Archives/gnosygnu_xowa
1
0
Fork 0
mirror of https://github.com/gnosygnu/xowa.git synced 2024-10-27 20:34:16 +00:00
Code Issues Projects Releases Wiki Activity

Parser: Trim url-encoded space from the end of a title [#567]

Browse Source
This commit is contained in:
gnosygnu 2019-09-07 08:23:46 -04:00
parent ce5fd232d8
commit 9e302b19b5
3 changed files with 36 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
400_xowa/src/gplx/xowa/Xoa_ttl.java
View File

@ -447,7 +447,8 @@ public class Xoa_ttl { // PAGE:en.w:http://en.wikipedia.org/wiki/Help:Link; REF.
} }
if (txt_bb_len == 0) return false; if (txt_bb_len == 0) return false;
if (wik_bgn == -1 && page_bgn == txt_bb_len) return false; // if no wiki, but page_bgn is at end, then ttl is ns only; EX: "Help:"; NOTE: "fr:", "fr:Help" is allowed if (wik_bgn == -1 && page_bgn == txt_bb_len) return false; // if no wiki, but page_bgn is at end, then ttl is ns only; EX: "Help:"; NOTE: "fr:", "fr:Help" is allowed
full_txt = bfr.To_bry_and_clear(); full_txt = bfr.To_bry_and_clear_and_trim(); // trim trailing spaces; EX:A  PAGE:it.w:Italo_Toni ISSUE#:567; DATE:2019-09-06
if (full_txt.length == 0) return false; // handle  
if ( ns.Case_match() == Xow_ns_case_.Tid__1st if ( ns.Case_match() == Xow_ns_case_.Tid__1st
&& wik_bgn == -1 ) { // do not check case if xwiki; EX: "fr:" would have a wik_bgn of 0 (and a wik_end of 3); "A" (and any non-xwiki ttl) would have a wik_bgn == -1 && wik_bgn == -1 ) { // do not check case if xwiki; EX: "fr:" would have a wik_bgn of 0 (and a wik_end of 3); "A" (and any non-xwiki ttl) would have a wik_bgn == -1
byte char_1st = full_txt[page_bgn]; byte char_1st = full_txt[page_bgn];

5
400_xowa/src/gplx/xowa/wikis/ttls/Xow_ttl__html_entity_tst.java
View File

@ -16,7 +16,8 @@ Apache License: https://github.com/gnosygnu/xowa/blob/master/LICENSE-APACHE2.txt
package gplx.xowa.wikis.ttls; import gplx.*; import gplx.xowa.*; import gplx.xowa.wikis.*; package gplx.xowa.wikis.ttls; import gplx.*; import gplx.xowa.*; import gplx.xowa.wikis.*;
import org.junit.*; import gplx.xowa.wikis.nss.*; import org.junit.*; import gplx.xowa.wikis.nss.*;
public class Xow_ttl__html_entity_tst { public class Xow_ttl__html_entity_tst {
@Before public void init() {fxt.Reset();} private Xow_ttl_fxt fxt = new Xow_ttl_fxt(); private final Xow_ttl_fxt fxt = new Xow_ttl_fxt();
@Before public void init() {fxt.Reset();}
@Test public void Eacute() {fxt.Init_ttl("é").Expd_page_txt("é").Test();} //É @Test public void Eacute() {fxt.Init_ttl("é").Expd_page_txt("é").Test();} //É
@Test public void Amp_at_end() {fxt.Init_ttl("Bisc &").Expd_page_txt("Bisc &").Test();} @Test public void Amp_at_end() {fxt.Init_ttl("Bisc &").Expd_page_txt("Bisc &").Test();}
@Test public void Ncr_dec() {fxt.Init_ttl("Ab").Expd_page_txt("Ab").Test();} @Test public void Ncr_dec() {fxt.Init_ttl("Ab").Expd_page_txt("Ab").Test();}
@ -24,4 +25,6 @@ public class Xow_ttl__html_entity_tst {
// @Test public void Ncr_hex_ns() {fxt.Init_ttl("Help:A").Expd_ns_id(Xow_ns_.Tid__help).Expd_page_txt("A").Test();} // @Test public void Ncr_hex_ns() {fxt.Init_ttl("Help:A").Expd_ns_id(Xow_ns_.Tid__help).Expd_page_txt("A").Test();}
@Test public void Nbsp() {fxt.Init_ttl("A b").Expd_page_txt("A b").Test();} // NOTE: &nbsp must convert to space; EX:w:United States [[Image:Dust Bowl - Dallas, South Dakota 1936.jpg|220px|alt=]] @Test public void Nbsp() {fxt.Init_ttl("A b").Expd_page_txt("A b").Test();} // NOTE: &nbsp must convert to space; EX:w:United States [[Image:Dust Bowl - Dallas, South Dakota 1936.jpg|220px|alt=]]
@Test public void Amp() {fxt.Init_ttl("A&b").Expd_page_txt("A&b").Test();} // PURPOSE: A&B -> A&B; PAGE:en.w:Amadou Bagayoko?redirect=n; DATE:2014-09-23 @Test public void Amp() {fxt.Init_ttl("A&b").Expd_page_txt("A&b").Test();} // PURPOSE: A&B -> A&B; PAGE:en.w:Amadou Bagayoko?redirect=n; DATE:2014-09-23
@Test public void Space_only() {fxt.Init_ttl(" ").Expd_invalid_y_().Test();}
@Test public void Space_at_end() {fxt.Init_ttl("A ").Expd_page_txt("A").Test();}
} }

