diff options
| -rw-r--r-- | Makefile | 11 | ||||
| -rw-r--r-- | README.md | 5 | ||||
| -rwxr-xr-x | adapter/protocol/man | 258 | ||||
| -rw-r--r-- | adapter/protocol/man.nim | 308 | ||||
| -rw-r--r-- | doc/cha.1 | 2 | ||||
| -rw-r--r-- | doc/protocols.md | 4 | ||||
| -rw-r--r-- | lib/quickjs/quickjs.c | 4 | ||||
| -rw-r--r-- | res/license.html | 18 | ||||
| -rw-r--r-- | src/js/regex.nim | 43 | ||||
| -rw-r--r-- | todo | 8 |
10 files changed, 349 insertions, 312 deletions
@@ -38,6 +38,8 @@ else ifeq ($(TARGET),release1) FLAGS += -d:release --debugger:native endif +QJSOBJ = $(OBJDIR)/quickjs + .PHONY: all all: $(OUTDIR_BIN)/cha $(OUTDIR_BIN)/mancha $(OUTDIR_CGI_BIN)/http \ $(OUTDIR_CGI_BIN)/gmifetch $(OUTDIR_LIBEXEC)/gmi2html \ @@ -107,9 +109,13 @@ $(OUTDIR_CGI_BIN)/cha-finger: adapter/protocol/cha-finger @mkdir -p $(OUTDIR_CGI_BIN) cp adapter/protocol/cha-finger $(OUTDIR_CGI_BIN) -$(OUTDIR_CGI_BIN)/man: adapter/protocol/man +$(OUTDIR_CGI_BIN)/man: adapter/protocol/man.nim $(QJSOBJ)/libregexp.o \ + $(QJSOBJ)/libunicode.o $(QJSOBJ)/cutils.o src/js/regex.nim \ + src/bindings/libregexp.nim src/types/opt.nim src/utils/twtstr.nim @mkdir -p $(OUTDIR_CGI_BIN) - cp adapter/protocol/man $(OUTDIR_CGI_BIN) + $(NIMC) $(FLAGS) --nimcache:"$(OBJDIR)/$(TARGET)/man" \ + --passL:"$(QJSOBJ)/libregexp.o $(QJSOBJ)/cutils.o $(QJSOBJ)/libunicode.o" \ + -o:"$(OUTDIR_CGI_BIN)/man" adapter/protocol/man.nim $(OUTDIR_CGI_BIN)/spartan: adapter/protocol/spartan @mkdir -p $(OUTDIR_CGI_BIN) @@ -164,7 +170,6 @@ $(OUTDIR_LIBEXEC)/urlenc: adapter/tools/urlenc.nim src/utils/twtstr.nim -o:"$(OUTDIR_LIBEXEC)/urlenc" adapter/tools/urlenc.nim CFLAGS = -fwrapv -g -Wall -O2 -DCONFIG_VERSION=\"$(shell cat lib/quickjs/VERSION)\" -QJSOBJ = $(OBJDIR)/quickjs # Dependencies $(QJSOBJ)/cutils.o: lib/quickjs/cutils.h @@ -35,14 +35,13 @@ supported yet.) 6. If you want manpages, run `make manpage`. (This requires pandoc to be installed.) 7. Finally, install using `make install` (e.g. `sudo make install`). -8. (Optional): install Perl so that the man page viewer (`mancha`) works too. Then, try: ```bash $ cha -V # open in visual mode for a list of default keybindings $ cha example.org # open your favorite website directly from the shell -$ mancha cha # read the cha(1) man page using `mancha' (requires Perl) +$ mancha cha # read the cha(1) man page using `mancha' ``` ## Features @@ -66,7 +65,7 @@ Currently implemented features are: * supports several protocols: HTTP(S), FTP, Gopher, Gemini, Finger, etc. * can load user-defined protocols/file formats using [local CGI](doc/localcgi.md), [urimethodmap](doc/urimethodmap.md) and [mailcap](doc/mailcap.md) -* man page viewer (based on w3mman) +* man page viewer (like w3mman) * mouse support ...with a lot more [planned](todo). diff --git a/adapter/protocol/man b/adapter/protocol/man deleted file mode 100755 index c322a6d2..00000000 --- a/adapter/protocol/man +++ /dev/null @@ -1,258 +0,0 @@ -#!