Archives/gnosygnu_xowa
1
0
Fork 0
mirror of https://github.com/gnosygnu/xowa.git synced 2026-03-02 03:49:30 +00:00
Code Issues Projects Releases Wiki Activity

v3.1.1.1

Browse Source
This commit is contained in:
gnosygnu
2016-01-03 21:27:38 -05:00
parent 9509363f46
commit 096045614c
647 changed files with 11693 additions and 7648 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
400_xowa/src/gplx/langs/htmls/docs/Gfh_atr.java Normal file
View File

@@ -0,0 +1,47 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
public class Gfh_atr implements gplx.core.brys.Bfr_arg {
public Gfh_atr(int idx, int atr_bgn, int atr_end, byte[] key, byte[] val, byte[] src, int val_bgn, int val_end) {
this.idx = idx; this.atr_bgn = atr_bgn; this.atr_end = atr_end; this.key = key; this.val = val;
this.src = src; this.val_bgn = val_bgn; this.val_end = val_end;
}
public byte[] Src() {return src;} private final byte[] src;
public int Idx() {return idx;} private final int idx;
public int Atr_bgn() {return atr_bgn;} private final int atr_bgn;
public int Atr_end() {return atr_end;} private final int atr_end;
public byte[] Key() {return key;} private final byte[] key;
public int Val_bgn() {return val_bgn;} private final int val_bgn;
public int Val_end() {return val_end;} private final int val_end;
public boolean Val_dat_exists() {return val_end != -1;}
public boolean Val_dat_missing() {return val_end == -1;}
public byte[] Val() {
if (val == null)
val = Bry_.Mid(src, val_bgn, val_end);
return val;
} private byte[] val;
public void Html__add(Bry_bfr bfr) {
if (val_end > val_bgn)
bfr.Add_mid(src, val_bgn, val_end);
}
public void Bfr_arg__add(Bry_bfr bfr) {
if (Val_dat_exists())
bfr.Add_mid(src, val_bgn, val_end);
}
public static final Gfh_atr Noop = new Gfh_atr(-1, -1, -1, Bry_.Empty, Bry_.Empty, Bry_.Empty, -1, -1);
}

53
400_xowa/src/gplx/langs/htmls/docs/Gfh_doc_parser.java Normal file
View File

@@ -0,0 +1,53 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
import gplx.core.btries.*;
public class Gfh_doc_parser {
private final Btrie_slim_mgr trie = Btrie_slim_mgr.cs();
private final Gfh_txt_wkr txt_wkr;
public Gfh_doc_parser(Gfh_txt_wkr txt_wkr, Gfh_doc_wkr... wkr_ary) {
this.txt_wkr = txt_wkr;
for (Gfh_doc_wkr wkr : wkr_ary)
trie.Add_obj(wkr.Hook(), wkr);
}
public void Parse(byte[] page_url, byte[] src, int src_bgn, int src_end) {
int txt_bgn = -1;
int pos = src_bgn;
while (pos < src_end) {
Object o = trie.Match_bgn(src, pos, src_end);
if (o == null) { // not a known hook; add to txt
if (txt_bgn == -1) txt_bgn = pos;
++pos;
}
else { // known hook
if (txt_bgn != -1) { // txt pending; handle it
txt_wkr.Parse(txt_bgn, pos);
txt_bgn = -1;
}
Gfh_doc_wkr wkr = (Gfh_doc_wkr)o;
try {pos = wkr.Parse(src, src_bgn, src_end, pos);}
catch (Exception e) {
Gfh_utl.Log(e, "html parse failed", page_url, src, pos);
txt_bgn = pos; // set txt_bgn to hook_bgn which is "pos"; i.e.: txt resumes from start of failed hook
pos = trie.Match_pos(); // set pos to hook_end
}
}
}
if (txt_bgn != -1) txt_wkr.Parse(txt_bgn, src_end); // handle add pending txt at EOS
}
}

