Bug #20150: Memory leak in grapheme clusters

• Status: Closed
• Assignee: -
• Target version: -
• Backport: 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: DONE

[peterzhu2118 (Peter Zhu)]

GitHub PR: https://github.com/ruby/ruby/pull/9414

String#grapheme_cluters and String#each_grapheme_cluster leaks memory because if the string is not UTF-8, then the created regex will not be freed.

For example:

str = "hello world".encode(Encoding::UTF_32LE)

10.times do
  1_000.times do
    str.grapheme_clusters
  end

  puts `ps -o rss= -p #{$$}`
end

Before:

26000
42256
59008
75792
92528
109232
125936
142672
159392
176160

After:

9264
9504
9808
10000
10128
10224
10352
10544
10704
10896

[duerst (Martin Dürst)]

• Related to Feature #19908: Update to Unicode 15.1 added

[jeremyevans0 (Jeremy Evans)]

• Status changed from Open to Closed

Fixed by b3d612804946e841e47d14e09b6839224a79c1a4

[nagachika (Tomoyuki Chikanaga)]

• Backport changed from 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: REQUIRED to 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: REQUIRED

ruby_3_2 b4f8623441a8be53b643fed826ba44e933cafd7e merged revision(s) b3d612804946e841e47d14e09b6839224a79c1a4.

Hello everybody (but in particular Tomoyuki Chikanaga and Yui Naruse),

On 2024-01-18 12:21, nagachika (Tomoyuki Chikanaga) via ruby-core wrote:

Issue #20150 has been updated by nagachika (Tomoyuki Chikanaga).

Backport changed from 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: REQUIRED, 3.3: REQUIRED to 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: REQUIRED

I was under the impression that backports of bug fixes had to "trickle
down", i.e. first being applied in the main branch, then 3.3, then 3.2,
and so on (of course unless they were not needed for a specific branch).

The above "3.2: DONE, 3.3: REQUIRED" shows that the backport first
occurred in 3.2, before 3.3.

Can somebody please confirm or restate the actual backport policy now in
effect?

Thanks and regards, Martin.

ruby_3_2 b4f8623441a8be53b643fed826ba44e933cafd7e merged revision(s) b3d612804946e841e47d14e09b6839224a79c1a4.

---

Bug #20150: Memory leak in grapheme clusters
https://bugs.ruby-lang.org/issues/20150#change-106310

• Author: peterzhu2118 (Peter Zhu)
• Status: Closed
• Priority: Normal
• Backport: 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: REQUIRED

---

GitHub PR: https://github.com/ruby/ruby/pull/9414

String#grapheme_cluters and String#each_grapheme_cluster leaks memory because if the string is not UTF-8, then the created regex will not be freed.

---

[nagachika (Tomoyuki Chikanaga)]

Hello, Martin-sensei.

In my understandings, there's no explicit rule regarding the order of backporting to each stable branch.
In this case, I backported the changeset to the 3.2 branch ahead of the 3.3 branch because I hoped to include some obvious bug-fixes in ruby-3.2.3 released yesterday. I also think these fixes should be backported to 3.3 branch before release of ruby-3.3.1, but it's up to naruse-san, the current 3.3 branch maintainer.

Best Regards,

Hello Tomoyuki,

Many thanks for your careful explanation!

Regards, Martin.

On 2024-01-19 17:14, nagachika (Tomoyuki Chikanaga) via ruby-core wrote:

Issue #20150 has been updated by nagachika (Tomoyuki Chikanaga).

Hello, Martin-sensei.

In my understandings, there's no explicit rule regarding the order of backporting to each stable branch.
In this case, I backported the changeset to the 3.2 branch ahead of the 3.3 branch because I hoped to include some obvious bug-fixes in ruby-3.2.3 released yesterday. I also think these fixes should be backported to 3.3 branch before release of ruby-3.3.1, but it's up to naruse-san, the current 3.3 branch maintainer.

Best Regards,

---

Bug #20150: Memory leak in grapheme clusters
https://bugs.ruby-lang.org/issues/20150#change-106342

• Author: peterzhu2118 (Peter Zhu)
• Status: Closed
• Priority: Normal
• Backport: 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: REQUIRED

---

GitHub PR: https://github.com/ruby/ruby/pull/9414

String#grapheme_cluters and String#each_grapheme_cluster leaks memory because if the string is not UTF-8, then the created regex will not be freed.

For example:

str = "hello world".encode(Encoding::UTF_32LE)

10.times do
   1_000.times do
     str.grapheme_clusters
   end

   puts `ps -o rss= -p #{$$}`
end

Before:

26000
42256
59008
75792
92528
109232
125936
142672
159392
176160

After:

9264
9504
9808
10000
10128
10224
10352
10544
10704
10896

---

[naruse (Yui NARUSE)]

• Backport changed from 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: REQUIRED to 3.0: UNKNOWN, 3.1: REQUIRED, 3.2: DONE, 3.3: DONE

ruby_3_3 62de3eb5a2e5b1f0f1516dc99241c4c54a1bf691 merged revision(s) b3d612804946e841e47d14e09b6839224a79c1a4.