/usr/bin/env perl -# -# From w3m. -# -# Note that this script has licensing terms different from those of Chawan. -# See /res/license.html#w3m for details. -# -# Usage: install perl, then look up man pages using: -# -# $ cha man:cha # view in any manual (man cha) -# $ cha 'man:cha(1)' # view in a specific manual (man -s 1 cha) -# $ cha man-k:cha # search in any manual (man -k cha) -# $ cha 'man-k:cha(1)' # search in a specific manual (man -k cha -s 1) -# -# You may also use the `mancha` wrapper. - -$MAN = $ENV{'MANCHA_MAN'} || '/usr/bin/man'; -$QUERY = $ENV{'QUERY_STRING'} || $ARGV[0]; -$SCRIPT_NAME = $ENV{'SCRIPT_NAME'} || $0; -$CGI = "man:"; -$CGI2 = "file:"; -# $CGI2 = "file:///\$LIB/hlink.cgi?"; -$SQUEEZE = 1; -$ENV{'PAGER'} = 'cat'; - -if ($QUERY =~ /^man-k:/) { - $QUERY =~ s/^man-k://; - my $keyword = &form_decode($QUERY); - my $sectionopt = ""; - if ($keyword =~ s/(.*)\((\w+)\)$//) { - $keyword = $1; - $sectionopt = "-s $2 "; - } - my $k = &html_quote($keyword); - print <<EOF; -Content-Type: text/html - -<html> -<head><title>man $sectionopt-k $k</title></head> -<body> -<h2>man $sectionopt-k <b>$k</b></h2> -<ul> -EOF - $keyword =~ s:([^-\w\200-\377.,])::g; - open(F, "$MAN $sectionopt -k $keyword 2> /dev/null |"); - while(<F>) { - chop; - $_ = &html_quote($_); - s/(\s+-.*)$//; - $title = $1; - s@(\w[\w.\-]*(\s*\,\s*\w[\w.\-]*)*)\s*(\([\dn]\w*\))@&keyword_ref($1, $3)@ge; - print "<li>$_$title\n"; - } - close(F); - print <<EOF; -</ul> -</body> -</html> -EOF - exit; -} elsif ($QUERY =~ /^man-l:/) { - $QUERY =~ s/^man-l://; - $file = &form_decode($QUERY); - open(F, "GROFF_NO_SGR=1 MAN_KEEP_FORMATTING=1 $MAN $file 2> /dev/null |"); -} else { - $QUERY =~ s/^man://; - my $man = &form_decode($QUERY); - if ($man =~ s/\((\w+)\)$//) { - $section = $1; - $man_section = "$man($1)"; - } else { - $section = ""; - $man_section = "$man"; - } - - $section =~ s:([^-\w\200-\377.,])::g; - $man =~ s:([^-\w\200-\377.,])::g; - open(F, "GROFF_NO_SGR=1 MAN_KEEP_FORMATTING=1 $MAN $section $man 2> /dev/null |"); -} - -$ok = 0; -undef $header; -$blank = -1; -$cmd = ""; -$prev = ""; -while(<F>) { - if (! defined($header)) { - /^\s*$/ && next; - $header = $_; - $space = $header; - chop $space; - $space =~ s/\S.*//; - } elsif ($_ eq $header) { # delete header - $blank = -1; - next; - } elsif (!/\010/ && /^$space[\w\200-\377].*\s\S/o) { # delete footer - $blank = -1; - next; - } - if ($SQUEEZE) { - if (/^\s*$/) { - $blank || $blank++; - next; - } elsif ($blank) { - $blank > 0 && print "\n"; - $blank = 0; - } - } - - s/\&/\&/g; - s/\</\</g; - s/\>/\>/g; - # non ASCII UTF-8 codepoint - my $utf8="[\300-\337][\200-\277]|[\340-\357][\200-\277]{2}|[\360-\367][\200-\277]{3}|[\370-\373][\200-\277]{4}|[\374\375][\200-\277]{5}"; - - s@($utf8)(\010\1)+@<b>$1</b>@g; - s@(\&\w+;|.)(\010\1)+@<b>$1</b>@g; - s@_\010((\<b\>)?($utf8)(\</b\>)?)@<u>$1</u>@g; - s@_\010((\<b\>)?(\&\w+\;|.)(\</b\>)?)@<u>$1</u>@g; - s@((\<b\>)?($utf8)(\</b\>)?)\010_@<u>$1</u>@g; - s@((\<b\>)?(\&\w+\;|.)(\</b\>)?)\010_@<u>$1</u>@g; - s@.\010(.)