22
400_xowa/src/gplx/langs/htmls/docs/Gfh_doc_wkr.java Normal file
View File

@@ -0,0 +1,22 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
public interface Gfh_doc_wkr {
byte[] Hook();
int Parse(byte[] src, int src_bgn, int src_end, int pos);
}

154
400_xowa/src/gplx/langs/htmls/docs/Gfh_tag.java Normal file
View File

@@ -0,0 +1,154 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
import gplx.xowa.parsers.htmls.*; import gplx.langs.htmls.styles.*; import gplx.langs.htmls.clses.*;
public class Gfh_tag implements Mwh_atr_wkr {
private Gfh_tag_rdr tag_rdr;
private Ordered_hash atrs_hash; private boolean atrs_null; private int atrs_bgn, atrs_end;
public Gfh_tag Init(Gfh_tag_rdr tag_rdr, byte[] src, boolean tag_is_tail, boolean tag_is_inline, int src_bgn, int src_end, int atrs_bgn, int atrs_end, int name_id, byte[] name_bry) {
this.tag_rdr = tag_rdr; this.src = src; this.atrs_null = true;
this.tag_is_tail = tag_is_tail; this.tag_is_inline = tag_is_inline;
this.atrs_bgn = atrs_bgn; this.atrs_end = atrs_end;
this.name_id = name_id; this.name_bry = name_bry; this.src_bgn = src_bgn; this.src_end = src_end;
return this;
}
public Gfh_tag Copy() {
Gfh_tag rv = new Gfh_tag().Init(tag_rdr, src, tag_is_tail, tag_is_inline, src_bgn, src_end, atrs_bgn, atrs_end, name_id, name_bry);
rv.atrs_null = false;
rv.atrs_hash = Copy(atrs_hash);
return rv;
}
public int Name_id() {return name_id;} private int name_id;
public byte[] Name_bry() {return name_bry;} private byte[] name_bry;
public Gfh_tag Chk_name_or_fail(int chk) {
if (!Chk_name(chk)) tag_rdr.Err_wkr().Fail("name_id chk failed", "expecting", Gfh_tag_.To_str(chk));
return this;
}
public boolean Chk_name(int chk) {
return ( chk == name_id
|| (name_id != Gfh_tag_.Id__eos && Int_.In(chk, Gfh_tag_.Id__any, Gfh_tag_.Id__comment)));
}
public boolean Chk(int chk_name, byte[] chk_cls) {return name_id == chk_name && Atrs__cls_has(chk_cls);}
public byte[] Src() {return src;} private byte[] src;
public int Src_bgn() {return src_bgn;} private int src_bgn;
public int Src_end() {return src_end;} private int src_end;
public boolean Src_exists() {return src_end > src_bgn;} // NOTE: only true if EOS where src_end == src_bgn == src_len
public boolean Tag_is_tail() {return tag_is_tail;} private boolean tag_is_tail;
public boolean Tag_is_inline() {return tag_is_inline;} private boolean tag_is_inline;
public Ordered_hash Atrs__hash() {if (atrs_null) Atrs__make(); return atrs_hash;}
public int Atrs__len() {if (atrs_null) Atrs__make(); return atrs_hash.Count();}
public boolean Atrs__match_pair(byte[] key, byte[] val) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(key);
return rv == null ? false : Bry_.Eq(val, rv.Val());
}
public boolean Atrs__cls_has(byte[] val) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(Gfh_atr_.Bry__class); if (rv == null) return false;
byte[] rv_val = rv.Val();
return Gfh_class_.Has(rv_val, 0, rv_val.length, val);
}
public boolean Atrs__cls_eq(byte[] val) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(Gfh_atr_.Bry__class); if (rv == null) return false;
return Bry_.Eq(val, rv.Val());
}
public byte Atrs__cls_find_or_fail(Hash_adp_bry hash) {
