RFR: 8291801: IGV: Broken button "Hide graphs which are the same as the previous graph" [v4]

Roberto Castañeda Lozano rcastanedalo at openjdk.org
Fri Jan 20 10:24:31 UTC 2023


On Wed, 18 Jan 2023 03:00:45 GMT, Koichi Sakata <ksakata at openjdk.org> wrote:

>> This pull request makes the "Hide graphs..." button working.
>> 
>> The parser for bgv files, `BinaryParaser` class, enables the "hide graphs" function. But the xml parser, `Parser` class, doesn't have any code about the function. So I wrote code by referring to BinaryParser class.
>> 
>> # Tests
>> 
>> I tested manually. Screenshots are as follows.
>> 
>> <img width="1175" alt="スクリーンショット 2022-09-27 16 52 16" src="https://user-images.githubusercontent.com/60008/192467385-328b8298-9055-484d-813d-1c50200e2239.png">
>> 
>> In this case, there are 13 graphs and 10 graphs from "After Parsing" to "Before matching" are the same. So only 4 graphs are shown after we push the button. 
>> 
>> <img width="1175" alt="スクリーンショット 2022-09-27 16 52 23" src="https://user-images.githubusercontent.com/60008/192467331-fbeb4911-460d-476c-9445-098805c39d27.png">
>> 
>> Pushing it again or opening the graph that is hidden in the view restores the view to its original state. I attached [the graph file that I used in the test](https://github.com/openjdk/jdk/files/9653179/hello.zip).
>> 
>> Furthermore, I added a test method for the class I changed. Running `mvn test` command was successful.
>> 
>> # Concerns
>> 
>> I'm concerned about the equivalence of 2 graphs. I regard them as the same graphs when all fields in `InputGraph` class are equal. But in the test class, 2 graphs are equal when nodes, edges, blocks and blocks of each `InputNode` are equal. Here is an extract of the test code.
>> 
>> 
>> // src/utils/IdealGraphVisualizer/Data/src/test/java/com/sun/hotspot/igv/data/Util.java
>> 
>>     public static void assertGraphEquals(InputGraph a, InputGraph b) {
>> 
>>         if(!a.getNodesAsSet().equals(b.getNodesAsSet())) {
>>             fail();
>>         }
>> 
>>         if (!a.getEdges().equals(b.getEdges())) {
>>             fail();
>>         }
>> 
>>         if (a.getBlocks().equals(b.getBlocks())) {
>>             fail();
>>         }
>> 
>>         for (InputNode n : a.getNodes()) {
>>             assertEquals(a.getBlock(n), b.getBlock(n));
>>         }
>>     }
>> 
>> 
>> But opening a graph is very slow when I compare blocks of each InputNode. So I didn't add that comparison in `isSameContent` method.
>
> Koichi Sakata has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Remove extra whitespaces

Thanks for coming back to this PR, Koichi! I did some testing and it seems to work deterministically now. However, as a user I find that using a different definition of graph equality compared to the graph difference functionality is confusing. Would it be possible to use the same logic as in `com.sun.hotspot.igv.difference.Difference`? I recognize that reusing the same code might be difficult, but one could at least infer and apply the same definition of graph equality.

Here is a concrete example of the problem, on a graph from [hello.xml](https://github.com/openjdk/jdk/files/9653179/hello.zip): if I open graph 1-6, then right-click on 1-7 in the `Outline` window and select `Difference to current graph`, node `0 Root` is marked as different (since its `dump_spec` property changes from empty to `inner`). However, when I select `Hide graphs which are the same as the previous graphs`, graph 1-7 is hidden as it is determined to be equal to 1-6. Ideally, a graph should only be hidden as a duplicate if and only if diffing it to the previous graph shows no differences.

-------------

Changes requested by rcastanedalo (Reviewer).

PR: https://git.openjdk.org/jdk/pull/10440


More information about the hotspot-compiler-dev mailing list