@$1@g; - - s@\</b\>\</u\>\<b\>_\</b\>\<u\>\<b\>@_@g; - s@\</u\>\<b\>_\</b\>\<u\>@_@g; - s@\</u\>\<u\>@@g; - s@\</b\>\<b\>@@g; - - if (! $ok) { - /^No/ && last; - print <<EOF; -Content-Type: text/html - -<html> -<head><title>man $man_section</title></head> -<body> -<pre> -EOF - print; - $ok = 1; - next; - } - - s@(https?|ftp)://[\w.\-/~]+[\w/]@<a href="$&">$&</a>@g; - s@\b(mailto:|)(\w[\w.\-]*\@\w[\w.\-]*\.[\w.\-]*\w)@<a href="mailto:$2">$1$2</a>@g; - s@(\W)(\~?/[\w.][\w.\-/~]*)@$1 . &file_ref($2)@ge; - s@(include(<\/?[bu]\>|\s)*\<)([\w.\-/]+)@$1 . &include_ref($3)@ge; - if ($prev && m@^\s*(\<[bu]\>)*(\w[\w.\-]*)(\</[bu]\>)*(\([\dm]\w*\))@) { - $cmd .= "$2$4"; - $prev =~ s@(\w[\w.\-]*-)((\</[bu]\>)*\s*)$@<a href="$CGI$cmd">$1</a>$2@; - print $prev; - $prev = ''; - s@^(\s*(\<[bu]\>)*)(\w[\w.\-]*)@@; - print "$1<a href=\"$CGI$cmd\">$3</a>"; - } elsif ($prev) { - print $prev; - $prev = ''; - } - s@(\w[\w.\-]*)((\</[bu]\>)*)(\([\dm]\w*\))@<a href="$CGI$1$4">$1</a>$2$4@g; - if (m@(\w[\w.\-]*)-(\</[bu]\>)*\s*$@) { - $cmd = $1; - $prev = $_; - next; - } - print; -} -if ($prev) { - print $prev; -} -close(F); -if (! $ok) { - if ($file) { - print "Cha-Control: ConnectionError 4 file $file not found"; - } else { - print "Cha-Control: ConnectionError 4 no manual entry for $man_section"; - } - exit 1; -} -print <<EOF; -</pre> -</body> -</html> -EOF - -sub is_command { - local($_) = @_; - local($p); - - (! -d && -x) || return 0; - if (! %PATH) { - for $p (split(":", $ENV{'PATH'})) { - $p =~ s@/+$@@; - $PATH{$p} = 1; - } - } - s@/[^/]*$@@; - return defined($PATH{$_}); -} - -sub file_ref { - local($_) = @_; - - if (&is_command($_)) { - ($man = $_) =~ s@.*/@@; - return "<a href=\"$CGI$man\">$_</a>"; - } - if (/^\~/ || -f || -d) { - ($file = $_) =~ s/^\~/$ENV{"HOME"}/; - return "<a href=\"$CGI2$file\">$_</a>"; - } - return $_; -} - -sub include_ref { - local($_) = @_; - local($d); - - for $d ( - "/usr/include", - "/usr/local/include", - "/usr/X11R6/include", - "/usr/X11/include", - "/usr/X/include", - "/usr/include/X11" - ) { - -f "$d/$_" && return "<a href=\"$CGI2$d/$_\">$_</a>"; - } - return $_; -} - -sub keyword_ref { - local($_, $s) = @_; - local(@a) = (); - - for (split(/\s*,\s*/)) { - push(@a, "<a href=\"$CGI$_$s\">$_</a>"); - } - return join(", ", @a) . $s; -} - -sub html_quote { - local($_) = @_; - local(%QUOTE) = ( - '<', '<', - '>', '>', - '&', '&', - '"', '"', - ); - s/[<>&"]/$QUOTE{$&}/g; - return $_; -} - -sub form_decode { - local($_) = @_; - s/\+/ /g; - s/%([\da-f][\da-f])/pack('c', hex($1))/egi; - return $_; -} diff --git a/adapter/protocol/man.nim b/adapter/protocol/man.nim new file mode 100644 index 00000000..4fc418d4 --- /dev/null +++ b/adapter/protocol/man.nim @@ -0,0 +1,308 @@ +import std/os +import std/posix +import std/strutils +import std/unicode + +import bindings/libregexp +import js/regex +import types/opt +import utils/twtstr + +proc lre_check_stack_overflow(opaque: pointer; alloca_size: csize_t): cint + {.exportc.