Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upAn iframe element's name and its contained child browsing context's name interact improperly #25057
Labels
Projects
Comments
|
In an unmerged branch where the window named getter is generally able to return browsing contexts (based on #21869 but with the iframe TODOs filled in), html/browsers/windows/the-window-object/window-named-properties.html fails cases "Dynamic name" and "Ghost name" and it seems like another manifestation of the same cause as the above problem. |
pshaughn
changed the title
|
Despite the test name, html/browsers/the-window-object/named-access-on-the-window-object/window-null-names.html is actually failing for reasons relating to this (in the aforementioned branch). Changing it to use another character instead of the null-byte doesn't make it pass; the problem is about the iframe and its child browsing context not having the right name interaction, not about the null byte. |
|
I'm going to have to put a subset of html/browsers/windows/nested-browsing-contexts/name-attribute.window.html into the _mozilla tests so we can observe it passing or failing; in an automated run, the test harness times out on the embed and object cases so we don't get to see what happens to the iframe cases. |
bors-servo
added a commit
that referenced
this issue
Jan 22, 2020
Make name content attributes consistently atoms and put them in rare_data for fast access <!-- Please describe your changes on the following line: --> All codepaths setting the name content attribute now use an atom, which is also stored in rare_data for direct lookup by a get_name method. Paralleling the get_name method, I added a get_id method, which makes some internal id-lookup cases nicer. A new test tests for a name setter on every HTML element type. In addition to its overt and upstreamable purpose of checking IDL property reflection semantics, for us this test also hits some Servo assertions that make sure the name is an atom in every case. If the test doesn't crash, even a failed test case still has the attribute as an atom rather than some other type. The failed cases are for elements that we have unimplemented or completely stubbed; I added a few missing name IDL properties to otherwise implemented elements. I also made a _mozilla copy of the working part of an existing but not-working test, as I explain in #25057 (comment) --- <!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: --> - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes fix #25570 and make progress on #25057 <!-- Either: --> - [X] There are tests for these changes <!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.--> <!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
bors-servo
added a commit
that referenced
this issue
Jan 22, 2020
Make name content attributes consistently atoms and put them in rare_data for fast access <!-- Please describe your changes on the following line: --> All codepaths setting the name content attribute now use an atom, which is also stored in rare_data for direct lookup by a get_name method. Paralleling the get_name method, I added a get_id method, which makes some internal id-lookup cases nicer. A new test tests for a name setter on every HTML element type. In addition to its overt and upstreamable purpose of checking IDL property reflection semantics, for us this test also hits some Servo assertions that make sure the name is an atom in every case. If the test doesn't crash, even a failed test case still has the attribute as an atom rather than some other type. The failed cases are for elements that we have unimplemented or completely stubbed; I added a few missing name IDL properties to otherwise implemented elements. I also made a _mozilla copy of the working part of an existing but not-working test, as I explain in #25057 (comment) --- <!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: --> - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes fix #25570 and make progress on #25057 <!-- Either: --> - [X] There are tests for these changes <!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.--> <!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
bors-servo
added a commit
that referenced
this issue
Feb 12, 2020
Make name content attributes consistently atoms and put them in rare_data for fast access <!-- Please describe your changes on the following line: --> All codepaths setting the name content attribute now use an atom, which is also stored in rare_data for direct lookup by a get_name method. Paralleling the get_name method, I added a get_id method, which makes some internal id-lookup cases nicer. A new test tests for a name setter on every HTML element type. In addition to its overt and upstreamable purpose of checking IDL property reflection semantics, for us this test also hits some Servo assertions that make sure the name is an atom in every case. If the test doesn't crash, even a failed test case still has the attribute as an atom rather than some other type. The failed cases are for elements that we have unimplemented or completely stubbed; I added a few missing name IDL properties to otherwise implemented elements. --- <!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: --> - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes fix #25570 and make progress on #25057 <!-- Either: --> - [X] There are tests for these changes <!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.--> <!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
bors-servo
added a commit
that referenced
this issue
Feb 13, 2020
Make name content attributes consistently atoms and put them in rare_data for fast access <!-- Please describe your changes on the following line: --> All codepaths setting the name content attribute now use an atom, which is also stored in rare_data for direct lookup by a get_name method. Paralleling the get_name method, I added a get_id method, which makes some internal id-lookup cases nicer. A new test tests for a name setter on every HTML element type. In addition to its overt and upstreamable purpose of checking IDL property reflection semantics, for us this test also hits some Servo assertions that make sure the name is an atom in every case. If the test doesn't crash, even a failed test case still has the attribute as an atom rather than some other type. The failed cases are for elements that we have unimplemented or completely stubbed; I added a few missing name IDL properties to otherwise implemented elements. --- <!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: --> - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes fix #25570 and make progress on #25057 <!-- Either: --> - [X] There are tests for these changes <!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.--> <!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
bors-servo
added a commit
that referenced
this issue
Feb 13, 2020
Make name content attributes consistently atoms and put them in rare_data for fast access <!-- Please describe your changes on the following line: --> All codepaths setting the name content attribute now use an atom, which is also stored in rare_data for direct lookup by a get_name method. Paralleling the get_name method, I added a get_id method, which makes some internal id-lookup cases nicer. A new test tests for a name setter on every HTML element type. In addition to its overt and upstreamable purpose of checking IDL property reflection semantics, for us this test also hits some Servo assertions that make sure the name is an atom in every case. If the test doesn't crash, even a failed test case still has the attribute as an atom rather than some other type. The failed cases are for elements that we have unimplemented or completely stubbed; I added a few missing name IDL properties to otherwise implemented elements. --- <!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `___` with appropriate data: --> - [X] `./mach build -d` does not report any errors - [X] `./mach test-tidy` does not report any errors - [X] These changes fix #25570 and make progress on #25057 <!-- Either: --> - [X] There are tests for these changes <!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.--> <!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
|
#25572 fixed this. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This is the iframe case of html/browsers/windows/nested-browsing-contexts/name-attribute.window.html (#25056 for other cases). iframe.setAttribute("name","meh") causes code in the iframe to see window.name as "meh". WPT expects the name that the iframe originally had to be visible as window,name, but for subsequent iframe.setAttributes by the parent to not change window.name. Firefox fails this in the same way; other browsers pass.