byte rv = Atrs__cls_find_or(hash, Byte_.Max_value_127); if (rv == Byte_.Max_value_127) tag_rdr.Err_wkr().Fail("cls missing");
return rv;
}
public byte Atrs__cls_find_or(Hash_adp_bry hash, byte or) {
if (atrs_null) Atrs__make();
Gfh_atr cls_atr = (Gfh_atr)atrs_hash.Get_by(Gfh_atr_.Bry__class); if (cls_atr == null) return or;
byte rv = Gfh_class_.Find_1st(src, cls_atr.Val_bgn(), cls_atr.Val_end(), hash); if (rv == Byte_.Max_value_127) return or;
return rv;
}
private static final Gfh_style_wkr__val_as_int style_wkr = new Gfh_style_wkr__val_as_int();
public int Atrs__style_get_as_int(byte[] key) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(Gfh_atr_.Bry__style); if (rv == null) return -1;
byte[] rv_val = rv.Val();
return style_wkr.Parse(rv_val, 0, rv_val.length, key);
}
public byte[] Atrs__get_as_bry(byte[] key) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(key);
return rv == null ? Bry_.Empty : rv.Val();
}
public int Atrs__get_as_int(byte[] key) {
int rv = Atrs__get_as_int_or(key, Int_.Min_value); if (rv == Int_.Min_value) tag_rdr.Err_wkr().Fail("atr missing", "key", key);
return rv;
}
public int Atrs__get_as_int_or(byte[] key, int or) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(key); if (rv == null) return or;
return Bry_.To_int_or(src, rv.Val_bgn(), rv.Val_end(), or);
}
public Gfh_atr Atrs__get_at(int i) {return (Gfh_atr)atrs_hash.Get_at(i);}
public Gfh_atr Atrs__get_by_or_fail(byte[] key) {return Atrs__get_by_or_fail(key, Bool_.Y);}
public Gfh_atr Atrs__get_by_or_empty(byte[] key) {return Atrs__get_by_or_fail(key, Bool_.N);}
public Gfh_atr Atrs__get_by_or_fail(byte[] key, boolean fail_if_null) {
if (atrs_null) Atrs__make();
Gfh_atr rv = (Gfh_atr)atrs_hash.Get_by(key);
if (rv == null) {
if (fail_if_null) tag_rdr.Err_wkr().Fail("atr missing", "key", key);
else return Gfh_atr.Noop;
}
return rv;
}
public String Atrs__print() {
if (atrs_null) Atrs__make();
Bry_bfr bfr = Bry_bfr.new_();
int len = atrs_hash.Count();
for (int i = 0; i < len; ++i) {
Gfh_atr atr = (Gfh_atr)atrs_hash.Get_at(i);
bfr.Add(atr.Key()).Add_byte_eq().Add(atr.Val()).Add_byte_nl();
}
return bfr.To_str();
}
private void Atrs__make() {
atrs_null = false;
if (atrs_hash == null) atrs_hash = Ordered_hash_.New_bry();
else atrs_hash.Clear();
tag_rdr.Atrs__make(this, atrs_bgn, atrs_end);
}
public void On_atr_each (Mwh_atr_parser mgr, byte[] src, int nde_tid, boolean valid, boolean repeated, boolean key_exists, byte[] key_bry, byte[] val_bry_manual, int[] itm_ary, int itm_idx) {
if (!valid) return;
byte[] val_bry = val_bry_manual;
int val_bgn = -1, val_end = -1;
int atr_bgn = itm_ary[itm_idx + Mwh_atr_mgr.Idx_atr_bgn];
int atr_end = itm_ary[itm_idx + Mwh_atr_mgr.Idx_atr_end];
if (key_exists) {
val_bgn = itm_ary[itm_idx + Mwh_atr_mgr.Idx_val_bgn];
val_end = itm_ary[itm_idx + Mwh_atr_mgr.Idx_val_end];
}
else
val_bry = key_bry;
Gfh_atr atr = new Gfh_atr(atrs_hash.Count(), atr_bgn, atr_end, key_bry, val_bry, src, val_bgn, val_end);
atrs_hash.Add(key_bry, atr);
}
private static Ordered_hash Copy(Ordered_hash src) {
Ordered_hash rv = Ordered_hash_.New();
int len = src.Count();
for (int i = 0; i < len; ++i) {
Gfh_atr atr = (Gfh_atr)src.Get_at(i);
rv.Add(atr.Key(), atr);
}
return rv;
}
}