} = + return 0 + +proc parseSection(query: string): tuple[page, section: string] = + var section = "" + if query.len > 0 and query[^1] == ')': + for i in countdown(query.high, 0): + if query[i] == '(': + section = query.substr(i + 1, query.high - 1) + break + if section != "": + return (query.substr(0, query.high - 2 - section.len), section) + return (query, "") + +func processBackspace(line: string): string = + var s = "" + var i = 0 + var thiscs = 0 .. -1 + var bspace = false + var inU = false + var inB = false + var pendingInU = false + var pendingInB = false + template flushChar = + if pendingInU != inU: + s &= (if inU: "</u>" else: "<u>") + inU = pendingInU + if pendingInB != inB: + s &= (if inB: "</b>" else: "<b>") + inB = pendingInB + if thiscs.len > 0: + let cs = case line[thiscs.a] + of '&': "&" + of '<': "<" + of '>': ">" + else: line.substr(thiscs.a, thiscs.b) + s &= cs + thiscs = i ..< i + n + pendingInU = false + pendingInB = false + while i < line.len: + # this is the same "sometimes works" algorithm as in ansi2html + if line[i] == '\b' and thiscs.len > 0: + bspace = true + inc i + continue + let n = line.runeLenAt(i) + if thiscs.len == 0: + thiscs = i ..< i + n + i += n + continue + if bspace and thiscs.len > 0: + if line[i] == '_' and not pendingInU and line[thiscs.a] != '_': + pendingInU = true + elif line[thiscs.a] == '_' and not pendingInU and line[i] != '_': + # underscore comes first; set thiscs to the current charseq + thiscs = i ..< i + n + pendingInU = true + elif line[i] == '_' and line[thiscs.a] == '_': + if pendingInB: + pendingInU = true + else: + pendingInB = true + elif not pendingInB: + pendingInB = true + bspace = false + else: + flushChar + i += n + let n = 0 + flushChar + pendingInU = false + pendingInB = false + flushChar + return s + +proc isCommand(paths: seq[string]; s: string): bool = + for p in paths: + if fileExists(p & s): + return true + false + +iterator myCaptures(captures: var seq[RegexCapture]; target: int; + offset: var int): RegexCapture = + for cap in captures.mitems: + if cap.i == target: + cap.s += offset + cap.e += offset + yield cap + +proc processManpage(file: File) = + template re(s: static string): Regex = + let r = s.compileRegex({LRE_FLAG_GLOBAL, LRE_FLAG_UTF16}) + if r.isNone: + stdout.write(s & ": " & r.error) + return + r.get + # regexes partially from w3mman2html + let linkRe = re"(https?|ftp)://[\w/~.-]+" + let mailRe = re"(mailto:|)(\w[\w.-]*@[\w-]+\.[\w.-]*)" + let fileRe = re"(file:)?[/~][\w/~.-]+[\w/]" + let includeRe = re"#include(</?[bu]>|\s)*<([\w./-]+)" + let manRe = re"(</?[bu]>)*(\w[\w.-]*)(</?[bu]>)*(\([0-9nlx]\w*\))" + var paths: seq[string] = @[] + for p in getEnv("PATH").split(':'): + var i = p.high + while i > 0 and p[i] == '/': + dec i + paths.add(p.substr(0, i) & "/") + var line: string + var wasBlank = false + while file.readLine(line): + if line == "": + if wasBlank: + continue + wasBlank = true + else: + wasBlank = false + var line = line.processBackspace() + if (var res = linkRe.exec(line); res.success): + var offset = 0 + for cap in res.captures.myCaptures(0, offset): + let s = line[cap.s..<cap.e] + let link = "<a href='" & s & "'>" & s & "</a>" + line[cap.s..