move whitespace stripping to parser, not renderers
Russ Ross russ@russross.com
Wed, 29 Jun 2011 15:38:35 -0600
4 files changed,
133 insertions(+),
148 deletions(-)
M
block.go
→
block.go
@@ -258,7 +258,12 @@
if j > 0 { size := i + j if doRender { - parser.r.BlockHtml(out, data[:size]) + // trim newlines + end := size + for end > 0 && data[end-1] == '\n' { + end-- + } + parser.r.BlockHtml(out, data[:end]) } return size }@@ -280,7 +285,12 @@ j = parser.isEmpty(data[i:])
if j > 0 { size := i + j if doRender { - parser.r.BlockHtml(out, data[:size]) + // trim newlines + end := size + for end > 0 && data[end-1] == '\n' { + end-- + } + parser.r.BlockHtml(out, data[:end]) } return size }@@ -326,7 +336,12 @@ }
// the end of the block has been found if doRender { - parser.r.BlockHtml(out, data[:i]) + // trim newlines + end := i + for end > 0 && data[end-1] == '\n' { + end-- + } + parser.r.BlockHtml(out, data[:end]) } return i@@ -931,11 +946,11 @@ end++
} // get working buffers - var work bytes.Buffer - var inter bytes.Buffer + var rawItem bytes.Buffer + var parsed bytes.Buffer // put the first line into the working buffer - work.Write(data[beg:end]) + rawItem.Write(data[beg:end]) beg = end // process the following lines@@ -984,7 +999,7 @@ break
} if sublist == 0 { - sublist = work.Len() + sublist = rawItem.Len() } } else { // how about a nested prefix header?@@ -1002,7 +1017,7 @@ *flags |= LIST_ITEM_END_OF_LIST
break } else { if containsBlankLine { - work.WriteByte('\n') + rawItem.WriteByte('\n') containsBlock = true } }@@ -1012,7 +1027,7 @@
containsBlankLine = false // add the line into the working buffer without prefix - work.Write(data[beg+i : end]) + rawItem.Write(data[beg+i : end]) beg = end }@@ -1021,27 +1036,32 @@ if containsBlock {
*flags |= LIST_ITEM_CONTAINS_BLOCK } - workbytes := work.Bytes() + rawItemBytes := rawItem.Bytes() if *flags&LIST_ITEM_CONTAINS_BLOCK != 0 { // intermediate render of block li - if sublist > 0 && sublist < len(workbytes) { - parser.parseBlock(&inter, workbytes[:sublist]) - parser.parseBlock(&inter, workbytes[sublist:]) + if sublist > 0 && sublist < len(rawItemBytes) { + parser.parseBlock(&parsed, rawItemBytes[:sublist]) + parser.parseBlock(&parsed, rawItemBytes[sublist:]) } else { - parser.parseBlock(&inter, workbytes) + parser.parseBlock(&parsed, rawItemBytes) } } else { // intermediate render of inline li - if sublist > 0 && sublist < len(workbytes) { - parser.parseInline(&inter, workbytes[:sublist]) - parser.parseBlock(&inter, workbytes[sublist:]) + if sublist > 0 && sublist < len(rawItemBytes) { + parser.parseInline(&parsed, rawItemBytes[:sublist]) + parser.parseBlock(&parsed, rawItemBytes[sublist:]) } else { - parser.parseInline(&inter, workbytes) + parser.parseInline(&parsed, rawItemBytes) } } // render li itself - parser.r.ListItem(out, inter.Bytes(), *flags) + parsedBytes := parsed.Bytes() + parsedEnd := len(parsedBytes) + for parsedEnd > 0 && parsedBytes[parsedEnd-1] == '\n' { + parsedEnd-- + } + parser.r.ListItem(out, parsedBytes[:parsedEnd], *flags) return beg }
M
html.go
→
html.go
@@ -19,6 +19,7 @@ import (
"bytes" "fmt" "strconv" + "strings" ) const (@@ -126,10 +127,7 @@ }
func (options *Html) Header(out *bytes.Buffer, text func() bool, level int) { marker := out.Len() - - if marker > 0 { - out.WriteByte('\n') - } + doubleSpace(out) if options.flags&HTML_TOC != 0 { // headerCount is incremented in htmlTocHeader@@ -157,28 +155,13 @@ if options.flags&HTML_SKIP_HTML != 0 {
return } - sz := len(text) - for sz > 0 && text[sz-1] == '\n' { - sz-- - } - org := 0 - for org < sz && text[org] == '\n' { - org++ - } - if org >= sz { - return - } - if out.Len() > 0 { - out.WriteByte('\n') - } - out.Write(text[org:sz]) + doubleSpace(out) + out.Write(text) out.WriteByte('\n') } func (options *Html) HRule(out *bytes.Buffer) { - if out.Len() > 0 { - out.WriteByte('\n') - } + doubleSpace(out) out.WriteString("<hr") out.WriteString(options.closeTag) }@@ -192,44 +175,33 @@ }
} func (options *Html) BlockCodeNormal(out *bytes.Buffer, text []byte, lang string) { - if out.Len() > 0 { - out.WriteByte('\n') - } - - if lang != "" { - out.WriteString("<pre><code class=\"") - - for i, cls := 0, 0; i < len(lang); i, cls = i+1, cls+1 { - for i < len(lang) && isspace(lang[i]) { - i++ - } - - if i < len(lang) { - org := i - for i < len(lang) && !isspace(lang[i]) { - i++ - } - - if lang[org] == '.' { - org++ - } + doubleSpace(out) - if cls > 0 { - out.WriteByte(' ') - } - attrEscape(out, []byte(lang[org:])) - } + // parse out the language names/classes + count := 0 + for _, elt := range strings.Fields(lang) { + if elt[0] == '.' { + elt = elt[1:] } - - out.WriteString("\">") - } else { - out.WriteString("<pre><code>") + if len(elt) == 0 { + continue + } + if count == 0 { + out.WriteString("<pre><code class=\"") + } else { + out.WriteByte(' ') + } + attrEscape(out, []byte(elt)) + count++ } - if len(text) > 0 { - attrEscape(out, text) + if count == 0 { + out.WriteString("<pre><code>") + } else { + out.WriteString("\">") } + attrEscape(out, text) out.WriteString("</code></pre>\n") }@@ -252,33 +224,29 @@ * E.g.
