emelie/tidy-html5
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

Merge pull request #553 from htacg/new_picklists

Browse source 
New picklists and parsers
This commit is contained in:
Jim Derry 2017-05-13 19:50:20 -04:00 committed by GitHub
parent 66bed8b9a0 a399725a1e
commit b6bf48c24a
5 changed files with 386 additions and 458 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
README/OPTIONS.md
View file

@ -38,7 +38,7 @@ struct _tidy_option
TidyOptionType type; /* string, int or bool */
ulong dflt; /* default for TidyInteger and TidyBoolean */
ParseProperty* parser; /* parsing method, read-only if NULL */
const ctmbstr* pickList; /* pick list */
PickListItems* pickList; /* pick list */
ctmbstr pdflt; /* default for TidyString */
};
~~~
@ -78,7 +78,7 @@ Care, each of these enumeration strings have been equated to two uppercase lette
The next item is the `default` value for a boolean, tristate or integer. Note tidy set `no=0` and `yes=1` as its own `Bool` enumeration.
There are a number of `parser` for the options. Likewise a number of `pickList`. Find another option similar to your new option and use the same values.
There are a number of `parser` for the options. Likewise a number of `pickList`. Find another option similar to your new option and use the same values. The `parser` is the function that parses config file or command line text input, and the `picklist` constitutes the canonical values for the option. Some types of values logically don't have picklists, such as strings or pure integers.
Presently no options have the final `default` string, and it is left out of the table. The compiler will add a NULL.

30
include/tidyenum.h
View file

@ -725,6 +725,36 @@ typedef enum
TidyCR /**< Use Macintosh style: CR */
} TidyLineEnding;
/** TidyEncodingOptions option values specify the input and/or output encoding.
** @remark This enum's starting value is guaranteed to remain stable.
*/
typedef enum
{
TidyEncRaw = 0,
TidyEncAscii,
TidyEncLatin0,
TidyEncLatin1,
TidyEncUtf8,
#ifndef NO_NATIVE_ISO2022_SUPPORT
TidyEncIso2022,
#endif
TidyEncMac,
TidyEncWin1252,
TidyEncIbm858,
#if SUPPORT_UTF16_ENCODINGS
TidyEncUtf16le,
TidyEncUtf16be,
TidyEncUtf16,
#endif
#if SUPPORT_ASIAN_ENCODINGS
TidyEncBig5,
TidyEncShiftjis
#endif
} TidyEncodingOptions;
/** Mode controlling treatment of doctype
** @remark This enum's starting value is guaranteed to remain stable.

704
src/config.c
View file

@ -1,22 +1,12 @@
/*
config.c -- read config file and manage config properties
(c) 1998-2008 (W3C) MIT, ERCIM, Keio University
See tidy.h for the copyright notice.
config files associate a property name with a value.
// comments can start at the beginning of a line
# comments can start at the beginning of a line
name: short values fit onto one line
name: a really long value that
continues on the next line
property names are case insensitive and should be less than
60 characters in length and must start at the begining of
the line, as whitespace at the start of a line signifies a
line continuation.
*/
/* config.c
* Read configuration files and manage configuration properties.
*
* Copyright (c) 1998-2017 World Wide Web Consortium (Massachusetts
* Institute of Technology, European Research Consortium for Informatics
* and Mathematics, Keio University) and HTACG.
*
* See tidy.h for the copyright notice.
*/
#include "config.h"
#include "tidy-int.h"
@ -29,8 +19,9 @@
#else
#ifdef DMALLOC
/*
macro for valloc() in dmalloc.h may conflict with declaration for valloc() in unistd.h -
we don't need (debugging for) valloc() here. dmalloc.h should come last but it doesn't.
macro for valloc() in dmalloc.h may conflict with declaration for valloc()
in unistd.h - we don't need (debugging for) valloc() here. dmalloc.h should
come last but it doesn't.
*/
#ifdef valloc
#undef valloc
@ -56,103 +47,98 @@ void TY_(FreeConfig)( TidyDocImpl* doc )
}
/* Arrange so index can be cast to enum
/*
Arrange so index can be cast to enum. Note that the value field in the
following structures is not currently used in code; they're present for
documentation purposes currently. The arrays must be populated in enum order.
*/
static const ctmbstr boolPicks[] =
{
"no",
"yes",
NULL
static PickListItems boolPicks = {
{ "no", TidyNoState, { "0", "n", "f", "no", "false", NULL } },
{ "yes", TidyYesState, { "1", "y", "t", "yes", "true", NULL } },
{ NULL }
};
static const ctmbstr autoBoolPicks[] =
{
"no",
"yes",
"auto",
NULL
static PickListItems autoBoolPicks = {
{ "no", TidyNoState, { "0", "n", "f", "no", "false", NULL } },
{ "yes", TidyYesState, { "1", "y", "t", "yes", "true", NULL } },
{ "auto", TidyYesState, { "auto", NULL } },
{ NULL }
};
static const ctmbstr repeatAttrPicks[] =