223
400_xowa/src/gplx/langs/htmls/docs/Gfh_tag_rdr.java Normal file
View File

@@ -0,0 +1,223 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
import gplx.core.primitives.*; import gplx.core.brys.*; import gplx.core.btries.*;
import gplx.xowa.parsers.htmls.*;
public class Gfh_tag_rdr {
private final Hash_adp_bry name_hash = Gfh_tag_.Hash;
private final Mwh_atr_parser atr_parser = new Mwh_atr_parser();
private final Gfh_tag tag__tmp__move = new Gfh_tag(), tag__tmp__peek = new Gfh_tag(), tag__eos = new Gfh_tag(), tag__comment = new Gfh_tag();
private final Int_obj_ref tmp_depth = Int_obj_ref.zero_();
public byte[] Src() {return src;} private byte[] src;
public int Src_end() {return src_end;} private int src_end;
public Bry_err_wkr Err_wkr() {return err_wkr;} private final Bry_err_wkr err_wkr = new Bry_err_wkr();
public void Init(byte[] ctx, byte[] src, int src_bgn, int src_end) {
this.src = src; this.pos = src_bgn; this.src_end = src_end;
tag__eos.Init(this, src, Bool_.N, Bool_.N, src_end, src_end, src_end, src_end, Gfh_tag_.Id__eos, Bry_.Empty);
err_wkr.Init_by_page(String_.new_u8(ctx), src);
}
public int Pos() {return pos;} private int pos;
public void Pos_(int v) {this.pos = v;}
public void Atrs__make(Mwh_atr_wkr atr_wkr, int head_bgn, int head_end) {atr_parser.Parse(atr_wkr, -1, -1, src, head_bgn, head_end);}
public Gfh_tag Tag__move_fwd_head() {return Tag__find(Bool_.Y, Bool_.N, Bool_.N, pos, src_end, Gfh_tag_.Id__any);}
public Gfh_tag Tag__move_fwd_head(int match_name_id) {return Tag__find(Bool_.Y, Bool_.N, Bool_.N, pos, src_end, match_name_id);}
public Gfh_tag Tag__move_fwd_tail(int match_name_id) {return Tag__find(Bool_.Y, Bool_.N, Bool_.Y, pos, src_end, match_name_id);}
public Gfh_tag Tag__peek_fwd_head() {return Tag__find(Bool_.N, Bool_.N, Bool_.N, pos, src_end, Gfh_tag_.Id__any);}
public Gfh_tag Tag__peek_fwd_head(int match_name_id) {return Tag__find(Bool_.N, Bool_.N, Bool_.N, pos, src_end, match_name_id);}
public Gfh_tag Tag__peek_fwd_tail(int match_name_id) {return Tag__find(Bool_.N, Bool_.N, Bool_.Y, pos, src_end, match_name_id);}
public Gfh_tag Tag__peek_bwd_tail(int match_name_id) {return Tag__find(Bool_.N, Bool_.Y, Bool_.Y, pos, src_end, match_name_id);}
public Gfh_tag Tag__peek_bwd_head() {return Tag__find(Bool_.N, Bool_.Y, Bool_.Y, pos, src_end, Gfh_tag_.Id__any);}
public Gfh_tag Tag__find_fwd_head(int bgn, int end, int match_name_id) {return Tag__find(Bool_.N, Bool_.N, Bool_.N, bgn, end, match_name_id);}
private Gfh_tag Tag__find(boolean move, boolean bwd, boolean tail, int rng_bgn, int rng_end, int match_name_id) {