* ~~~~ {.python .numbered} => <pre lang="python"><code> */ func (options *Html) BlockCodeGithub(out *bytes.Buffer, text []byte, lang string) { - if out.Len() > 0 { - out.WriteByte('\n') - } - - if len(lang) > 0 { - out.WriteString("<pre lang=\"") + doubleSpace(out) - i := 0 - for i < len(lang) && !isspace(lang[i]) { - i++ + // parse out the language name + count := 0 + for _, elt := range strings.Fields(lang) { + if elt[0] == '.' { + elt = elt[1:] } - - if lang[0] == '.' { - attrEscape(out, []byte(lang[1:i])) - } else { - attrEscape(out, []byte(lang[:i])) + if len(elt) == 0 { + continue } - + out.WriteString("<pre lang=\"") + attrEscape(out, []byte(elt)) out.WriteString("\"><code>") - } else { - out.WriteString("<pre><code>") + count++ + break } - if len(text) > 0 { - attrEscape(out, text) + if count == 0 { + out.WriteString("<pre><code>") } + attrEscape(out, text) out.WriteString("</code></pre>\n") }@@ -290,29 +258,23 @@ out.WriteString("</blockquote>")
} func (options *Html) Table(out *bytes.Buffer, header []byte, body []byte, columnData []int) { - if out.Len() > 0 { - out.WriteByte('\n') - } - out.WriteString("<table><thead>\n") + doubleSpace(out) + out.WriteString("<table>\n<thead>\n") out.Write(header) - out.WriteString("\n</thead><tbody>\n") + out.WriteString("\n</thead>\n<tbody>\n") out.Write(body) - out.WriteString("\n</tbody></table>") + out.WriteString("\n</tbody>\n</table>") } func (options *Html) TableRow(out *bytes.Buffer, text []byte) { - if out.Len() > 0 { - out.WriteByte('\n') - } + doubleSpace(out) out.WriteString("<tr>\n") out.Write(text) out.WriteString("\n</tr>") } func (options *Html) TableCell(out *bytes.Buffer, text []byte, align int) { - if out.Len() > 0 { - out.WriteByte('\n') - } + doubleSpace(out) switch align { case TABLE_ALIGNMENT_LEFT: out.WriteString("<td align=\"left\">")@@ -330,10 +292,8 @@ }
func (options *Html) List(out *bytes.Buffer, text func() bool, flags int) { marker := out.Len() + doubleSpace(out) - if marker > 0 { - out.WriteByte('\n') - } if flags&LIST_TYPE_ORDERED != 0 { out.WriteString("<ol>\n") } else {@@ -352,19 +312,13 @@ }
func (options *Html) ListItem(out *bytes.Buffer, text []byte, flags int) { out.WriteString("<li>") - size := len(text) - for size > 0 && text[size-1] == '\n' { - size-- - } - out.Write(text[:size]) + out.Write(text) out.WriteString("</li>\n") } func (options *Html) Paragraph(out *bytes.Buffer, text func() bool) { marker := out.Len() - if marker > 0 { - out.WriteByte('\n') - } + doubleSpace(out) out.WriteString("<p>") if !text() {@@ -375,10 +329,11 @@ out.WriteString("</p>\n")
} func (options *Html) AutoLink(out *bytes.Buffer, link []byte, kind int) { - if len(link) == 0 { - return - } if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) && kind != LINK_TYPE_EMAIL { + // mark it but don't link it if it is not a safe link: no smartypants + out.WriteString("<tt>") + attrEscape(out, link) + out.WriteString("</tt>") return }@@ -389,16 +344,14 @@ }
attrEscape(out, link) out.WriteString("\">") - /* - * Pretty print: if we get an email address as - * an actual URI, e.g. `mailto:foo@bar.com`, we don't - * want to print the `mailto:` prefix - */ + // Pretty print: if we get an email address as + // an actual URI, e.g. `mailto:foo@bar.com`, we don't + // want to print the `mailto:` prefix switch { case bytes.HasPrefix(link, []byte("mailto://")): - attrEscape(out, link[9:]) + attrEscape(out, link[len("mailto://"):]) case bytes.HasPrefix(link, []byte("mailto:")): - attrEscape(out, link[7:]) + attrEscape(out, link[len("mailto:"):]) default: attrEscape(out, link) }@@ -413,9 +366,6 @@ out.WriteString("</code>")
