Attachment 507324 Details for Bug 355178

[patch] Patch (v1)

355178.patch (text/plain), 11.44 KB, created by (no longer active)

(hide)

Description:

Filename:

MIME Type:

Creator: (no longer active)

Size: 11.44 KB

patch

obsolete

># HG changeset patch
># Parent c7f5a1af072b58299a9fa13c35e36e2968d107ce
># User Ehsan Akhgari <ehsan@mozilla.com>
>Bug 355178 - Correct the tokenization of spell checker so that word boundary rules can be defined using the affix file
>
>
>diff --git a/extensions/spellcheck/locales/en-US/hunspell/dictionary-sources/upstream-hunspell.diff b/extensions/spellcheck/locales/en-US/hunspell/dictionary-sources/upstream-hunspell.diff
>--- a/extensions/spellcheck/locales/en-US/hunspell/dictionary-sources/upstream-hunspell.diff
>+++ b/extensions/spellcheck/locales/en-US/hunspell/dictionary-sources/upstream-hunspell.diff
>@@ -9458,106 +9458,108 @@ 41415a47226
> 41914c47725
> < schnaps's
> ---
> > schnaps/M
> 41949c47760
> < schrod's
> ---
> > schrod/SM
>-42883,42885c48694
>+41998a42010
>+> scot-free
>+42883,42885c48695
> < shit's
> < shit/S!
> < shite/S!
> ---
> > shit/MS!
>-42887,42888c48696,48697
>+42887,42888c48697,48698
> < shithead/S!
> < shitload/!
> ---
> > shithead/MS!
> > shitload/MS!
>-42891c48700
>+42891c48701
> < shitty/RT!
> ---
> > shitty/TR!
>-42976a48786
>+42976a48787
> > should've
>-43008c48818
>+43008c48819
> < showtime
> ---
> > showtime/MS
>-43724,43726c49534
>+43724,43726c49535
> < smoulder's
> < smouldered
> < smoulders
> ---
> > smoulder/GSMD
>-44062c49870
>+44062c49871
> < sonofabitch
> ---
> > sonofabitch/!
>-44371a50180
>+44371a50181
> > spick/S!
>-44383c50192
>+44383c50193
> < spik/S
> ---
> > spik/S!
>-46106a51916
>+46106a51917
> > syllabi
>-46160c51970
>+46160c51971
> < synch/GMD
> ---
> > synch/GMDS
>-46167d51976
>+46167d51977
> < synchs
>-46203,46204c52012,52013
>+46203,46204c52013,52014
> < sysadmin/S
> < sysop/S
> ---
> > sysadmin/MS
> > sysop/MS
>-46752a52562
>+46752a52563
> > terabit/MS
>-46753a52564,52565
>+46753a52565,52566
> > terahertz/M
> > terapixel/MS
>-46817a52630
>+46817a52631
> > testcase/MS
>-46831a52645
>+46831a52646
> > testsuite/MS
>-46925a52740
>+46925a52741
> > theremin/MS
>-47755a53571
>+47755a53572
> > transfect/DSMG
>-47774a53591,53592
>+47774a53592,53593
> > transgenderism
> > transgene/MS
>-47951c53769
>+47951c53770
> < triage/M
> ---
> > triage/MG
>-48869a54688
>+48869a54689
> > unlikeable
>-49211c55030
>+49211c55031
> < vagina/M
> ---
> > vagina/MS
>-49368,49369c55187
>+49368,49369c55188
> < velour's
> < velours's
> ---
> > velour/MS
>-49478a55297
>+49478a55298
> > vertices
>-50148a55968
>+50148a55969
> > weaponize/DSG
>-50260,50261d56079
>+50260,50261d56080
> < werwolf/M
> < werwolves
>-50728c56546
>+50728c56547
> < women
> ---
> > women/M
>-50794c56612
>+50794c56613
> < wop/S!
> ---
> > wop/MS!
>diff --git a/extensions/spellcheck/locales/en-US/hunspell/en-US.aff b/extensions/spellcheck/locales/en-US/hunspell/en-US.aff
>--- a/extensions/spellcheck/locales/en-US/hunspell/en-US.aff
>+++ b/extensions/spellcheck/locales/en-US/hunspell/en-US.aff
>@@ -7,17 +7,52 @@ COMPOUNDMIN 1
> # only in compounds: 1th, 2th, 3th
> ONLYINCOMPOUND c
> # compound rules:
> # 1. [0-9]*1[0-9]th (10th, 11th, 12th, 56714th, etc.)
> # 2. [0-9]*[02-9](1st|2nd|3rd|[4-9]th) (21st, 22nd, 123rd, 1234th, etc.)
> COMPOUNDRULE 2
> COMPOUNDRULE n*1t
> COMPOUNDRULE n*mp
>-WORDCHARS 0123456789
>+WORDCHARS 0123456789'
>+
>+BREAK 33
>+BREAK `
>+BREAK "
>+BREAK [
>+BREAK ]
>+BREAK (
>+BREAK )
>+BREAK {
>+BREAK }
>+BREAK <
>+BREAK >
>+BREAK :
>+BREAK ,
>+BREAK !
>+BREAK .
>+BREAK ?
>+BREAK /
>+BREAK ;
>+BREAK -
>+BREAK ^-
>+BREAK -$
>+BREAK @
>+BREAK ~
>+BREAK `
>+BREAK #
>+BREAK $
>+BREAK %
>+BREAK ^
>+BREAK &
>+BREAK *
>+BREAK =
>+BREAK +
>+BREAK |
>+BREAK \
> 
> PFX A Y 1
> PFX A   0     re         .
> 
> PFX I Y 1
