Bugzilla

I'm not sure if it's a determining factor in causing the issue or not, but one interesting aspect of the testcase is that it's got *extremely* deeply nested inline-level elements, due to the testcase having a whole bunch of unescaped `<stuff>` tokens  which end up getting interpreted as HTML tags (equivalent to `<span>` effectively).  These weren't intended as HTML tags, so they're never closed, which means the nesting level gets deeper and deeper as you go further down the page.

The first such inadvertent HTML tag is from a mention of `<thread.h>`, and then the rest are mostly email addresses from replies in the quoted discussions there (e.g. `John Doe <johndoe@example.org> wrote:` -- the `<>` expression there gets interpreted as an HTML tag and adds a layer of nesting).

I'm not sure if it's a determining factor in causing the issue or not, but one interesting aspect of the testcase is that it's got *extremely* deeply nested inline-level elements, due to the testcase having a whole bunch of unescaped `<stuff>` tokens  which end up getting interpreted as HTML tags (equivalent to `<span>` effectively).  These weren't intended as HTML tags, so they're never closed, which means the nesting level gets deeper and deeper as you go further down the page.

The first such inadvertent HTML tag is from a mention of `<thread.h>`, and then the rest are mostly email addresses from replies in the quoted discussions there (e.g. `John Doe <johndoe@example.org> wrote:` -- the `<>` expression there gets interpreted as an HTML tag and adds a layer of nesting for the rest of the document that follows it).

I'm not sure if it's a determining factor in causing the issue or not, but one interesting aspect of the testcase is that it's got *extremely* deeply nested inline-level elements, due to the testcase having a whole bunch of unescaped `<stuff>` tokens  which end up getting interpreted as HTML tags (equivalent to `<span>` effectively).  These weren't intended as HTML tags, so they're never closed, which means the nesting level gets deeper and deeper as you go further down the page.

The first such inadvertent HTML tag is from a mention of `<thread.h>`, and then the rest are mostly email addresses from replies in the quoted discussions there (e.g. `John Doe <johndoe@example.org> wrote: [some quoted email]` -- the `<>` expression there gets interpreted as an HTML tag and adds a layer of nesting for the rest of the document that follows it).

I'm not sure if it's a determining factor in causing the issue or not, but one interesting aspect of the testcase is that it's got *extremely* deeply nested inline-level elements, due to the testcase having a whole bunch of unescaped `<stuff>` tokens  which end up getting interpreted as HTML tags (equivalent to `<span>` effectively).  These weren't intended as HTML tags (they're meant to be human-readable text and markup in quoted newsgroup posts), so they're never closed, which means the nesting level gets deeper and deeper as you go further down the page.

(This is essentially what roc observed in comment 35 here.)

The first such inadvertent HTML tag is from a mention of `<thread.h>`, and then the rest are mostly email addresses from replies in the quoted discussions there (e.g. `John Doe <johndoe@example.org> wrote: [some quoted email]` -- the `<>` expression there gets interpreted as an HTML tag and adds a layer of nesting for the rest of the document that follows it).