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

Mw_parse: Add support for gobbling single-line comments

Browse Source
This commit is contained in:
gnosygnu 2017-01-11 08:35:16 -05:00
parent 25f74d7d80
commit 75972fcff0
3 changed files with 22 additions and 21 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
100_core/src/gplx/Bry_find_.java
View File

@ -159,14 +159,14 @@ public class Bry_find_ {
return Bry_find_.Not_found; return Bry_find_.Not_found;
} }
public static int Find_bwd__while_space_or_tab(byte[] src, int cur, int end) { // get pos of 1st char that is not \t or \s public static int Find_bwd__while_space_or_tab(byte[] src, int cur, int end) { // get pos of 1st char that is not \t or \s
if (cur >= src.length) return Bry_find_.Not_found; if (cur < 0 || cur >= src.length) return Bry_find_.Not_found;
for (int i = cur; i >= end; i--) { for (int i = cur - 1; i >= end; i--) {
byte b = src[i]; byte b = src[i];
switch (b) { switch (b) {
case Byte_ascii.Space: case Byte_ascii.Tab: case Byte_ascii.Space: case Byte_ascii.Tab:
break; break;
default: default:
return i; return i + 1;
} }
} }
return Bry_find_.Not_found; return Bry_find_.Not_found;

27
400_xowa/src/gplx/xowa/parsers/mws/prepros/Xomw_prepro_wkr.java
View File

@ -23,7 +23,6 @@ public class Xomw_prepro_wkr { // TS.UNSAFE:caching for repeated calls
private final Hash_adp_bry xmlish_elems = Hash_adp_bry.ci_a7(); private final Hash_adp_bry xmlish_elems = Hash_adp_bry.ci_a7();
private final Hash_adp_bry xmlish_allow_missing_end_tag = Hash_adp_bry.cs().Add_many_str("includeonly", "noinclude", "onlyinclude"); private final Hash_adp_bry xmlish_allow_missing_end_tag = Hash_adp_bry.cs().Add_many_str("includeonly", "noinclude", "onlyinclude");
private final Hash_adp_bry no_more_closing_tag = Hash_adp_bry.cs(); private final Hash_adp_bry no_more_closing_tag = Hash_adp_bry.cs();
// private final Btrie_slim_mgr search_dflt_trie = Btrie_slim_mgr.cs().Add_many_int(0, "[", "{", "<", "\n"); // $searchBase = "[{<\n";
private final Btrie_slim_mgr elements_trie = Btrie_slim_mgr.ci_a7(); private final Btrie_slim_mgr elements_trie = Btrie_slim_mgr.ci_a7();
private final Xomw_prepro_stack stack = new Xomw_prepro_stack(); private final Xomw_prepro_stack stack = new Xomw_prepro_stack();
private Bry_bfr accum = Bry_bfr_.New(); private Bry_bfr accum = Bry_bfr_.New();
@ -286,7 +285,11 @@ public class Xomw_prepro_wkr { // TS.UNSAFE:caching for repeated calls
// Search forwards for trailing whitespace // Search forwards for trailing whitespace
// $wsEnd will be the position of the last space (or the '>' if there's none) // $wsEnd will be the position of the last space (or the '>' if there's none)
int ws_end = Bry_find_.Find_fwd_while_space_or_tab(src, end_pos + 3, src_len); // PORTED: $wsEnd = $endPos + 2 + strspn( $text, " \t", $endPos + 3 );
int ws_end = end_pos + 2; // set pos to ">"
int ws_end2 = Bry_find_.Find_fwd_while_space_or_tab(src, end_pos + 3, src_len);
if (ws_end2 != ws_end + 1) // if ws after ">"
ws_end = ws_end2 - 1; // set to "last space"
// Keep looking forward as long as we're finding more // Keep looking forward as long as we're finding more
// comments. // comments.
@ -314,11 +317,12 @@ public class Xomw_prepro_wkr { // TS.UNSAFE:caching for repeated calls
// Remove leading whitespace from the end of the accumulator // Remove leading whitespace from the end of the accumulator
// Sanity check first though // Sanity check first though
int ws_len = i - ws_bgn; int ws_len = i - ws_bgn;
byte[] accum_bry = accum.To_bry(); if (ws_len > 0) {
if ( ws_len > 0 // PORTED:"&& strspn( $accum, " \t", -$wsLength ) === $wsLength"
&& Bry_find_.Find_fwd_while_space_or_tab(accum_bry, -ws_len, src_len) == ws_len int accum_bry_len = accum.Len();
) { int ws_end_lhs = Bry_find_.Find_bwd__while_space_or_tab(accum.Bfr(), accum_bry_len, 0);
accum.Clear().Add(Bry_.Mid(accum_bry, 0, -ws_len)); if (accum_bry_len - ws_end_lhs == ws_len)
accum.Del_by(ws_len);
} }
// Dump all but the last comment to the accumulator // Dump all but the last comment to the accumulator
@ -660,10 +664,11 @@ public class Xomw_prepro_wkr { // TS.UNSAFE:caching for repeated calls
for (int j = 0; j < parts_len; j++) { for (int j = 0; j < parts_len; j++) {
Xomw_prepro_part part = (Xomw_prepro_part)parts.Get_at(j); Xomw_prepro_part part = (Xomw_prepro_part)parts.Get_at(j);
if (part.Eqpos != -1) { if (part.Eqpos != -1) {
byte[] part_bry = part.bfr.To_bry(); Bry_bfr part_bfr = part.bfr;
byte[] arg_key = Bry_.Mid(part_bry, 0, part.Eqpos); byte[] part_bfr_bry = part_bfr.Bfr();
byte[] arg_val = Bry_.Mid(part_bry, part.Eqpos + 1); tmp_bfr.Add_str_a7("<part><name>").Add_mid(part_bfr_bry, 0, part.Eqpos);
tmp_bfr.Add_str_a7("<part><name>").Add(arg_key).Add_str_a7("</name>=<value>").Add(arg_val).Add_str_a7("</value></part>"); tmp_bfr.Add_str_a7("</name>=<value>").Add_mid(part_bfr_bry, part.Eqpos + 1, part_bfr.Len());
tmp_bfr.Add_str_a7("</value></part>");
} }
else { else {
tmp_bfr.Add_str_a7("<part><name index=\"").Add_int_variable(arg_idx).Add_str_a7("\" /><value>").Add(part.bfr.To_bry()).Add_str_a7("</value></part>"); tmp_bfr.Add_str_a7("<part><name index=\"").Add_int_variable(arg_idx).Add_str_a7("\" /><value>").Add(part.bfr.To_bry()).Add_str_a7("</value></part>");