15
400_xowa/src/gplx/xowa/wikis/ttls/Xow_ttl_fxt.java
View File

@ -40,7 +40,7 @@ class Xow_ttl_fxt {
public Xow_ttl_fxt Expd_base_txt_wo_qarg(String v) {expd_base_txt_wo_qarg = v; return this;} private String expd_base_txt_wo_qarg; public Xow_ttl_fxt Expd_base_txt_wo_qarg(String v) {expd_base_txt_wo_qarg = v; return this;} private String expd_base_txt_wo_qarg;
public Xow_ttl_fxt Expd_leaf_txt_wo_qarg(String v) {expd_leaf_txt_wo_qarg = v; return this;} private String expd_leaf_txt_wo_qarg; public Xow_ttl_fxt Expd_leaf_txt_wo_qarg(String v) {expd_leaf_txt_wo_qarg = v; return this;} private String expd_leaf_txt_wo_qarg;
public Xow_ttl_fxt Expd_force_literal_link(int v) {expd_force_literal_link = v; return this;} private int expd_force_literal_link = -1; public Xow_ttl_fxt Expd_force_literal_link(int v) {expd_force_literal_link = v; return this;} private int expd_force_literal_link = -1;
public Xow_ttl_fxt Expd_err(Gfo_msg_itm v) {expd_err = v; return this;} private Gfo_msg_itm expd_err; public Xow_ttl_fxt Expd_invalid_y_() {expd_invalid = true; return this;} private boolean expd_invalid;
public Xowe_wiki Wiki() {return fxt.Wiki();} public Xowe_wiki Wiki() {return fxt.Wiki();}
public void Reset() { public void Reset() {
fxt.Reset(); fxt.Reset();
@ -50,13 +50,18 @@ class Xow_ttl_fxt {
expd_xwik_txt = expd_full_txt = expd_full_url = expd_page_txt = expd_page_url = expd_leaf_txt = expd_leaf_url = expd_base_txt = expd_base_url expd_xwik_txt = expd_full_txt = expd_full_url = expd_page_txt = expd_page_url = expd_leaf_txt = expd_leaf_url = expd_base_txt = expd_base_url
= expd_root_txt = expd_rest_txt = expd_talk_txt = expd_talk_url = expd_subj_txt = expd_subj_url = expd_anch_txt = expd_root_txt = expd_rest_txt = expd_talk_txt = expd_talk_url = expd_subj_txt = expd_subj_url = expd_anch_txt
= expd_base_txt_wo_qarg = expd_leaf_txt_wo_qarg = expd_qarg_txt = null; = expd_base_txt_wo_qarg = expd_leaf_txt_wo_qarg = expd_qarg_txt = null;
expd_err = null;
expd_force_literal_link = -1; expd_force_literal_link = -1;
expd_invalid = false;
fxt.Log_clear(); fxt.Log_clear();
} }
public void Test() { public void Test() {
Xoa_ttl actl = Xoa_ttl.Parse(fxt.Wiki(), Bry_.new_u8(test_raw)); Xoa_ttl actl = Xoa_ttl.Parse(fxt.Wiki(), Bry_.new_u8(test_raw));
if (expd_err == null) { if (expd_invalid) {
if (actl == null)
return;
else
Tfds.Fail("invalid expected; " + test_raw);
}
if (expd_ns_id != Int_.Min_value) Tfds.Eq(expd_ns_id, actl.Ns().Id(), "ns"); if (expd_ns_id != Int_.Min_value) Tfds.Eq(expd_ns_id, actl.Ns().Id(), "ns");
if (expd_xwik_txt != null) Tfds.Eq(expd_xwik_txt, String_.new_u8(actl.Wik_txt()), "Wiki"); if (expd_xwik_txt != null) Tfds.Eq(expd_xwik_txt, String_.new_u8(actl.Wik_txt()), "Wiki");
if (expd_page_txt != null) Tfds.Eq(expd_page_txt, String_.new_u8(actl.Page_txt()), "Page_txt"); if (expd_page_txt != null) Tfds.Eq(expd_page_txt, String_.new_u8(actl.Page_txt()), "Page_txt");
@ -80,8 +85,4 @@ class Xow_ttl_fxt {
if (expd_leaf_txt_wo_qarg != null) Tfds.Eq(expd_leaf_txt_wo_qarg, String_.new_u8(actl.Leaf_txt_wo_qarg()), "Expd_leaf_txt_wo_qarg"); if (expd_leaf_txt_wo_qarg != null) Tfds.Eq(expd_leaf_txt_wo_qarg, String_.new_u8(actl.Leaf_txt_wo_qarg()), "Expd_leaf_txt_wo_qarg");
if (expd_force_literal_link != -1) Tfds.Eq(expd_force_literal_link == 1, actl.ForceLiteralLink(), "ForceLiteralLink"); if (expd_force_literal_link != -1) Tfds.Eq(expd_force_literal_link == 1, actl.ForceLiteralLink(), "ForceLiteralLink");
} }
else {
Tfds.Eq_ary(String_.Ary(String_.new_u8(expd_err.Owner().Path()) + "." + String_.new_u8(expd_err.Key_bry())), fxt.Log_xtoAry());
}
}
} }