<cap.e] = link + offset += link.len - (cap.e - cap.s) + if (var res = mailRe.exec(line); res.success): + var offset = 0 + for cap in res.captures.myCaptures(2, offset): + let s = line[cap.s..<cap.e] + let link = "<a href='mailto:" & s & "'>" & s & "</a>" + line[cap.s..<cap.e] = link + offset += link.len - (cap.e - cap.s) + if (var res = fileRe.exec(line); res.success): + var offset = 0 + for cap in res.captures.myCaptures(0, offset): + let s = line[cap.s..<cap.e] + let target = s.expandTilde() + if not fileExists(target) and not symlinkExists(target) and + not dirExists(target): + continue + let link = if paths.isCommand(target): + "<a href='man:" & target.afterLast('/') & "'>" & s & "</a>" + else: + "<a href='file:" & target & "'>" & s & "</a>" + line[cap.s..<cap.e] = link + offset += link.len - (cap.e - cap.s) + if (var res = includeRe.exec(line); res.success): + var offset = 0 + for cap in res.captures.myCaptures(2, offset): + let s = line[cap.s..<cap.e] + const includePaths = [ + "/usr/include/", + "/usr/local/include/", + "/usr/X11R6/include/", + "/usr/X11/include/", + "/usr/X/include/", + "/usr/include/X11/" + ] + for path in includePaths: + let file = path & s + if fileExists(file): + let link = "<a href='file:" & file & "'>" & s & "</a>" + line[cap.s..<cap.e] = link + offset += link.len - (cap.e - cap.s) + break + if (var res = manRe.exec(line); res.success): + var offset = 0 + for j, cap in res.captures.mpairs: + if cap.i == 0: + cap.s += offset + cap.e += offset + var manCap = res.captures[j + 2] + manCap.s += offset + manCap.e += offset + var secCap = res.captures[j + 4] + secCap.s += offset + secCap.e += offset + let man = line[manCap.s..<manCap.e] + let cat = man & line[secCap.s..<secCap.e] + let link = "<a href='man:" & cat & "'>" & man & "</a>" + line[manCap.s..<manCap.e] = link + offset += link.len - (manCap.e - manCap.s) + stdout.write(line & "\n") + +proc doMan(man, keyword, section: string) = + let sectionOpt = if section == "": "" else: " -s " & section + let cmd = "GROFF_NO_SGR=1 MAN_KEEP_FORMATTING=1 " & + man & sectionOpt & " " & keyword & " 2>&1" + let file = popen(cstring(cmd), "r") + if file == nil: + stdout.write("Cha-Control: ConnectionError 1 failed to run " & cmd) + return + var line0: string + if not file.readLine(line0) or file.endOfFile(): + discard file.pclose() + if line0.startsWith("man: "): + line0 = line0.after(' ') + stdout.write("Cha-Control: ConnectionError 4 " & line0) + return + var manword = keyword + if section != "": + manword &= '(' & section & ')' + stdout.write("""Content-Type: text/html + +<title>man """ & manword & """</title> +<pre>""" & line0 & "\n") + file.processManpage() + discard file.pclose() + +proc doLocal(man, keyword: string) = + let cmd = "GROFF_NO_SGR=1 MAN_KEEP_FORMATTING=1 " & + man & " -l " & keyword & " 2>/dev/null" + let file = popen(cstring(cmd), "r") + if file == nil: + stdout.write("Cha-Control: ConnectionError 1 failed to run " & cmd) + return + stdout.write("""Content-Type: text/html + +<title>man -l """ & keyword & """</title> +<h1>man -l <b>""" & keyword & """</b></h1> +<pre>""") + file.processManpage() + discard