} func (options *Html) DoubleEmphasis(out *bytes.Buffer, text []byte) { - if len(text) == 0 { - return - } out.WriteString("<strong>") out.Write(text) out.WriteString("</strong>")@@ -435,9 +385,6 @@ if options.flags&HTML_SKIP_IMAGES != 0 {
return } - if len(link) == 0 { - return - } out.WriteString("<img src=\"") attrEscape(out, link) out.WriteString("\" alt=\"")@@ -461,10 +408,18 @@ }
func (options *Html) Link(out *bytes.Buffer, link []byte, title []byte, content []byte) { if options.flags&HTML_SKIP_LINKS != 0 { + // write the link text out but don't link it, just mark it with typewriter font + out.WriteString("<tt>") + attrEscape(out, content) + out.WriteString("</tt>") return } if options.flags&HTML_SAFELINK != 0 && !isSafeLink(link) { + // write the link text out but don't link it, just mark it with typewriter font + out.WriteString("<tt>") + attrEscape(out, content) + out.WriteString("</tt>") return }@@ -497,18 +452,12 @@ out.Write(text)
} func (options *Html) TripleEmphasis(out *bytes.Buffer, text []byte) { - if len(text) == 0 { - return - } out.WriteString("<strong><em>") out.Write(text) out.WriteString("</em></strong>") } func (options *Html) StrikeThrough(out *bytes.Buffer, text []byte) { - if len(text) == 0 { - return - } out.WriteString("<del>") out.Write(text) out.WriteString("</del>")@@ -718,3 +667,9 @@ }
return isspace(tag[i]) || tag[i] == '>' } + +func doubleSpace(out *bytes.Buffer) { + if out.Len() > 0 { + out.WriteByte('\n') + } +}
M
inline.go
→
inline.go
@@ -139,7 +139,9 @@ fEnd--
} // render the code span - parser.r.CodeSpan(out, data[fBegin:fEnd]) + if fBegin != fEnd { + parser.r.CodeSpan(out, data[fBegin:fEnd]) + } return end@@ -409,7 +411,7 @@ uLink = uLinkBuf.Bytes()
} // links need something to click on and somewhere to go - if len(uLink) == 0 || content.Len() == 0 { + if len(uLink) == 0 || (!isImg && content.Len() == 0) { return 0 }@@ -439,7 +441,9 @@ if end > 2 {
if altype != LINK_TYPE_NOT_AUTOLINK { var uLink bytes.Buffer unescapeText(&uLink, data[1:end+1-2]) - parser.r.AutoLink(out, uLink.Bytes(), altype) + if uLink.Len() > 0 { + parser.r.AutoLink(out, uLink.Bytes(), altype) + } } else { parser.r.RawHtmlTag(out, data[:end]) }@@ -611,7 +615,9 @@
var uLink bytes.Buffer unescapeText(&uLink, data[:linkEnd]) - parser.r.AutoLink(out, uLink.Bytes(), LINK_TYPE_NORMAL) + if uLink.Len() > 0 { + parser.r.AutoLink(out, uLink.Bytes(), LINK_TYPE_NORMAL) + } return linkEnd - rewind }@@ -879,11 +885,13 @@ if i+1 < len(data) && data[i] == c && data[i+1] == c && i > 0 && !isspace(data[i-1]) {
var work bytes.Buffer parser.parseInline(&work, data[:i]) - // pick the right renderer - if c == '~' { - parser.r.StrikeThrough(out, work.Bytes()) - } else { - parser.r.DoubleEmphasis(out, work.Bytes()) + if work.Len() > 0 { + // pick the right renderer + if c == '~' { + parser.r.StrikeThrough(out, work.Bytes()) + } else { + parser.r.DoubleEmphasis(out, work.Bytes()) + } } return i + 2 }@@ -915,7 +923,9 @@ // triple symbol found
var work bytes.Buffer parser.parseInline(&work, data[:i]) - parser.r.TripleEmphasis(out, work.Bytes()) + if work.Len() > 0 { + parser.r.TripleEmphasis(out, work.Bytes()) + } return i + 3 case (i+1 < len(data) && data[i+1] == c): // double symbol found, hand over to emph1
M
inline_test.go
→
inline_test.go
@@ -224,7 +224,7 @@ "a single multi-tick marker with ``` no text\n",
"<p>a single multi-tick marker with ``` no text</p>\n", "markers with ` ` a space\n", - "<p>markers with <code></code> a space</p>\n", + "<p>markers with a space</p>\n", "`source code` and a `stray\n", "<p><code>source code</code> and a `stray</p>\n",