nReading...100%"
+  const lines = [];
+  for (const part of line.split('\r')) {
+    if (part === '') continue;
+    const partHtml = ansiLogRender.toHtml(part);
+    if (partHtml !== '') {
+      lines.push(partHtml);
+    }
+  }
+  // the log message element is with "white-space: break-spaces;", so use "\n" to break lines
+  return lines.join('\n');
+}
+
 </script>
 
 <style scoped>