Spring fails to determine XML is XSD-based if DOCTYPE appears in a comment

[shooye]

Affects: 5.3.14 and earlier

If you want to load a DTD configuration file, <! -- Different positions from --> will affect the loading of spring.

If my profile is written like this:

<?xml version="1.0" encoding="UTF-8"?>

<!--

DOCTYPE -->  <!--  -->
<beans xmlns="http://www.springframework.org/schema/beans"
       xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
       xsi:schemaLocation="http://www.springframework.org/schema/beans
            http://www.springframework.org/schema/beans/spring-beans-2.5.xsd">

</beans>

Then, when Spring is used to load the configuration file, an exception will appear.

This is because there is a bug when determining the profile validation pattern.

In org.springframework.util.xml.XmlValidationModeDetector,Spring analyzes programs one line at a time.

However, each analysis focuses only on the first <!-- in each line. Spring will put <!-- The previous content is not considered as a comment.
Thus, the situation in the above example is ignored.

I think that can result in org.springframework.util.xml.XmlValidationModeDetector#consumeCommentTokens(String line) return to add the following code before:

@Nullable
private String consumeCommentTokens(String line) {
    int indexOfStartComment = line.indexOf(START_COMMENT);
    if (indexOfStartComment == -1 && !line.contains(END_COMMENT)) {
        return line;
    }

    String result = "";
    String currLine = line;
    if (indexOfStartComment >= 0) {
        result = line.substring(0, indexOfStartComment);
        currLine = line.substring(indexOfStartComment);
    }

    while ((currLine = consume(currLine)) != null) {
        if (!this.inComment && !currLine.trim().startsWith(START_COMMENT)) {
            int index = result.indexOf(END_COMMENT);
            if(index != -1) result = result.substring(index + END_COMMENT.length());
            return result + currLine;
        }
    }
    return null;
}

It can be removed in front of '-->'.

[sbrannen]

This is closely related to:

• Spring fails to determine that XML is DTD-based if DTD declaration is followed by a comment #23605

Thanks for raising the issue. We'll look into it.

[sbrannen]

The proposed change results in a StringIndexOutOfBoundsException for an existing test in XmlValidationModeDetectorTests. So we'll investigate an alternative solution.

[shooye]

The proposed change results in a StringIndexOutOfBoundsException for an existing test in XmlValidationModeDetectorTests. So we'll investigate an alternative solution.

@sbrannen

Sorry, my carelessness caused StringIndexOutOfBoundsException.
I tried to modify the org.springframework.util.xml.XmlValidationModeDetector.Probably changed two or three methods:

I created 15 XML files and generated several random XML files using the program.

These files even contain the following extreme cases:

<!DOC<!-- comment -->TYPE beans PUBLIC "-//SPRING//DTD BEAN 2.0//EN" "https://www.springframework.org/dtd/spring-beans-2.0.dtd">

All files passed the test.

I package and upload the modified code and these test files.I hope my method can solve this problem.

XmlValidationModeDetector.zip

[sbrannen]

Hi @shooye,

Thanks for working on it and providing the ZIP file.

I took a look at it, and it appears to fix the issue; however, I noticed that you introduced a second boolean flag to track whether we're "in a comment", and I thought we would still only need a single flag for that.

In any case, the proper way to submit a fix is via a PR so that we can be sure you have signed the contributor agreement.

If you have time to submit a PR, that would be great. Otherwise, I will likely attempt to fix the issue myself.

Cheers,

Sam