file.pclose() + +proc doKeyword(man, keyword, section: string) = + let sectionOpt = if section == "": "" else: " -s " & section + let cmd = man & sectionOpt & " -k " & keyword & " 2>/dev/null" + let file = popen(cstring(cmd), "r") + if file == nil: + stdout.write("Cha-Control: ConnectionError 1 failed to run " & cmd) + return + stdout.write("Content-Type: text/html\n\n") + stdout.write("<title>man" & sectionOpt & " -k " & keyword & "</title>\n") + stdout.write("<h1>man" & sectionOpt & " -k <b>" & keyword & "</b></h1>\n") + stdout.write("<ul>") + var line: string + template die = + stdout.write("Error parsing line! " & line) + return + while file.readLine(line): + if line.len == 0: + stdout.write("\n") + continue + # collect titles + var titles: seq[string] = @[] + var i = 0 + while true: + let title = line.until({'(', ','}, i) + i += title.len + titles.add(title) + if i >= line.len or line[i] == '(': + break + i = line.skipBlanks(i + 1) + # collect section + if line[i] != '(': die + let sectionText = line.substr(i, line.find(')', i)) + i += sectionText.len + # create line + var section = sectionText.until(',') # for multiple sections, take first + if section[^1] != ')': + section &= ')' + var s = "<li>" + for i, title in titles: + let title = title.htmlEscape() + s &= "<a href='man:" & title & section & "'>" & title & "</a>" + if i < titles.high: + s &= ", " + s &= sectionText + s &= line.substr(i) + s &= "\n" + stdout.write(s) + +proc main() = + var man = getEnv("MANCHA_MAN") + if man == "": + man = "/usr/bin/man" + if not fileExists(man): + man = "/bin/man" + if not fileExists(man): + man = "/usr/local/bin/man" + if not fileExists(man): + man = "/usr/bin/env man" + doAssert getAppFilename() != man # don't accidentally fork bomb ourselves + let query = percentDecode(getEnv("QUERY_STRING")) + if query.startsWith("man:"): + let (keyword, section) = parseSection(query.after(':')) + doMan(man, keyword, section) + elif query.startsWith("man-k:"): + let (keyword, section) = parseSection(query.after(':')) + doKeyword(man, keyword, section) + elif query.startsWith("man-l:"): + doLocal(man, query.after(':')) + else: + stdout.write("Cha-Control: ConnectionError 1 invalid invocation") + +main() @@ -1,4 +1,4 @@ -.TH CHA 1 +.TH cha 1 .SH NAME cha - the Chawan text-mode browser .SH SYNOPSIS diff --git a/doc/protocols.md b/doc/protocols.md index 2f1c7afd..84a8dde1 100644 --- a/doc/protocols.md +++ b/doc/protocols.md @@ -125,8 +125,8 @@ and `about:license`. `man:`, `man-k:` and `man-l:` are wrappers around the commands `man`, `man -k` and `man -l`. These look up man pages using `/usr/bin/man` and turn on-page references into links. A wrapper command `mancha` also exists; this has an -interface similar to `man`. Note that Perl is required for these protocols -to work. +interface similar to `man`. Note: this used to be based on w3mman2html.cgi, but +it has been rewritten in Nim (and therefore no longer depends on Perl either). `data:` decodes a data URL as defined in RFC 2397. diff --git a/lib/quickjs/quickjs.c