> PFX I   0     in         .
> 
> PFX U Y 1
>diff --git a/extensions/spellcheck/locales/en-US/hunspell/en-US.dic b/extensions/spellcheck/locales/en-US/hunspell/en-US.dic
>--- a/extensions/spellcheck/locales/en-US/hunspell/en-US.dic
>+++ b/extensions/spellcheck/locales/en-US/hunspell/en-US.dic
>@@ -1,9 +1,9 @@
>-57434
>+57435
> 0/nm
> 0th/pt
> 1/n1
> 1st/p
> 1th/tc
> 2/nm
> 2nd/p
> 2th/tc
>@@ -48136,16 +48136,17 @@ scorecard/MS
> scorekeeper/MS
> scoreless
> scoreline/S
> scorer/M
> scorn/MDRSZG
> scorner/M
> scornful/Y
> scorpion/MS
>+scot-free
> scotch/MDSG
> scotchs
> scoundrel/MS
> scour/DRSZG
> scourer/M
> scourge/DSMG
> scout/MDRSZG
> scouting/M
>diff --git a/extensions/spellcheck/src/mozInlineSpellWordUtil.cpp b/extensions/spellcheck/src/mozInlineSpellWordUtil.cpp
>--- a/extensions/spellcheck/src/mozInlineSpellWordUtil.cpp
>+++ b/extensions/spellcheck/src/mozInlineSpellWordUtil.cpp
>@@ -894,22 +894,64 @@ WordSplitState::ClassifyCharacter(PRInt3
>       return CHAR_CLASS_SEPARATOR;
>     if (ClassifyCharacter(aIndex + 1, false) != CHAR_CLASS_WORD)
>       return CHAR_CLASS_SEPARATOR;
> 
>     // char on either side is a word, this counts as a word
>     return CHAR_CLASS_WORD;
>   }
> 
>+  // The dot character, if appearing at the end of a word, should
>+  // be considered part of that word.  Example: "etc.", or
>+  // abbreviations
>+  if (aIndex > 0 &&
>+      mDOMWordText[aIndex] == '.' &&
>+      mDOMWordText[aIndex - 1] != '.' &&
>+      ClassifyCharacter(aIndex - 1, false) != CHAR_CLASS_WORD) {
>+    return CHAR_CLASS_WORD;
>+  }
>+
>   // all other punctuation
>   if (charCategory == nsIUGenCategory::kSeparator ||
>       charCategory == nsIUGenCategory::kOther ||
>       charCategory == nsIUGenCategory::kPunctuation ||
>-      charCategory == nsIUGenCategory::kSymbol)
>-    return CHAR_CLASS_SEPARATOR;
>+      charCategory == nsIUGenCategory::kSymbol) {
>+    PRBool start = PR_TRUE, end = PR_TRUE;
>+
>+    PRInt32 tmpIndex = aIndex - 1;
>+    if (tmpIndex >= 0) {
>+      nsIUGenCategory::nsUGenCategory
>+        charCategoryTmp = mWordUtil->GetCategories()->Get(PRUint32(mDOMWordText[tmpIndex]));
>+      if (charCategoryTmp != nsIUGenCategory::kSeparator &&
>+          charCategoryTmp != nsIUGenCategory::kOther &&
>+          charCategoryTmp != nsIUGenCategory::kPunctuation &&
>+          charCategoryTmp != nsIUGenCategory::kSymbol) {
>+        start = PR_FALSE;
>+      }
>+    }
>+    if (start) {
>+      return CHAR_CLASS_SEPARATOR;
>+    }
>+
>+    tmpIndex = aIndex + 1;
>+    if (tmpIndex < PRInt32(mDOMWordText.Length())) {
>+      nsIUGenCategory::nsUGenCategory
>+        charCategoryTmp = mWordUtil->GetCategories()->Get(PRUint32(mDOMWordText[tmpIndex]));
>+      if (charCategoryTmp != nsIUGenCategory::kSeparator &&
>+          charCategoryTmp != nsIUGenCategory::kOther &&
>+          charCategoryTmp != nsIUGenCategory::kPunctuation &&
>+          charCategoryTmp != nsIUGenCategory::kSymbol) {
>+        end = PR_FALSE;
>+      }
>+    }
>+
>+    if (end) {
>+      return CHAR_CLASS_SEPARATOR;
>+    }
>+  }
> 
>   // any other character counts as a word
>   return CHAR_CLASS_WORD;
> }
> 
> 
> // WordSplitState::Advance
> 
>diff --git a/layout/reftests/editor/reftest.list b/layout/reftests/editor/reftest.list
>--- a/layout/reftests/editor/reftest.list
>+++ b/layout/reftests/editor/reftest.list
>@@ -43,8 +43,14 @@ asserts(1) == passwd-4.html passwd-ref.h
> == caret_on_focus.html caret_on_focus-ref.html
> != caret_on_textarea_lastline.html caret_on_textarea_lastline-ref.html
> == input-text-onfocus-reframe.html input-text-onfocus-reframe-ref.html
> == input-text-notheme-onfocus-reframe.html input-text-notheme-onfocus-reframe-ref.html
> == caret_after_reframe.html caret_after_reframe-ref.html
> == nobogusnode-1.html nobogusnode-ref.html
> == nobogusnode-2.html nobogusnode-ref.html