10
400_xowa/src/gplx/xowa/parsers/mws/prepros/Xomw_prepro_wkr__tst.java
View File

@ -34,6 +34,9 @@ public class Xomw_prepro_wkr__tst {
@Test public void Comment() { @Test public void Comment() {
fxt.Test__parse("a<!--b-->c", "<root>a<comment>&lt;!--b--&gt;</comment>c</root>"); fxt.Test__parse("a<!--b-->c", "<root>a<comment>&lt;!--b--&gt;</comment>c</root>");
} }
@Test public void Comment__nl__ws() {
fxt.Test__parse("xo\n <!--1--> \n <!--2--> \nz", "<root>xo\n<comment> &lt;!--1--&gt; \n</comment><comment> &lt;!--2--&gt; \n</comment>z</root>");
}
@Test public void Ext__pre() { @Test public void Ext__pre() {
fxt.Test__parse("a<pre id=\"1\">b</pre>c", "<root>a<ext><name>pre</name><attr> id=&quot;1&quot;</attr><inner>b</inner><close>&lt;/pre&gt;</close></ext>c</root>"); fxt.Test__parse("a<pre id=\"1\">b</pre>c", "<root>a<ext><name>pre</name><attr> id=&quot;1&quot;</attr><inner>b</inner><close>&lt;/pre&gt;</close></ext>c</root>");
} }
@ -43,13 +46,6 @@ TODO:
* heading.general * heading.general
* heading.EOS: "==a" (no closing ==) * heading.EOS: "==a" (no closing ==)
* ignored tags * ignored tags
* FIX:
if ( ws_len > 0
&& Bry_find_.Find_fwd_while_space_or_tab(accum_bry, -ws_len, src_len) == ws_len
) {
accum.Clear().Add(Bry_.Mid(accum_bry, 0, -ws_len));
}
*/ */
} }
class Xomw_prepro_wkr__fxt { class Xomw_prepro_wkr__fxt {