int tmp = rng_bgn;
int stop_pos = rng_end; int adj = 1;
if (bwd) {
stop_pos = -1;
adj = -1;
--tmp; // subtract 1 from tmp; needed when pos is at src_len, else array error below
}
tmp_depth.Val_zero_();
Gfh_tag rv = null;
while (tmp != stop_pos) {
if (src[tmp] == Byte_ascii.Angle_bgn) {
rv = Tag__extract(move, tail, match_name_id, tmp);
if (Tag__match(move, bwd, tail, match_name_id, tmp_depth, rv))
break;
else {
tmp = bwd ? rv.Src_bgn() - 1 : rv.Src_end();
rv = null;
}
}
else
tmp += adj;
}
if (rv == null) {
if (move && tail && !bwd)
err_wkr.Fail("move failed", "tag_name", Gfh_tag_.To_str(match_name_id));
else
return Tag__eos(rng_bgn);
}
if (move) pos = rv.Src_end();
return rv;
}
private boolean Tag__match(boolean move, boolean bwd, boolean tail, int match_name_id, Int_obj_ref depth_obj, Gfh_tag tag) {
int tag_name_id = tag.Name_id();
if ( tag_name_id != match_name_id // tag doesn't match requested
&& match_name_id != Gfh_tag_.Id__any // requested is not wildcard
) return false;
if (tag_name_id == Gfh_tag_.Id__comment) {
if (match_name_id == Gfh_tag_.Id__comment)
return true;
else
return false;
}
int depth = depth_obj.Val();
boolean tag_is_tail = tag.Tag_is_tail();
if (tail == tag_is_tail) {
if (depth == 0)
return true;
else {
if (match_name_id == tag_name_id)
depth_obj.Val_add(-1);
return false;
}
}
else {
if (!bwd && tail && !tag_is_tail && !tag.Tag_is_inline()) {
if (match_name_id == tag_name_id)
depth_obj.Val_add(1);
return false;
}
else
return false;
}
}
public Gfh_tag Tag__extract(boolean move, boolean tail, int match_name_id, int tag_bgn) {
int name_bgn = tag_bgn + 1; if (name_bgn == src_end) return Tag__eos(tag_bgn); // EX: "<EOS"
byte name_0 = src[name_bgn];
boolean cur_is_tail = false;
switch (name_0) {
case Byte_ascii.Bang:
if (Bry_.Match(src, name_bgn + 1, name_bgn + 3, Bry__comment__mid)) // skip comment; EX: "<!"
return Tag__comment(tag_bgn);
break;
case Byte_ascii.Slash:
++name_bgn; if (name_bgn == src_end) return Tag__eos(tag_bgn); // EX: "</EOS"
name_0 = src[name_bgn];
cur_is_tail = true;
break;
}
int name_end = -1, atrs_end = -1, tag_end = -1, name_pos = name_bgn;
byte name_byte = name_0; boolean inline = false;
boolean loop = true;
while (true) {
switch (name_byte) {
case Byte_ascii.Angle_end: // EX: "<a>"
name_end = atrs_end = name_pos;
tag_end = name_end + 1;
loop = false;
break;
case Byte_ascii.Slash: // EX: "<a/>"
name_end = name_pos;
tag_end = name_pos + 1; if (tag_end == src_end) return Tag__eos(tag_bgn);// EX: "<a/EOS"
if (src[tag_end] == Byte_ascii.Angle_end) {
atrs_end = name_end;
inline = true;