{
"keep-first",
"keep-last",
NULL
static PickListItems repeatAttrPicks = {
{ "keep-first", TidyNoState, { "keep-first", NULL } },
{ "keep-last", TidyYesState, { "keep-last", NULL } },
{ NULL }
};
static const ctmbstr accessPicks[] =
{
"0 (Tidy Classic)",
"1 (Priority 1 Checks)",
"2 (Priority 2 Checks)",
"3 (Priority 3 Checks)",
NULL
static PickListItems accessPicks = {
{ "0 (Tidy Classic)", 0, { "0", "0 (Tidy Classic)", NULL } },
{ "1 (Priority 1 Checks)", 1, { "1", "1 (Priority 1 Checks)", NULL } },
{ "2 (Priority 2 Checks)", 2, { "2", "2 (Priority 2 Checks)", NULL } },
{ "3 (Priority 3 Checks)", 3, { "3", "3 (Priority 3 Checks)", NULL } },
{ NULL }
};
static const ctmbstr charEncPicks[] =
{
"raw",
"ascii",
"latin0",
"latin1",
"utf8",
static PickListItems charEncPicks = {
{ "raw", TidyEncRaw, { "raw", NULL } },
{ "ascii", TidyEncAscii, { "ascii", NULL } },
{ "latin0", TidyEncLatin0, { "latin0", NULL } },
{ "latin1", TidyEncLatin1, { "latin1", NULL } },
{ "utf8", TidyEncUtf8, { "utf8", NULL } },
#ifndef NO_NATIVE_ISO2022_SUPPORT
"iso2022",
{ "iso2022", TidyEncIso2022, { "iso2022", NULL } },
#endif
"mac",
"win1252",
"ibm858",
{ "mac", TidyEncMac, { "mac", NULL } },
{ "win1252", TidyEncWin1252, { "win1252", NULL } },
{ "ibm858", TidyEncIbm858, { "ibm858", NULL } },
#if SUPPORT_UTF16_ENCODINGS
"utf16le",
"utf16be",
"utf16",
{ "utf16le", TidyEncUtf16le, { "utf16le", NULL } },
{ "utf16be", TidyEncUtf16be, { "utf16be", NULL } },
{ "utf16", TidyEncUtf16, { "utf16", NULL } },
#endif
#if SUPPORT_ASIAN_ENCODINGS
"big5",
"shiftjis",
{ "big5", TidyEncBig5, { "big5", NULL } },
{ "shiftjis", TidyEncShiftjis, { "shiftjis", NULL } },
#endif
NULL
{ NULL }
};
static const ctmbstr newlinePicks[] =
{
"LF",
"CRLF",
"CR",
NULL
static PickListItems newlinePicks = {
{ "LF", TidyLF, { "lf", NULL } },
{ "CRLF", TidyCRLF, { "crlf", NULL } },
{ "CR", TidyCR, { "cr", NULL } },
{ NULL }
};
static const ctmbstr doctypePicks[] =
{
"html5",
"omit",
"auto",
"strict",
"transitional",
"user",
NULL
static PickListItems doctypePicks = {
{ "html5", TidyDoctypeHtml5, { "html5", NULL } },
{ "omit", TidyDoctypeOmit, { "omit", NULL } },
{ "auto", TidyDoctypeAuto, { "auto", NULL } },
{ "strict", TidyDoctypeStrict, { "strict", NULL } },
{ "transitional", TidyDoctypeLoose, { "loose", "transitional", NULL } },
{ "user", TidyDoctypeUser, { "user", NULL } },
{ NULL }
};
static const ctmbstr sorterPicks[] =
{
"none",
"alpha",
NULL
static PickListItems sorterPicks = {
{ "none", TidySortAttrNone, { "none", NULL } },
{ "alpha", TidySortAttrAlpha, { "alpha", NULL } },
{ NULL }
};
static const ctmbstr customTagsPicks[] =
{
"no",
"blocklevel",
"empty",
"inline",
"pre",
NULL
static PickListItems customTagsPicks = {
{"no", TidyCustomNo, { "no", "n", NULL } },
{"blocklevel", TidyCustomBlocklevel, { "blocklevel", NULL } },
{"empty", TidyCustomEmpty, { "empty", NULL } },
{"inline", TidyCustomInline, { "inline", "y", "yes", NULL } },
{"pre", TidyCustomPre, { "pre", NULL } },
{ NULL }
};
#define MU TidyMarkup
#define DG TidyDiagnostics
#define PP TidyPrettyPrint
@ -171,7 +157,7 @@ static const ctmbstr customTagsPicks[] =
/* If Accessibility checks not supported, make config setting read-only */
#if SUPPORT_ACCESSIBILITY_CHECKS
#define ParseAcc ParseInt
#define ParseAcc ParsePickList
#else
#define ParseAcc NULL
#endif
@ -181,13 +167,6 @@ static void AdjustConfig( TidyDocImpl* doc );
/* parser for integer values */
static ParseProperty ParseInt;
/* parser for 't'/'f', 'true'/'false', 'y'/'n', 'yes'/'no' or '1'/'0' */
static ParseProperty ParseBool;
/* parser for 't'/'f', 'true'/'false', 'y'/'n', 'yes'/'no', '1'/'0'
or 'auto' */
static ParseProperty ParseAutoBool;
/* a string excluding whitespace */
static ParseProperty ParseName;
@ -200,145 +179,135 @@ static ParseProperty ParseString;
/* a space or comma separated list of tag names */
static ParseProperty ParseTagNames;
/* alpha */
static ParseProperty ParseSorter;
/* RAW, ASCII, LATIN0, LATIN1, UTF8, ISO2022, MACROMAN,
/* RAW, ASCII, LATIN0, LATIN1, UTF8, ISO2022, MACROMAN,
WIN1252, IBM858, UTF16LE, UTF16BE, UTF16, BIG5, SHIFTJIS
*/
static ParseProperty ParseCharEnc;
static ParseProperty ParseNewline;
/* html5 | omit | auto | strict | loose | <fpi> */
static ParseProperty ParseDocType;
/* keep-first or keep-last? */
static ParseProperty ParseRepeatAttr;
/*\
* 20150515 - support using tabs instead of spaces - Issue #108
* (a) parser for 't'/'f', 'true'/'false', 'y'/'n', 'yes'/'no' or '1'/'0'
* (b) sets the TidyIndentSpaces to 1 if 'yes'
* (c) sets the indent_char to '\t' or ' '
\*/
/* 20150515 - support using tabs instead of spaces - Issue #108
*/
static ParseProperty ParseTabs;