> == caret_on_input_edge.html caret_on_input_edge-ref.html
>+== spellcheck-hyphen-valid.html spellcheck-hyphen-valid-ref.html
>+!= spellcheck-hyphen-invalid.html spellcheck-hyphen-invalid-ref.html
>+== spellcheck-slash-valid.html spellcheck-slash-valid-ref.html
>+== spellcheck-period-valid.html spellcheck-period-valid-ref.html
>+== spellcheck-space-valid.html spellcheck-space-valid-ref.html
>+== spellcheck-comma-valid.html spellcheck-comma-valid-ref.html
>diff --git a/layout/reftests/editor/spellcheck-comma-valid-ref.html b/layout/reftests/editor/spellcheck-comma-valid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-comma-valid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">good,nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-comma-valid.html b/layout/reftests/editor/spellcheck-comma-valid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-comma-valid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>good,nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-hyphen-invalid-ref.html b/layout/reftests/editor/spellcheck-hyphen-invalid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-hyphen-invalid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">dddf-gggy</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-hyphen-invalid.html b/layout/reftests/editor/spellcheck-hyphen-invalid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-hyphen-invalid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>dddf-gggy</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-hyphen-valid-ref.html b/layout/reftests/editor/spellcheck-hyphen-valid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-hyphen-valid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">scot-free</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-hyphen-valid.html b/layout/reftests/editor/spellcheck-hyphen-valid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-hyphen-valid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>scot-free</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-period-valid-ref.html b/layout/reftests/editor/spellcheck-period-valid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-period-valid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">good.nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-period-valid.html b/layout/reftests/editor/spellcheck-period-valid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-period-valid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>good.nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-slash-valid-ref.html b/layout/reftests/editor/spellcheck-slash-valid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-slash-valid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">good/nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-slash-valid.html b/layout/reftests/editor/spellcheck-slash-valid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-slash-valid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>good/nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-space-valid-ref.html b/layout/reftests/editor/spellcheck-space-valid-ref.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-space-valid-ref.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus spellcheck="false">good nice</textarea>
>+  </body>
>+</html>
>diff --git a/layout/reftests/editor/spellcheck-space-valid.html b/layout/reftests/editor/spellcheck-space-valid.html
>new file mode 100644
>--- /dev/null
>+++ b/layout/reftests/editor/spellcheck-space-valid.html
>@@ -0,0 +1,6 @@
>+<!DOCTYPE html>
>+<html>
>+  <body>
>+    <textarea autofocus>good nice</textarea>
>+  </body>
>+</html>

Flags:

smaug: review+
ehsan.akhgari: approval2.0-

Actions: View | Diff | Review

Attachments on bug 355178: 418524 | 418548 | 418549 | 418984 | 419239 | 431603 | 507324 | 507912 | 507917 | 508164 | 510837 | 526149