msgconv/matrixfmt: fix trimming space from whitespace-only string

mautrix · Sep 13, 2024 · adee821 · adee821
1 parent 4ede5dc
commit adee821
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 5 deletions.
diff --git a/pkg/msgconv/matrixfmt/convert_test.go b/pkg/msgconv/matrixfmt/convert_test.go
@@ -65,6 +65,13 @@ func TestParse_HTML(t *testing.T) {
  Length: 5,
  Value: signalfmt.StyleBold,
  }}},
+ {name: "UnnecessaryWhitespace", in: "<strong> Hello </strong>, World!", out: "Hello, World!", ent: signalfmt.BodyRangeList{{
+ Start: 0,
+ Length: 5,
+ Value: signalfmt.StyleBold,
+ }}},
+ {name: "UnnecessaryWhitespaceParagraph", in: "<p> Hello </p>", out: "Hello"},
+ {name: "EmptyParagraph", in: "<p>Hello</p><p> </p>", out: "Hello"},
  {
  name: "MultiBasic",
  in: "<strong><em>Hell</em>o</strong>, <del>Wo<span data-mx-spoiler>rld</span></del><code>!</code>",

diff --git a/pkg/msgconv/matrixfmt/html.go b/pkg/msgconv/matrixfmt/html.go
@@ -81,22 +81,26 @@ func (es *EntityString) TrimSpace() *EntityString {
  return nil
  }
  DebugLog("TRIMSPACE %q %+v\n", es.String, es.Entities)
- var cutEnd, cutStart int
- for cutStart = 0; cutStart < len(es.String); cutStart++ {
+ cutStart := 0
+ for ; cutStart < len(es.String); cutStart++ {
  switch es.String[cutStart] {
  case '\t', '\n', '\v', '\f', '\r', ' ', 0x85, 0xA0:
  continue
  }
  break
  }
- for cutEnd = len(es.String) - 1; cutEnd >= 0; cutEnd-- {
- switch es.String[cutEnd] {
+ cutEnd := len(es.String)
+ for ; cutEnd > cutStart; cutEnd-- {
+ switch es.String[cutEnd-1] {
  case '\t', '\n', '\v', '\f', '\r', ' ', 0x85, 0xA0:
  continue
  }
  break
  }
- cutEnd++
+ if cutEnd == cutStart {
+ DebugLog(" -> EMPTY\n")
+ return NewEntityString("")
+ }
  if cutStart == 0 && cutEnd == len(es.String) {
  DebugLog(" -> NOOP\n")
  return es