b/lib/quickjs/quickjs.c index a5e51d19..9c5e948c 100644 --- a/lib/quickjs/quickjs.c +++ b/lib/quickjs/quickjs.c @@ -43960,16 +43960,20 @@ fail: BOOL lre_check_stack_overflow(void *opaque, size_t alloca_size) { + if (!opaque) + return 0; JSContext *ctx = opaque; return js_check_stack_overflow(ctx->rt, alloca_size); } +#if 0 void *lre_realloc(void *opaque, void *ptr, size_t size) { JSContext *ctx = opaque; /* No JS exception is raised here */ return js_realloc_rt(ctx->rt, ptr, size); } +#endif static JSValue js_regexp_exec(JSContext *ctx, JSValueConst this_val, int argc, JSValueConst *argv) diff --git a/res/license.html b/res/license.html index e58a4876..405d6e2d 100644 --- a/res/license.html +++ b/res/license.html @@ -23,7 +23,6 @@ Table of contents: <LI><A HREF="#chawan">Chawan</A> <LI><A HREF="#quickjs">QuickJS</A> <LI><A HREF="#punycode">Punycode library</A> -<LI><A HREF="#w3mman">w3mman</A> </UL> <H2 id=chawan>Chawan</H2> <PRE> @@ -115,22 +114,5 @@ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. </PRE> -<H2 id=w3mman>w3mman</H2> -Chawan uses a customized version of w3mman2html.cgi from w3m for viewing man -pages. This script is distributed under the following terms: -<PRE> -(C) Copyright 1994-2002 by Akinori Ito -(C) Copyright 2002-2011 by Akinori Ito, Hironori Sakamoto, Fumitoshi Ukai - -Use, modification and redistribution of this software is hereby granted, -provided that this entire copyright notice is included on any copies of -this software and applications and derivations thereof. - -This software is provided on an "as is" basis, without warranty of any -kind, either expressed or implied, as to any matter including, but not -limited to warranty of fitness of purpose, or merchantability, or -results obtained from use of this software. -</PRE> - </BODY> </HTML> diff --git a/src/js/regex.nim b/src/js/regex.nim index fbbaf55f..e189b430 100644 --- a/src/js/regex.nim +++ b/src/js/regex.nim @@ -2,7 +2,6 @@ import std/unicode import bindings/libregexp -import bindings/quickjs import types/opt import utils/twtstr @@ -13,35 +12,37 @@ type bytecode: seq[uint8] buf: string + RegexCapture* = tuple # start, end, index + s, e: int + i: int32 + RegexResult* = object success*: bool - captures*: seq[tuple[s, e: int]] # start, end + captures*: seq[RegexCapture] RegexReplace* = object regex: Regex rule: string global: bool -var dummyRuntime = JS_NewRuntime() -var dummyContext = JS_NewContextRaw(dummyRuntime) - func `$`*(regex: Regex): string = regex.buf -proc compileRegex*(buf: string, flags: LREFlags = {}): Result[Regex, string] = - var error_msg_size = 64 - var error_msg = newString(error_msg_size) - prepareMutation(error_msg) +# this is hardcoded into quickjs, so we must override it here. +proc lre_realloc(opaque, p: pointer; size: csize_t): pointer {.exportc.