RFR: Minor changes

[Aleksey Shipilev]

On Thu, 8 Oct 2020 09:54:23 GMT, Sergey Ponomarev <github.com+415502+stokito at openjdk.org> wrote:

> Some non critical changes:
> 1. Currently the repo has `.hgignore` file but don't have a `.gitignore`. I copied it from `.hgignore` but not sure if
> we can remove it as not used. 2. I shamelessly added my IdeJOL plugin to README. There is also mentioned that we can
> create an issue here on GitHub but currently the Issues tab is disabled. The same is for JMH repo. Can you open the
> issues tab? 3. I noted that in `StringCompress` we have a minor bug where in println format pattern we have less
> placeholders than params. While investigated it turned out that it will be better to swap "model" and "hprof file"
> columns for better output. Now it looks like: "total", "String", "String+bool",    "String+oop",  "char[]-2b",
> "char[]-1b", "char[]-1b-comp", "savings(same)", "savings(bool)",
> "savings(oop)",                                                                "model", "hprof file" 3918376,
> 59992,         59992,           59992,            0,           0,               0,           0.000,
> 0.000,           0.000,                     "VM Layout Simulation (X32 model, 8-byte aligned)", "~/empty.hprof"
> 4323184,   119984,        119984,          119984,            0,           0,               0,
> 0.000,           0.000,           0.000,                     "VM Layout Simulation (X64 model, 8-byte aligned)",
> "~/empty.hprof" 4056392,   119984,        119984,          119984,            0,           0,
> 0,           0.000,           0.000,           0.000,   "VM Layout Simulation (X64 model (compressed oops), 8-byte
> aligned)", "~/empty.hprof" 4116336,   119984,        119984,          119984,            0,           0,
> 0,           0.000,           0.000,           0.000,  "VM Layout Simulation (X64 model (compressed oops), 16-byte
> aligned)", "~/empty.hprof"  Also I tried to add a unit test for heap dump analyzer because it currently doesn't have
> any coverage. The test should analyze a heap dump and compare results. But got few problems that want to discuss and
> since there is a closed Issues tab then I'll write it here. First if all I made a heap dump of a "Hello World"
> application on JDK14 and the resulted file even gzipped is more than 900Kb.  I don't want to add such a big file into
> test resources. So here I see two solutions: 1. Create a synthetic heap dump with only few rows. 2. Start a small
> application and get it's thread dump. This wont work because each time we'll get some classes changed. Can you create a
> small heap dump file for a testing?  Another problem I faced when tried to test `StringComprees` with my empty thread
> dump and got an exception `org.openjdk.jol.heap.HeapDumpException: String.value array 34355822528 is not char[] in
> empty.hprof.gz, skipping` This is probably a bug in the JOL itself so I'll attach the
> [empty.hprof.gz](https://github.com/openjdk/jol/files/5346882/empty.hprof.gz) for you to reproduce.

This looks fine, with minor comments. _In the future_, please don't hesitate to split PRs by area and have a
descriptive synopsis for them. For now, changing the synopsis to something like: "Cleanups: README typos and links,
.gitignore, leftover JDK migrations, column widths" is enough.

Say "/integrate" once bots clear the pre-integration checks.

(The previous review should have been explicit approval)

README.md line 166:

> 164:
> 165: * [IntelliJ IDEA JOL Plugin](https://github.com/stokito/IdeaJol) can estimate object size and has an inspection to
> find heavy classes 166: * Heap dump `*.hprof` files analysers:

Heap dump decoding is very experimental in JOL, and hardly a target for current development. So listing heap dump
analyzers looks excessive.

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

PR: https://git.openjdk.java.net/jol/pull/7Marked as reviewed by shade (Committer).