feat: update tmx writer, escaper, and unit test

[miurahr]

Pull request type

• Feature enhancement -> [enhancement]

Which ticket is resolved?

• TMX writer output in 6.1 should be the same as in 6.0
  • https://sourceforge.net/p/omegat/feature-requests/1749/

What does this PR change?

• update TMX test expectation for level2 write
  • use double quote for attribute
  • format indentation and empty line
• Add TmxEscapingWriterFactory class and unit test with ordinary string, "< & >", Surrogate Pair and invalid one.
• Use TmxEscapingWriterFactory in TMXWriter2 for TEXT writer

Other information

com.sun.xml.bind.marshaller#escape which is ultimately OmegaT 6.0 and before use is as follows;

public void escape(char[] ch, int start, int length, boolean isAttVal, Writer out) throws IOException {
        int limit = start + length;

        for(int i = start; i < limit; ++i) {
            char c = ch[i];
            if (c == '&' || c == '<' || c == '>' || c == '\r' || c == '\n' && isAttVal || c == '"' && isAttVal) {
                if (i != start) {
                    out.write(ch, start, i - start);
                }

                start = i + 1;
                switch (ch[i]) {
                    case '\n':
                    case '\r':
                        out.write("&#");
                        out.write(Integer.toString(c));
                        out.write(59);
                        break;
                    case '"':
                        out.write("&quot;");
                        break;
                    case '&':
                        out.write("&amp;");
                        break;
                    case '<':
                        out.write("&lt;");
                        break;
                    case '>':
                        out.write("&gt;");
                        break;
                    default:
                        throw new IllegalArgumentException("Cannot escape: '" + c + "'");
                }
            }
        }

        if (start != limit) {
            out.write(ch, start, limit - start);
        }

    }

[miurahr]

see https://github.com/FasterXML/woodstox/blob/master/src/main/java/com/ctc/wstx/sw/BufferingXmlWriter.java#L535

[miurahr]

@brandelune this PR is created by your request. Have you already lost the interest?

[miurahr]

The change here make a modification and an override of a deep and core part of XML writing, marshal. When there is a potential bug, it can break things easily. To reduce sub effect in another place, it only apply a TMXWriter2, TMX writing class. So there is no change in preferences, and other XML related features.

[brandelune]

@brandelune this PR is created by your request. Have you already lost the interest?

Thank you for the nice reminder :-)

[brandelune]

The change here make a modification and an override of a deep and core part of XML writing, marshal. When there is a potential bug, it can break things easily. To reduce sub effect in another place, it only apply a TMXWriter2, TMX writing class. So there is no change in preferences, and other XML related features.

@t-cordonnier do you have an opinion on this ?

[miurahr]

@t-cordonnier do you have an opinion on this ?

Maybe no opinion.

[miurahr]

Is it still needed?

[brandelune]

The change here make a modification and an override of a deep and core part of XML writing, marshal. When there is a potential bug, it can break things easily. To reduce sub effect in another place, it only apply a TMXWriter2, TMX writing class. So there is no change in preferences, and other XML related features.

Do you mean that there is no option in the new library to make the behavior the same as in 6.0?

I don’t think we should take risks with such overrides. So I’m still eager to have @t-cordonnier’s opinion.