} = + return realloc(p, size) + +proc compileRegex*(buf: string; flags: LREFlags = {}): Result[Regex, string] = + var errorMsg = newString(64) var plen: cint - let bytecode = lre_compile(addr plen, cstring(error_msg), - cint(error_msg_size), cstring(buf), csize_t(buf.len), flags.toCInt, - dummyContext) + let bytecode = lre_compile(addr plen, cstring(errorMsg), cint(errorMsg.len), + cstring(buf), csize_t(buf.len), flags.toCInt, nil) if bytecode == nil: - return err(error_msg.until('\0')) # Failed to compile. + return err(errorMsg.until('\0')) # Failed to compile. assert plen > 0 var bcseq = newSeqUninitialized[uint8](plen) copyMem(addr bcseq[0], bytecode, plen) - dummyRuntime.js_free_rt(bytecode) + dealloc(bytecode) let regex = Regex( buf: buf, bytecode: bcseq @@ -83,7 +84,7 @@ proc compileMatchRegex*(buf: string): Result[Regex, string] = buf2 &= "$" return compileRegex(buf2) -proc compileSearchRegex*(str: string, defaultFlags: LREFlags): +proc compileSearchRegex*(str: string; defaultFlags: LREFlags): Result[Regex, string] = # Emulate vim's \c/\C: override defaultFlags if one is found, then remove it # from str. @@ -108,13 +109,13 @@ proc compileSearchRegex*(str: string, defaultFlags: LREFlags): flags.incl(LRE_FLAG_GLOBAL) # for easy backwards matching return compileRegex(s, flags) -proc exec*(regex: Regex, str: string, start = 0, length = -1, nocaps = false): RegexResult = +proc exec*(regex: Regex; str: string; start = 0; length = -1; nocaps = false): + RegexResult = let length = if length == -1: str.len else: length - assert 0 <= start and start <= length - + assert start in 0 .. length let bytecode = unsafeAddr regex.bytecode[0] let captureCount = lre_get_capture_count(bytecode) var capture: ptr UncheckedArray[int] = nil @@ -126,7 +127,7 @@ proc exec*(regex: Regex, str: string, start = 0, length = -1, nocaps = false): R var start = start while true: let ret = lre_exec(cast[ptr ptr uint8](capture), bytecode, - cast[ptr uint8](cstr), cint(start), cint(length), cint(3), dummyContext) + cast[ptr uint8](cstr), cint(start), cint(length), cint(3), nil) if ret != 1: #TODO error handling? (-1) break result.success = true @@ -138,7 +139,7 @@ proc exec*(regex: Regex, str: string, start = 0, length = -1, nocaps = false): R for i in 0 ..< captureCount: let s = capture[i * 2] - cstrAddress let e = capture[i * 2 + 1] - cstrAddress - result.captures.add((s, e)) + result.captures.add((s, e, i)) if LRE_FLAG_GLOBAL notin flags: break if start >= str.len: @@ -148,5 +149,5 @@ proc exec*(regex: Regex, str: string, start = 0, length = -1, nocaps = false): R if captureCount > 0: dealloc(capture) -proc match*(regex: Regex, str: string, start = 0, length = str.len): bool = +proc match*(regex: Regex; str: string; start = 0; length = str.len): bool = return regex.exec(str, start, length, nocaps = true).success @@ -80,12 +80,8 @@ images: - incremental decoding (maybe implement streams first?) - separate image decoder process? or just run on a different thread? man: -- mancha: - * detect man directory automatically - * eventually rewrite in Nim -- man pages: - * add a DOM -> man page converter so that we do not depend on pandoc - for man page conversion +* add a DOM -> man page converter so that we do not depend on pandoc + for man page conversion gmifetch: - rewrite in Nim etc: |