Should enforce that tags start with ascii characters only

[jackila]

When I use Jsoup.parse('') to parse the html

<p>5.到推机制，推荐会员匹配成功超过24小时不打款帐号会被冻结，扣领导人
<会员挂单金额5%>动态做为惩罚</p>
<p><br></p>

The result come out like this:

<p>5.到推机制，推荐会员匹配成功超过24小时不打款帐号会被冻结，扣领导人
<会员挂单金额5%>
动态做为惩罚
</会员挂单金额5%></p>
<p><br /></p>

apparently <会员挂单金额5%> is not a tag ,is there any method can solve it?

[krystiangorecki]

If that's the only problematic tag you can escape it manually:

html = html.replaceAll("<会员挂单金额5%>", "&lt;会员挂单金额5%&gt;");

[ghost]

How to handle this issue if there are multiple problematic tags?

[SliverySky]

In this version, Chinese character is also considered as legal tag. e.g<一></一> this is not allow in chrome. String "<一>"will be mistaken as unclosed tag and be replace by <一></一>. The reason is that Character.isLetter() will consider all character in all language as letter.https://github.com/jhy/jsoup/pull/1390

[jhy]

Per the HTML spec, and in checking in Chrome: tags must start with an ascii character, and any character after that is allowed.

See tag open state followed by tag name state.

So, valid tags should be:

<a一>
<a会员挂单金额5>
<table(╯°□°)╯>

Invalid:

<一>
<会员挂单金额5>
<(╯°□°)╯>

[SliverySky]

Thank you for your replay which let me know the reason for this problem (why `<一>` can't work in Chrome). And I will standardize the format of PRs. Thank you for reminder!

[jhy]

Thanks, fixed now