loop = false;
}
else {
name_end = tag_end = -1;
}
break;
case Byte_ascii.Tab: case Byte_ascii.Nl: case Byte_ascii.Cr: case Byte_ascii.Space:
name_end = name_pos;
loop = false;
break;
}
if (!loop) break;
++name_pos; if (name_pos == src_end) return Tag__eos(tag_bgn); // EX: "<abEOS"
name_byte = src[name_pos];
}
if (tag_end == -1) {
tag_end = Bry_find_.Find_fwd(src, Byte_ascii.Angle_end, name_end, src_end);
if (tag_end == Bry_find_.Not_found) return Tag__eos(tag_bgn);
int prv_pos = tag_end - 1;
if (src[prv_pos] == Byte_ascii.Slash) {
atrs_end = prv_pos;
inline = true;
}
else
atrs_end = tag_end;
++tag_end; // position after ">"
}
Gfh_tag tmp = move ? tag__tmp__move : tag__tmp__peek;
return tmp.Init(this, src, cur_is_tail, inline, tag_bgn, tag_end, name_end, atrs_end, name_hash.Get_as_int_or(src, name_bgn, name_end, -1), Bry_.Mid(src, name_bgn, name_end));
}
public boolean Read_and_move(byte match) {
byte b = src[pos];
if (b == match) {
++pos;
return true;
}
else
return false;
}
public int Read_int_to(byte to_char) {
int rv = Read_int_to(to_char, Int_.Max_value); if (rv == Int_.Max_value) err_wkr.Fail("invalid int", "pos", pos);
return rv;
}
public int Read_int_to(byte to_char, int or_int) {
int bgn = pos;
int rv = 0;
int negative = 1;
while (pos < src_end) {
byte b = src[pos++];
switch (b) {
case Byte_ascii.Num_0: case Byte_ascii.Num_1: case Byte_ascii.Num_2: case Byte_ascii.Num_3: case Byte_ascii.Num_4:
case Byte_ascii.Num_5: case Byte_ascii.Num_6: case Byte_ascii.Num_7: case Byte_ascii.Num_8: case Byte_ascii.Num_9:
rv = (rv * 10) + (b - Byte_ascii.Num_0);
break;
case Byte_ascii.Dash:
if (negative == -1) // 2nd negative
return or_int; // return or_int
else // 1st negative
negative = -1; // flag negative
break;
default: {
boolean match = b == to_char;
if (to_char == Byte_ascii.Null) {// hack for Read_int_to_non_num
--pos;
match = true;
}
return match ? rv * negative : or_int;
}
}
}
return bgn == pos ? or_int : rv * negative;
}
private Gfh_tag Tag__comment(int tag_bgn) {
int tag_end = Bry_find_.Move_fwd(src, gplx.langs.htmls.Gfh_tag_.Comm_end, tag_bgn, src_end); if (tag_end == Bry_find_.Not_found) tag_end = src_end;
return tag__comment.Init(this, src, Bool_.N, Bool_.N, tag_bgn, tag_end, tag_end, tag_end, Gfh_tag_.Id__comment, Bry_.Empty);
}
private Gfh_tag Tag__eos(int tag_bgn) {
int tag_end = tag_bgn + 255; if (tag_end > src_end) tag_end = src_end;
return tag__comment.Init(this, src, Bool_.N, Bool_.N, tag_bgn, tag_end, tag_end, tag_end, Gfh_tag_.Id__eos, Bry_.Empty);
}
private static final byte[] Bry__comment__mid = Bry_.new_a7("--");
}