/* Parse the value of TidyUseCustomTags */
static ParseProperty ParseUseCustomTags;
/* General parser for options having picklists */
static ParseProperty ParsePickList;
/* Ensure struct order is same order as tidyenum.h:TidyOptionId! */
static const TidyOptionImpl option_defs[] =
{
{ TidyUnknownOption, MS, "unknown!", IN, 0, NULL, NULL },
{ TidyAccessibilityCheckLevel, DG, "accessibility-check", IN, 0, ParseAcc, accessPicks },
{ TidyAltText, MU, "alt-text", ST, 0, ParseString, NULL },
{ TidyAnchorAsName, MU, "anchor-as-name", BL, yes, ParseBool, boolPicks },
{ TidyAsciiChars, CE, "ascii-chars", BL, no, ParseBool, boolPicks },
{ TidyBlockTags, MU, "new-blocklevel-tags", ST, 0, ParseTagNames, NULL },
{ TidyBodyOnly, MU, "show-body-only", IN, no, ParseAutoBool, autoBoolPicks },
{ TidyBreakBeforeBR, PP, "break-before-br", BL, no, ParseBool, boolPicks },
{ TidyCharEncoding, CE, "char-encoding", IN, UTF8, ParseCharEnc, charEncPicks },
{ TidyCoerceEndTags, MU, "coerce-endtags", BL, yes, ParseBool, boolPicks },
{ TidyCSSPrefix, MU, "css-prefix", ST, 0, ParseCSS1Selector, NULL },
{ TidyCustomTags, IR, "new-custom-tags", ST, 0, ParseTagNames, NULL }, /* 20170309 - Issue #119 */
{ TidyDecorateInferredUL, MU, "decorate-inferred-ul", BL, no, ParseBool, boolPicks },
{ TidyDoctype, MU, "doctype", ST, 0, ParseDocType, doctypePicks },
{ TidyUnknownOption, MS, "unknown!", IN, 0, NULL, NULL },
{ TidyAccessibilityCheckLevel, DG, "accessibility-check", IN, 0, ParseAcc, &accessPicks },
{ TidyAltText, MU, "alt-text", ST, 0, ParseString, NULL },
{ TidyAnchorAsName, MU, "anchor-as-name", BL, yes, ParsePickList, &boolPicks },
{ TidyAsciiChars, CE, "ascii-chars", BL, no, ParsePickList, &boolPicks },
{ TidyBlockTags, MU, "new-blocklevel-tags", ST, 0, ParseTagNames, NULL },
{ TidyBodyOnly, MU, "show-body-only", IN, no, ParsePickList, &autoBoolPicks },
{ TidyBreakBeforeBR, PP, "break-before-br", BL, no, ParsePickList, &boolPicks },
{ TidyCharEncoding, CE, "char-encoding", IN, UTF8, ParseCharEnc, &charEncPicks },
{ TidyCoerceEndTags, MU, "coerce-endtags", BL, yes, ParsePickList, &boolPicks },
{ TidyCSSPrefix, MU, "css-prefix", ST, 0, ParseCSS1Selector, NULL },
{ TidyCustomTags, IR, "new-custom-tags", ST, 0, ParseTagNames, NULL }, /* 20170309 - Issue #119 */
{ TidyDecorateInferredUL, MU, "decorate-inferred-ul", BL, no, ParsePickList, &boolPicks },
{ TidyDoctype, MU, "doctype", ST, 0, ParseDocType, &doctypePicks },
#ifndef DOXYGEN_SHOULD_SKIP_THIS
{ TidyDoctypeMode, IR, "doctype-mode", IN, TidyDoctypeAuto, NULL, doctypePicks },
{ TidyDoctypeMode, IR, "doctype-mode", IN, TidyDoctypeAuto, NULL, &doctypePicks },
#endif
{ TidyDropEmptyElems, MU, "drop-empty-elements", BL, yes, ParseBool, boolPicks },
{ TidyDropEmptyParas, MU, "drop-empty-paras", BL, yes, ParseBool, boolPicks },
{ TidyDropPropAttrs, MU, "drop-proprietary-attributes", BL, no, ParseBool, boolPicks },
{ TidyDuplicateAttrs, MU, "repeated-attributes", IN, TidyKeepLast, ParseRepeatAttr, repeatAttrPicks },
{ TidyEmacs, MS, "gnu-emacs", BL, no, ParseBool, boolPicks },
{ TidyDropEmptyElems, MU, "drop-empty-elements", BL, yes, ParsePickList, &boolPicks },
{ TidyDropEmptyParas, MU, "drop-empty-paras", BL, yes, ParsePickList, &boolPicks },
{ TidyDropPropAttrs, MU, "drop-proprietary-attributes", BL, no, ParsePickList, &boolPicks },
{ TidyDuplicateAttrs, MU, "repeated-attributes", IN, TidyKeepLast, ParsePickList, &repeatAttrPicks },
{ TidyEmacs, MS, "gnu-emacs", BL, no, ParsePickList, &boolPicks },
#ifndef DOXYGEN_SHOULD_SKIP_THIS
{ TidyEmacsFile, IR, "gnu-emacs-file", ST, 0, ParseString, NULL },
{ TidyEmacsFile, IR, "gnu-emacs-file", ST, 0, ParseString, NULL },
#endif
{ TidyEmptyTags, MU, "new-empty-tags", ST, 0, ParseTagNames, NULL },
{ TidyEncloseBlockText, MU, "enclose-block-text", BL, no, ParseBool, boolPicks },
{ TidyEncloseBodyText, MU, "enclose-text", BL, no, ParseBool, boolPicks },
{ TidyErrFile, MS, "error-file", ST, 0, ParseString, NULL },
{ TidyEscapeCdata, MU, "escape-cdata", BL, no, ParseBool, boolPicks },
{ TidyEscapeScripts, PP, "escape-scripts", BL, yes, ParseBool, boolPicks }, /* 20160227 - Issue #348 */
{ TidyFixBackslash, MU, "fix-backslash", BL, yes, ParseBool, boolPicks },
{ TidyFixComments, MU, "fix-bad-comments", BL, yes, ParseBool, boolPicks },
{ TidyFixUri, MU, "fix-uri", BL, yes, ParseBool, boolPicks },
{ TidyForceOutput, MS, "force-output", BL, no, ParseBool, boolPicks },
{ TidyGDocClean, MU, "gdoc", BL, no, ParseBool, boolPicks },
{ TidyHideComments, MU, "hide-comments", BL, no, ParseBool, boolPicks },
{ TidyHtmlOut, MU, "output-html", BL, no, ParseBool, boolPicks },
{ TidyInCharEncoding, CE, "input-encoding", IN, UTF8, ParseCharEnc, charEncPicks },