77
400_xowa/src/gplx/langs/htmls/docs/Gfh_tag_rdr_tst.java Normal file
View File

@@ -0,0 +1,77 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
import org.junit.*;
public class Gfh_tag_rdr_tst {
private final Gfh_tag_rdr_fxt fxt = new Gfh_tag_rdr_fxt();
@Test public void Basic() {
fxt.Init("1<div id='1'>2</div>3<div id='2'>4</div>5<div id='3'>6</div>7");
fxt.Test__move_fwd_head("<div id='1'>"); fxt.Test__pos("2");
fxt.Test__peek_fwd_head("<div id='2'>"); fxt.Test__pos("2");
fxt.Test__move_fwd_head("<div id='2'>"); fxt.Test__pos("4");
fxt.Test__peek_bwd_tail("</div>3") ; fxt.Test__pos("4");
}
@Test public void Comment() {
fxt.Init("1<!--2-->3<!--4-->5<div id='1'>6</div>");
fxt.Test__move_fwd_head(Gfh_tag_.Id__comment , "<!--2-->") ; fxt.Test__pos("3");
fxt.Test__move_fwd_head(Gfh_tag_.Id__any , "<div id='1'>") ; fxt.Test__pos("6");
}
@Test public void Meta() {
fxt.Init("<!DOCTYPE html>1<div id='1'>2</div>3");
fxt.Test__move_fwd_head(Gfh_tag_.Id__div , "<div id='1'>") ; fxt.Test__pos("2");
}
@Test public void Recursive__same_tags() {
fxt.Init("1<a>2<a>3</a>4</a>5");
fxt.Test__move_fwd_head(Gfh_tag_.Id__a , "<a>") ; fxt.Test__pos("2");
fxt.Test__move_fwd_tail(Gfh_tag_.Id__a , "</a>") ; fxt.Test__pos("5");
}
@Test public void Recursive__diff_tags() {
fxt.Init("1<div>2<a>3<img/>4</a>5</div>6");
fxt.Test__move_fwd_head(Gfh_tag_.Id__div , "<div>") ; fxt.Test__pos("2");
fxt.Test__move_fwd_tail(Gfh_tag_.Id__div , "</div>") ; fxt.Test__pos("6");
}
}
class Gfh_tag_rdr_fxt {
private final Gfh_tag_rdr rdr = new Gfh_tag_rdr();
public void Init(String src_str) {
byte[] src_bry = Bry_.new_u8(src_str);
rdr.Init(Bry_.Empty, src_bry, 0, src_bry.length);
}
public void Test__move_fwd_head(String expd) {Test__move_fwd_head(Gfh_tag_.Id__any, expd);}
public void Test__move_fwd_head(int match_name_id, String expd) {
Gfh_tag actl_tag = rdr.Tag__move_fwd_head(match_name_id).Chk_name_or_fail(match_name_id);
Tfds.Eq_str(expd, String_.new_u8(rdr.Src(), actl_tag.Src_bgn(), actl_tag.Src_end()));
}
public void Test__move_fwd_tail(int match_name_id, String expd) {
Gfh_tag actl_tag = rdr.Tag__move_fwd_tail(match_name_id);
Tfds.Eq_str(expd, String_.new_u8(rdr.Src(), actl_tag.Src_bgn(), actl_tag.Src_end()));
}
public void Test__peek_fwd_head(String expd) {
Gfh_tag actl_tag = rdr.Tag__peek_fwd_head();
Tfds.Eq_str(expd, String_.new_u8(rdr.Src(), actl_tag.Src_bgn(), actl_tag.Src_end()));
}
public void Test__peek_bwd_tail(String expd_str) {
byte[] expd_bry = Bry_.new_u8(expd_str);
Gfh_tag actl_tag = rdr.Tag__peek_bwd_tail(-1);
Tfds.Eq_bry(expd_bry, Bry_.Mid(rdr.Src(), actl_tag.Src_bgn(), actl_tag.Src_bgn() + expd_bry.length));
}
public void Test__pos(String expd_str) {
byte[] expd_bry = Bry_.new_u8(expd_str);
Tfds.Eq_bry(expd_bry, Bry_.Mid(rdr.Src(), rdr.Pos(), rdr.Pos() + expd_bry.length));
}
}

21
400_xowa/src/gplx/langs/htmls/docs/Gfh_txt_wkr.java Normal file
View File

@@ -0,0 +1,21 @@
/*
XOWA: the XOWA Offline Wiki Application
Copyright (C) 2012 gnosygnu@gmail.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU Affero General Public License as
published by the Free Software Foundation, either version 3 of the
License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU Affero General Public License for more details.
You should have received a copy of the GNU Affero General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
package gplx.langs.htmls.docs; import gplx.*; import gplx.langs.*; import gplx.langs.htmls.*;
public interface Gfh_txt_wkr {
void Parse(int rng_bgn, int rng_end);
}