{ TidyIndentAttributes, PP, "indent-attributes", BL, no, ParseBool, boolPicks },
{ TidyIndentCdata, MU, "indent-cdata", BL, no, ParseBool, boolPicks },
{ TidyIndentContent, PP, "indent", IN, TidyNoState, ParseAutoBool, autoBoolPicks },
{ TidyIndentSpaces, PP, "indent-spaces", IN, 2, ParseInt, NULL },
{ TidyInlineTags, MU, "new-inline-tags", ST, 0, ParseTagNames, NULL },
{ TidyJoinClasses, MU, "join-classes", BL, no, ParseBool, boolPicks },
{ TidyJoinStyles, MU, "join-styles", BL, yes, ParseBool, boolPicks },
{ TidyKeepFileTimes, MS, "keep-time", BL, no, ParseBool, boolPicks },
{ TidyLiteralAttribs, MU, "literal-attributes", BL, no, ParseBool, boolPicks },
{ TidyLogicalEmphasis, MU, "logical-emphasis", BL, no, ParseBool, boolPicks },
{ TidyLowerLiterals, MU, "lower-literals", BL, yes, ParseBool, boolPicks },
{ TidyMakeBare, MU, "bare", BL, no, ParseBool, boolPicks },
{ TidyMakeClean, MU, "clean", BL, no, ParseBool, boolPicks },
{ TidyMark, MS, "tidy-mark", BL, yes, ParseBool, boolPicks },
{ TidyMergeDivs, MU, "merge-divs", IN, TidyAutoState, ParseAutoBool, autoBoolPicks },
{ TidyMergeEmphasis, MU, "merge-emphasis", BL, yes, ParseBool, boolPicks },
{ TidyMergeSpans, MU, "merge-spans", IN, TidyAutoState, ParseAutoBool, autoBoolPicks },
{ TidyEmptyTags, MU, "new-empty-tags", ST, 0, ParseTagNames, NULL },
{ TidyEncloseBlockText, MU, "enclose-block-text", BL, no, ParsePickList, &boolPicks },
{ TidyEncloseBodyText, MU, "enclose-text", BL, no, ParsePickList, &boolPicks },
{ TidyErrFile, MS, "error-file", ST, 0, ParseString, NULL },
{ TidyEscapeCdata, MU, "escape-cdata", BL, no, ParsePickList, &boolPicks },
{ TidyEscapeScripts, PP, "escape-scripts", BL, yes, ParsePickList, &boolPicks }, /* 20160227 - Issue #348 */
{ TidyFixBackslash, MU, "fix-backslash", BL, yes, ParsePickList, &boolPicks },
{ TidyFixComments, MU, "fix-bad-comments", BL, yes, ParsePickList, &boolPicks },
{ TidyFixUri, MU, "fix-uri", BL, yes, ParsePickList, &boolPicks },
{ TidyForceOutput, MS, "force-output", BL, no, ParsePickList, &boolPicks },
{ TidyGDocClean, MU, "gdoc", BL, no, ParsePickList, &boolPicks },
{ TidyHideComments, MU, "hide-comments", BL, no, ParsePickList, &boolPicks },
{ TidyHtmlOut, MU, "output-html", BL, no, ParsePickList, &boolPicks },
{ TidyInCharEncoding, CE, "input-encoding", IN, UTF8, ParseCharEnc, &charEncPicks },
{ TidyIndentAttributes, PP, "indent-attributes", BL, no, ParsePickList, &boolPicks },
{ TidyIndentCdata, MU, "indent-cdata", BL, no, ParsePickList, &boolPicks },
{ TidyIndentContent, PP, "indent", IN, TidyNoState, ParsePickList, &autoBoolPicks },
{ TidyIndentSpaces, PP, "indent-spaces", IN, 2, ParseInt, NULL },
{ TidyInlineTags, MU, "new-inline-tags", ST, 0, ParseTagNames, NULL },
{ TidyJoinClasses, MU, "join-classes", BL, no, ParsePickList, &boolPicks },
{ TidyJoinStyles, MU, "join-styles", BL, yes, ParsePickList, &boolPicks },
{ TidyKeepFileTimes, MS, "keep-time", BL, no, ParsePickList, &boolPicks },
{ TidyLiteralAttribs, MU, "literal-attributes", BL, no, ParsePickList, &boolPicks },
{ TidyLogicalEmphasis, MU, "logical-emphasis", BL, no, ParsePickList, &boolPicks },
{ TidyLowerLiterals, MU, "lower-literals", BL, yes, ParsePickList, &boolPicks },
{ TidyMakeBare, MU, "bare", BL, no, ParsePickList, &boolPicks },
{ TidyMakeClean, MU, "clean", BL, no, ParsePickList, &boolPicks },
{ TidyMark, MS, "tidy-mark", BL, yes, ParsePickList, &boolPicks },
{ TidyMergeDivs, MU, "merge-divs", IN, TidyAutoState, ParsePickList, &autoBoolPicks },
{ TidyMergeEmphasis, MU, "merge-emphasis", BL, yes, ParsePickList, &boolPicks },
{ TidyMergeSpans, MU, "merge-spans", IN, TidyAutoState, ParsePickList, &autoBoolPicks },
#if SUPPORT_ASIAN_ENCODINGS
{ TidyNCR, MU, "ncr", BL, yes, ParseBool, boolPicks },
{ TidyNCR, MU, "ncr", BL, yes, ParsePickList, &boolPicks },
#endif
{ TidyNewline, CE, "newline", IN, DLF, ParseNewline, newlinePicks },
{ TidyNumEntities, MU, "numeric-entities", BL, no, ParseBool, boolPicks },
{ TidyOmitOptionalTags, MU, "omit-optional-tags", BL, no, ParseBool, boolPicks },
{ TidyOutCharEncoding, CE, "output-encoding", IN, UTF8, ParseCharEnc, charEncPicks },
{ TidyOutFile, MS, "output-file", ST, 0, ParseString, NULL },
{ TidyNewline, CE, "newline", IN, DLF, ParsePickList, &newlinePicks },
{ TidyNumEntities, MU, "numeric-entities", BL, no, ParsePickList, &boolPicks },
{ TidyOmitOptionalTags, MU, "omit-optional-tags", BL, no, ParsePickList, &boolPicks },
{ TidyOutCharEncoding, CE, "output-encoding", IN, UTF8, ParseCharEnc, &charEncPicks },
{ TidyOutFile, MS, "output-file", ST, 0, ParseString, NULL },
#if SUPPORT_UTF16_ENCODINGS
{ TidyOutputBOM, CE, "output-bom", IN, TidyAutoState, ParseAutoBool, autoBoolPicks },
{ TidyOutputBOM, CE, "output-bom", IN, TidyAutoState, ParsePickList, &autoBoolPicks },
#endif
{ TidyPPrintTabs, PP, "indent-with-tabs", BL, no, ParseTabs, boolPicks }, /* 20150515 - Issue #108 */
{ TidyPreserveEntities, MU, "preserve-entities", BL, no, ParseBool, boolPicks },
{ TidyPreTags, MU, "new-pre-tags", ST, 0, ParseTagNames, NULL },
{ TidyPPrintTabs, PP, "indent-with-tabs", BL, no, ParseTabs, &boolPicks }, /* 20150515 - Issue #108 */
{ TidyPreserveEntities, MU, "preserve-entities", BL, no, ParsePickList, &boolPicks },
{ TidyPreTags, MU, "new-pre-tags", ST, 0, ParseTagNames, NULL },
#if SUPPORT_ASIAN_ENCODINGS
{ TidyPunctWrap, PP, "punctuation-wrap", BL, no, ParseBool, boolPicks },
{ TidyPunctWrap, PP, "punctuation-wrap", BL, no, ParsePickList, &boolPicks },
#endif
{ TidyQuiet, MS, "quiet", BL, no, ParseBool, boolPicks },
{ TidyQuoteAmpersand, MU, "quote-ampersand", BL, yes, ParseBool, boolPicks },
{ TidyQuoteMarks, MU, "quote-marks", BL, no, ParseBool, boolPicks },
{ TidyQuoteNbsp, MU, "quote-nbsp", BL, yes, ParseBool, boolPicks },
{ TidyReplaceColor, MU, "replace-color", BL, no, ParseBool, boolPicks },
{ TidyShowErrors, DG, "show-errors", IN, 6, ParseInt, NULL },
{ TidyShowInfo, DG, "show-info", BL, yes, ParseBool, boolPicks },
{ TidyShowMarkup, PP, "markup", BL, yes, ParseBool, boolPicks },
{ TidyShowWarnings, DG, "show-warnings", BL, yes, ParseBool, boolPicks },
{ TidySkipNested, MU, "skip-nested", BL, yes, ParseBool, boolPicks }, /* 1642186 - Issue #65 */
{ TidySortAttributes, PP, "sort-attributes", IN, TidySortAttrNone,ParseSorter, sorterPicks },
{ TidyStrictTagsAttr, MU, "strict-tags-attributes", BL, no, ParseBool, boolPicks }, /* 20160209 - Issue #350 */
{ TidyTabSize, PP, "tab-size", IN, 8, ParseInt, NULL },
{ TidyUpperCaseAttrs, MU, "uppercase-attributes", BL, no, ParseBool, boolPicks },
{ TidyUpperCaseTags, MU, "uppercase-tags", BL, no, ParseBool, boolPicks },
{ TidyUseCustomTags, MU, "custom-tags", IN, TidyCustomNo, ParseUseCustomTags,customTagsPicks }, /* 20170309 - Issue #119 */
{ TidyVertSpace, PP, "vertical-space", IN, no, ParseAutoBool, autoBoolPicks }, /* #228 - tri option */
{ TidyWarnPropAttrs, MU, "warn-proprietary-attributes", BL, yes, ParseBool, boolPicks },
{ TidyWord2000, MU, "word-2000", BL, no, ParseBool, boolPicks },
{ TidyWrapAsp, PP, "wrap-asp", BL, yes, ParseBool, boolPicks },
{ TidyWrapAttVals, PP, "wrap-attributes", BL, no, ParseBool, boolPicks },
{ TidyWrapJste, PP, "wrap-jste", BL, yes, ParseBool, boolPicks },
{ TidyWrapLen, PP, "wrap", IN, 68, ParseInt, NULL },
{ TidyWrapPhp, PP, "wrap-php", BL, yes, ParseBool, boolPicks },
{ TidyWrapScriptlets, PP, "wrap-script-literals", BL, no, ParseBool, boolPicks },
{ TidyWrapSection, PP, "wrap-sections", BL, yes, ParseBool, boolPicks },
{ TidyWriteBack, MS, "write-back", BL, no, ParseBool, boolPicks },
{ TidyXhtmlOut, MU, "output-xhtml", BL, no, ParseBool, boolPicks },
{ TidyXmlDecl, MU, "add-xml-decl", BL, no, ParseBool, boolPicks },
{ TidyXmlOut, MU, "output-xml", BL, no, ParseBool, boolPicks },
{ TidyXmlPIs, MU, "assume-xml-procins", BL, no, ParseBool, boolPicks },
{ TidyXmlSpace, MU, "add-xml-space", BL, no, ParseBool, boolPicks },
{ TidyXmlTags, MU, "input-xml", BL, no, ParseBool, boolPicks },
{ N_TIDY_OPTIONS, XX, NULL, XY, 0, NULL, NULL }
{ TidyQuiet, MS, "quiet", BL, no, ParsePickList, &boolPicks },
{ TidyQuoteAmpersand, MU, "quote-ampersand", BL, yes, ParsePickList, &boolPicks },
{ TidyQuoteMarks, MU, "quote-marks", BL, no, ParsePickList, &boolPicks },
{ TidyQuoteNbsp, MU, "quote-nbsp", BL, yes, ParsePickList, &boolPicks },
{ TidyReplaceColor, MU, "replace-color", BL, no, ParsePickList, &boolPicks },
{ TidyShowErrors, DG, "show-errors", IN, 6, ParseInt, NULL },
{ TidyShowInfo, DG, "show-info", BL, yes, ParsePickList, &boolPicks },
{ TidyShowMarkup, PP, "markup", BL, yes, ParsePickList, &boolPicks },
{ TidyShowWarnings, DG, "show-warnings", BL, yes, ParsePickList, &boolPicks },
{ TidySkipNested, MU, "skip-nested", BL, yes, ParsePickList, &boolPicks }, /* 1642186 - Issue #65 */
{ TidySortAttributes, PP, "sort-attributes", IN, TidySortAttrNone,ParsePickList, &sorterPicks },
{ TidyStrictTagsAttr, MU, "strict-tags-attributes", BL, no, ParsePickList, &boolPicks }, /* 20160209 - Issue #350 */
{ TidyTabSize, PP, "tab-size", IN, 8, ParseInt, NULL },
{ TidyUpperCaseAttrs, MU, "uppercase-attributes", BL, no, ParsePickList, &boolPicks },
{ TidyUpperCaseTags, MU, "uppercase-tags", BL, no, ParsePickList, &boolPicks },
{ TidyUseCustomTags, MU, "custom-tags", IN, TidyCustomNo, ParsePickList, &customTagsPicks }, /* 20170309 - Issue #119 */
{ TidyVertSpace, PP, "vertical-space", IN, no, ParsePickList, &autoBoolPicks }, /* #228 - tri option */
{ TidyWarnPropAttrs, MU, "warn-proprietary-attributes", BL, yes, ParsePickList, &boolPicks },
{ TidyWord2000, MU, "word-2000", BL, no, ParsePickList, &boolPicks },
{ TidyWrapAsp, PP, "wrap-asp", BL, yes, ParsePickList, &boolPicks },
{ TidyWrapAttVals, PP, "wrap-attributes", BL, no, ParsePickList, &boolPicks },
{ TidyWrapJste, PP, "wrap-jste", BL, yes, ParsePickList, &boolPicks },
{ TidyWrapLen, PP, "wrap", IN, 68, ParseInt, NULL },
{ TidyWrapPhp, PP, "wrap-php", BL, yes, ParsePickList, &boolPicks },
{ TidyWrapScriptlets, PP, "wrap-script-literals", BL, no, ParsePickList, &boolPicks },
{ TidyWrapSection, PP, "wrap-sections", BL, yes, ParsePickList, &boolPicks },
{ TidyWriteBack, MS, "write-back", BL, no, ParsePickList, &boolPicks },
{ TidyXhtmlOut, MU, "output-xhtml", BL, no, ParsePickList, &boolPicks },
{ TidyXmlDecl, MU, "add-xml-decl", BL, no, ParsePickList, &boolPicks },
{ TidyXmlOut, MU, "output-xml", BL, no, ParsePickList, &boolPicks },
{ TidyXmlPIs, MU, "assume-xml-procins", BL, no, ParsePickList, &boolPicks },
{ TidyXmlSpace, MU, "add-xml-space", BL, no, ParsePickList, &boolPicks },
{ TidyXmlTags, MU, "input-xml", BL, no, ParsePickList, &boolPicks },
{ N_TIDY_OPTIONS, XX, NULL, XY, 0, NULL, NULL }
};
/* Should only be called by options set by name
** thus, it is cheaper to do a few scans than set
** up every option in a hash table.
@ -624,7 +593,7 @@ TidyTriState TY_(_cfgGetAutoBool)( TidyDocImpl* doc, TidyOptionId optId )
ulong val = TY_(_cfgGet)( doc, optId );
const TidyOptionImpl* opt = &option_defs[ optId ];
assert( opt && opt->type == TidyInteger
&& opt->parser == ParseAutoBool );
&& opt->parser == ParsePickList );
return (TidyTriState) val;
}
@ -1121,76 +1090,6 @@ Bool ParseInt( TidyDocImpl* doc, const TidyOptionImpl* entry )
return digits;
}
/* true/false or yes/no or 0/1 or "auto" only looks at 1st char */
static Bool ParseTriState( TidyTriState theState, TidyDocImpl* doc,
const TidyOptionImpl* entry, ulong* flag )
{
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
if (c == 't' || c == 'T' || c == 'y' || c == 'Y' || c == '1')
*flag = yes;
else if (c == 'f' || c == 'F' || c == 'n' || c == 'N' || c == '0')
*flag = no;
else if (theState == TidyAutoState && (c == 'a' || c =='A'))
*flag = TidyAutoState;
else
{
TY_(ReportBadArgument)( doc, entry->name );
return no;
}
return yes;
}
/* cr, lf or crlf */
Bool ParseNewline( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
int nl = -1;
tmbchar work[ 16 ] = {0};
tmbstr cp = work, end = work + sizeof(work);
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
while ( c!=EndOfStream && cp < end && !TY_(IsWhite)(c) && c != '\r' && c != '\n' )
{
*cp++ = (tmbchar) c;
c = AdvanceChar( cfg );
}
*cp = 0;
if ( TY_(tmbstrcasecmp)(work, "lf") == 0 )
nl = TidyLF;
else if ( TY_(tmbstrcasecmp)(work, "crlf") == 0 )
nl = TidyCRLF;
else if ( TY_(tmbstrcasecmp)(work, "cr") == 0 )
nl = TidyCR;
if ( nl < TidyLF || nl > TidyCR )
TY_(ReportBadArgument)( doc, entry->name );
else
TY_(SetOptionInt)( doc, entry->id, nl );
return ( nl >= TidyLF && nl <= TidyCR );
}
Bool ParseBool( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
ulong flag = 0;
Bool status = ParseTriState( TidyNoState, doc, entry, &flag );
if ( status )
TY_(SetOptionBool)( doc, entry->id, flag != 0 );
return status;
}
Bool ParseAutoBool( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
ulong flag = 0;
Bool status = ParseTriState( TidyAutoState, doc, entry, &flag );
if ( status )
TY_(SetOptionInt)( doc, entry->id, flag );
return status;
}
/* a string excluding whitespace */
Bool FUNC_UNUSED ParseName( TidyDocImpl* doc, const TidyOptionImpl* option )
{
@ -1242,15 +1141,77 @@ Bool ParseCSS1Selector( TidyDocImpl* doc, const TidyOptionImpl* option )
return yes;
}
/* A general parser for anything using pick lists. This provides the engine to
determine the proper option value, and can be used by parsers in addition to
ParsePickList that require special handling.
*/
Bool GetParsePickListValue( TidyDocImpl* doc, const TidyOptionImpl* entry, uint *result )
{
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
tmbchar work[ 16 ] = {0};
tmbstr cp = work, end = work + sizeof(work);
const PickListItem *item = NULL;
uint ix = 0;
while ( c!=EndOfStream && cp < end && !TY_(IsWhite)(c) && c != '\r' && c != '\n' )
{
*cp++ = (tmbchar) c;
c = AdvanceChar( cfg );
}
while ( (item = &(*entry->pickList)[ ix ]) && item->label )
{
ctmbstr input;
uint i = 0;
while ( ( input = &(*item->inputs[i]) ) )
{
if (TY_(tmbstrcasecmp)(work, input) == 0 )
{
*result = ix;
return yes;
}
++i;
}
++ix;
}
TY_(ReportBadArgument)( doc, entry->name );
return no;
}
/* A general parser for anything using pick lists that don't require special
handling.
*/
Bool ParsePickList( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
uint value;
if ( GetParsePickListValue( doc, entry, &value ) )
{
if ( entry->type == TidyBoolean )
TY_(SetOptionBool)( doc, entry->id, value );
else if ( entry->type == TidyInteger )
TY_(SetOptionInt)( doc, entry->id, value );
return yes;
}
TY_(ReportBadArgument)( doc, entry->name );
return no;
}
/*\
* 20150515 - support using tabs instead of spaces - Issue #108
* Sets the indent character to a tab if on, and set indent space count to 1
* and sets indent character to a space if off.
\*/
\*/
Bool ParseTabs( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
ulong flag = 0;
Bool status = ParseTriState( TidyNoState, doc, entry, &flag );
uint flag = 0;
Bool status = GetParsePickListValue( doc, entry, &flag );
if ( status ) {
Bool tabs = flag != 0 ? yes : no;
TY_(SetOptionBool)( doc, entry->id, tabs );
@ -1263,52 +1224,6 @@ Bool ParseTabs( TidyDocImpl* doc, const TidyOptionImpl* entry )
return status;
}
/* Parse the value of TidyUseCustomTags. Like other option values, we will
* look for the first character only, of no, blocklevel, empty, inline, pre.
*/
Bool ParseUseCustomTags( TidyDocImpl* doc, const TidyOptionImpl* entry )
{
uint value;
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
switch (c)
{
case 'n':
case 'N':
value = TidyCustomNo;
break;
case 'b':
value = TidyCustomBlocklevel;
break;
case 'e':
case 'E':
value = TidyCustomEmpty;
break;
case 'y':
case 'Y':
case 'i':
case 'I':
value = TidyCustomInline;
break;
case 'p':
case 'P':
value = TidyCustomPre;
break;
default:
TY_(ReportBadArgument)( doc, entry->name );
return no;
}
TY_(SetOptionInt)( doc, TidyUseCustomTags, value );
return yes;
}
/* Coordinates Config update and Tags data */
void TY_(DeclareUserTag)( TidyDocImpl* doc, TidyOptionId optId,
@ -1544,11 +1459,8 @@ ctmbstr TY_(CharEncodingOptName)( int encoding )
*/
Bool ParseDocType( TidyDocImpl* doc, const TidyOptionImpl* option )
{
tmbchar buf[ 32 ] = {0};
uint i = 0;
Bool status = yes;
TidyDoctypeModes dtmode = TidyDoctypeAuto;
uint value;
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
@ -1558,94 +1470,21 @@ Bool ParseDocType( TidyDocImpl* doc, const TidyOptionImpl* option )
{
status = ParseString(doc, option);
if (status)
{
TY_(SetOptionInt)( doc, TidyDoctypeMode, TidyDoctypeUser );
}
return status;
}
/* read first word */
while ( i < sizeof(buf)-1 && c != EndOfStream && !TY_(IsWhite)(c) )
if ( (status = GetParsePickListValue( doc, option, &value ) ) )
{
buf[i++] = (tmbchar) c;
c = AdvanceChar( cfg );
TY_(SetOptionInt)( doc, TidyDoctypeMode, value );
}
buf[i] = '\0';
if ( TY_(tmbstrcasecmp)(buf, "auto") == 0 )
dtmode = TidyDoctypeAuto;
else if ( TY_(tmbstrcasecmp)(buf, "html5") == 0 )
dtmode = TidyDoctypeHtml5;
else if ( TY_(tmbstrcasecmp)(buf, "omit") == 0 )
dtmode = TidyDoctypeOmit;
else if ( TY_(tmbstrcasecmp)(buf, "strict") == 0 )
dtmode = TidyDoctypeStrict;
else if ( TY_(tmbstrcasecmp)(buf, "loose") == 0 ||
TY_(tmbstrcasecmp)(buf, "transitional") == 0 )
dtmode = TidyDoctypeLoose;
else
{
TY_(ReportBadArgument)( doc, option->name );
status = no;
}
if ( status )
TY_(SetOptionInt)( doc, TidyDoctypeMode, dtmode );
return status;
}
Bool ParseRepeatAttr( TidyDocImpl* doc, const TidyOptionImpl* option )
{
Bool status = yes;
tmbchar buf[64] = {0};
uint i = 0;
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
while (i < sizeof(buf)-1 && c != EndOfStream && !TY_(IsWhite)(c))
{
buf[i++] = (tmbchar) c;
c = AdvanceChar( cfg );
}
buf[i] = '\0';
if ( TY_(tmbstrcasecmp)(buf, "keep-first") == 0 )
cfg->value[ TidyDuplicateAttrs ].v = TidyKeepFirst;
else if ( TY_(tmbstrcasecmp)(buf, "keep-last") == 0 )
cfg->value[ TidyDuplicateAttrs ].v = TidyKeepLast;
else
{
TY_(ReportBadArgument)( doc, option->name );
status = no;
}
return status;
}
Bool ParseSorter( TidyDocImpl* doc, const TidyOptionImpl* option )
{
Bool status = yes;
tmbchar buf[64] = {0};
uint i = 0;
TidyConfigImpl* cfg = &doc->config;
tchar c = SkipWhite( cfg );
while (i < sizeof(buf)-1 && c != EndOfStream && !TY_(IsWhite)(c))
{
buf[i++] = (tmbchar) c;
c = AdvanceChar( cfg );
}
buf[i] = '\0';
if ( TY_(tmbstrcasecmp)(buf, "alpha") == 0 )
cfg->value[ TidySortAttributes ].v = TidySortAttrAlpha;
else if ( TY_(tmbstrcasecmp)(buf, "none") == 0)
cfg->value[ TidySortAttributes ].v = TidySortAttrNone;
else
{
TY_(ReportBadArgument)( doc, option->name );
status = no;
}
return status;
}
@ -1691,12 +1530,22 @@ ctmbstr TY_(getNextOptionPick)( const TidyOptionImpl* option,
{
size_t ix;
ctmbstr val = NULL;
const PickListItem *item= NULL;
assert( option!=NULL && iter != NULL );
ix = (size_t) *iter;
if ( ix > 0 && ix < 16 && option->pickList )
val = option->pickList[ ix-1 ];
*iter = (TidyIterator) ( val && option->pickList[ix] ? ix + 1 : (size_t)0 );
if ( option->pickList )
{
if ( ix > 0 && ix < TIDY_PL_SIZE && option->pickList )
{
item = &(*option->pickList)[ ix-1 ];
val = item->label;
}
item = &(*option->pickList)[ ix ];
*iter = (TidyIterator) ( val && item->label ? ix + 1 : (size_t)0 );
}
return val;
}
@ -1730,12 +1579,19 @@ static int WriteOptionBool( const TidyOptionImpl* option, Bool bval, StreamOut*
static int WriteOptionPick( const TidyOptionImpl* option, uint ival, StreamOut* out )
{
uint ix;
const ctmbstr* val = option->pickList;
for ( ix=0; val[ix] && ix<ival; ++ix )
/**/;
if ( ix==ival && val[ix] )
return WriteOptionString( option, val[ix], out );
uint ix = 0;
const PickListItem *item = NULL;
if ( option-> pickList )
{
while ( (item = &(*option->pickList)[ ix ]) && item->label && ix<ival )
{
++ix;
}
if ( ix==ival && item->label )
return WriteOptionString( option, item->label, out );
}
return -1;
}
@ -1795,7 +1651,7 @@ static int SaveConfigToStream( TidyDocImpl* doc, StreamOut* out )
else
rc = WriteOptionPick( option, dtmode, out );
}
else if ( option->pickList )
else if ( option->pickList)
rc = WriteOptionPick( option, val->v, out );
else
{

87
src/config.h
View file

@ -1,30 +1,65 @@
#ifndef __CONFIG_H__
#define __CONFIG_H__
/* config.h -- read config file and manage config properties
(c) 1998-2006 (W3C) MIT, ERCIM, Keio University
See tidy.h for the copyright notice.
config files associate a property name with a value.
// comments can start at the beginning of a line
# comments can start at the beginning of a line
name: short values fit onto one line
name: a really long value that
continues on the next line
property names are case insensitive and should be less than
60 characters in length and must start at the begining of
the line, as whitespace at the start of a line signifies a
line continuation.
*/
/**************************************************************************//**
* @file
* Read configuration files and manage configuration properties.
*
* Config files associate a property name with a value.
*
* // comments can start at the beginning of a line
* # comments can start at the beginning of a line
* name: short values fit onto one line
* name: a really long value that
* continues on the next line
*
* Property names are case insensitive and should be less than 60 characters
* in length, and must start at the begining of the line, as whitespace at
* the start of a line signifies a line continuation.
*
* @author HTACG, et al (consult git log)
*
* @copyright
* Copyright (c) 1998-2017 World Wide Web Consortium (Massachusetts
* Institute of Technology, European Research Consortium for Informatics
* and Mathematics, Keio University) and HTACG.
* @par
* All Rights Reserved.
* @par
* See `tidy.h` for the complete license.
*
* @date Additional updates: consult git log
*
******************************************************************************/
#include "forward.h"
#include "tidy.h"
#include "streamio.h"
/** PickLists may have up to 16 items. For some reason,
** this limit has always been hard-coded into Tidy.
*/
#define TIDY_PL_SIZE 16
/** Structs of this type contain information needed in order to present pick lists,
** relate pick list entries to public enum values, and parse strings that are
** accepted in order to assign the value.
*/
typedef struct PickListItem {
ctmbstr label; /**< PickList label for this item. */
const int value; /**< The option value represented by this label. */
ctmbstr inputs[10]; /**< String values that can select this value. */
} PickListItem;
/** An array of PickListItems, fixed in size for in-code declarations.
** Arrays must be populated in 0 to 10 order, as the option value is assigned
** based on this index and *not* on the structures' value field. It remains
** a best practice, however, to assign a public enum value with the proper
** index value.
*/
typedef const PickListItem PickListItems[TIDY_PL_SIZE];
struct _tidy_option;
typedef struct _tidy_option TidyOptionImpl;
@ -33,13 +68,13 @@ typedef Bool (ParseProperty)( TidyDocImpl* doc, const TidyOptionImpl* opt );
struct _tidy_option
{
TidyOptionId id;
TidyConfigCategory category; /* put 'em in groups */
ctmbstr name; /* property name */
TidyOptionType type; /* string, int or bool */
ulong dflt; /* default for TidyInteger and TidyBoolean */
ParseProperty* parser; /* parsing method, read-only if NULL */
const ctmbstr* pickList; /* pick list */
ctmbstr pdflt; /* default for TidyString */
TidyConfigCategory category; /* put 'em in groups */
ctmbstr name; /* property name */
TidyOptionType type; /* string, int or bool */
ulong dflt; /* default for TidyInteger and TidyBoolean */
ParseProperty* parser; /* parsing method, read-only if NULL */
PickListItems* pickList; /* new style pick list */
ctmbstr pdflt; /* default for TidyString */
};
typedef union

19
src/tidylib.c
View file

@ -476,15 +476,22 @@ ctmbstr TIDY_CALL tidyOptGetEncName( TidyDoc tdoc, TidyOptionId optId )
ctmbstr TIDY_CALL tidyOptGetCurrPick( TidyDoc tdoc, TidyOptionId optId )
{
const TidyOptionImpl* option = TY_(getOption)( optId );
if ( option && option->pickList )
{
uint ix, pick = tidyOptGetInt( tdoc, optId );
const ctmbstr* pL = option->pickList;
for ( ix=0; *pL && ix < pick; ++ix )
++pL;
if ( *pL )
return *pL;
uint ix = 0;
uint pick = tidyOptGetInt( tdoc, optId );
const PickListItem *item = NULL;
// loop through the picklist until index matches the value
while ( (item = &(*option->pickList)[ ix ]) && item->label && ix<pick )
{
++ix;
}
if ( ix==pick && item->label )
return item->label;
}
return NULL;
}