|
1 | 1 | -*- coding: utf-8 -*- |
2 | 2 |
|
| 3 | +commit 7a5688e2a27668e48f8d6ff4af5b2208b98a2f5e |
| 4 | + Author: Takashi Kokubun < [email protected]> |
| 5 | + AuthorDate: 2025-10-08 01:42:48 +0900 |
| 6 | + Commit: Takashi Kokubun < [email protected]> |
| 7 | + CommitDate: 2025-10-08 01:42:48 +0900 |
| 8 | + |
| 9 | + v3.4.7 |
| 10 | + |
| 11 | +commit ef3ce711fb29f7273216edb2fe41bb2604c835c3 |
| 12 | + Author: Takashi Kokubun < [email protected]> |
| 13 | + AuthorDate: 2025-10-08 01:41:15 +0900 |
| 14 | + Commit: Takashi Kokubun < [email protected]> |
| 15 | + CommitDate: 2025-10-08 01:41:15 +0900 |
| 16 | + |
| 17 | + merge revision(s) 7ae67e8f6ad6e7fd0677b28a7a10961f79d55495: [Backport #21568] |
| 18 | + |
| 19 | + [PATCH] load.c: Fix dest and src of MEMMOVE |
| 20 | + |
| 21 | + When multiple files with the same name are required, the features_index |
| 22 | + hash stores the indexes in `$LOADED_FEATURES` array into a darray. |
| 23 | + The dest and src arguments for `MEMMOVE` were wrongly reversed when |
| 24 | + inserting a new index in the darray. |
| 25 | + |
| 26 | + [Bug #21568] |
| 27 | + |
| 28 | +commit a7eb7e7c6fe02f7127ff527fff18a0914a7aa888 |
| 29 | + Author: Takashi Kokubun < [email protected]> |
| 30 | + AuthorDate: 2025-10-07 02:53:24 +0900 |
| 31 | + Commit: Takashi Kokubun < [email protected]> |
| 32 | + CommitDate: 2025-10-08 01:40:55 +0900 |
| 33 | + |
| 34 | + tool/merger.rb: Fetch diff from GitHub instead of cgit |
| 35 | + |
| 36 | + Our cgit server has been shut down. |
| 37 | + |
| 38 | + tool/merger.rb: Fetch a diff in the patch format |
| 39 | + |
| 40 | + It expects "Subject:", so it needs to be a patch file. |
| 41 | + |
| 42 | +commit d2cd7eaa11f57f81cdbacf0e8d5162d37e28337b |
| 43 | + Author: Nobuyoshi Nakada < [email protected]> |
| 44 | + AuthorDate: 2025-10-07 10:30:16 +0900 |
| 45 | + Commit: Takashi Kokubun < [email protected]> |
| 46 | + CommitDate: 2025-10-08 01:15:46 +0900 |
| 47 | + |
| 48 | + Update rubyspec as of CVE-2025-27221 |
| 49 | + |
| 50 | +commit 71462666e0f1ea720062e12f97b2ea89166038d1 |
| 51 | + Author: Hiroshi SHIBATA < [email protected]> |
| 52 | + AuthorDate: 2025-10-07 18:29:36 +0900 |
| 53 | + Commit: Takashi Kokubun < [email protected]> |
| 54 | + CommitDate: 2025-10-08 01:15:46 +0900 |
| 55 | + |
| 56 | + Merge URI-1.0.4 |
| 57 | + |
| 58 | +commit 49d7295fc38583abf3018a16a2018667871d9f3d |
| 59 | + Author: Takashi Kokubun < [email protected]> |
| 60 | + AuthorDate: 2025-10-03 02:05:27 +0900 |
| 61 | + Commit: Takashi Kokubun < [email protected]> |
| 62 | + CommitDate: 2025-10-03 02:05:39 +0900 |
| 63 | + |
| 64 | + macos.yml: macOS 13 hosted runner image is closing down |
| 65 | + |
| 66 | + https://github.blog/changelog/2025-09-19-github-actions-macos-13-runner-image-is-closing-down/ |
| 67 | + |
| 68 | +commit 264b2d78c366491834c9ce4499a4ad7b7564b89d |
| 69 | + Author: Takashi Kokubun < [email protected]> |
| 70 | + AuthorDate: 2025-09-30 07:10:37 +0900 |
| 71 | + Commit: Takashi Kokubun < [email protected]> |
| 72 | + CommitDate: 2025-09-30 07:10:58 +0900 |
| 73 | + |
| 74 | + Reapply "merge revision(s) 62430c19c9f1ab49429cebe65f30588472648c95: [Backport #21342]" |
| 75 | + |
| 76 | + This reverts commit c414b9871f263331cde0af1c08cf5c1a47e1aedf. |
| 77 | + |
| 78 | +commit c414b9871f263331cde0af1c08cf5c1a47e1aedf |
| 79 | + Author: Takashi Kokubun < [email protected]> |
| 80 | + AuthorDate: 2025-09-30 07:08:47 +0900 |
| 81 | + Commit: Takashi Kokubun < [email protected]> |
| 82 | + CommitDate: 2025-09-30 07:08:47 +0900 |
| 83 | + |
| 84 | + Revert "merge revision(s) 62430c19c9f1ab49429cebe65f30588472648c95: [Backport #21342]" |
| 85 | + |
| 86 | + This reverts commit 4306c9048fb674d24b92dc46b6746a4749564147. |
| 87 | + |
| 88 | +commit 22c2262b83224f6798d60d539f1f8609d9722766 |
| 89 | + Author: Takashi Kokubun < [email protected]> |
| 90 | + AuthorDate: 2025-09-30 06:36:06 +0900 |
| 91 | + Commit: Takashi Kokubun < [email protected]> |
| 92 | + CommitDate: 2025-09-30 06:36:06 +0900 |
| 93 | + |
| 94 | + merge revision(s) 354d47ae5bc4edcc94db4a5391ed71a8b9844e57: [Backport #21569] |
| 95 | + |
| 96 | + IBF: Avoid unaligned load on 32 bit platforms |
| 97 | + |
| 98 | + [Bug #21569] |
| 99 | + |
| 100 | +commit 4306c9048fb674d24b92dc46b6746a4749564147 |
| 101 | + Author: Takashi Kokubun < [email protected]> |
| 102 | + AuthorDate: 2025-09-30 06:35:15 +0900 |
| 103 | + Commit: Takashi Kokubun < [email protected]> |
| 104 | + CommitDate: 2025-09-30 06:35:15 +0900 |
| 105 | + |
| 106 | + merge revision(s) 62430c19c9f1ab49429cebe65f30588472648c95: [Backport #21342] |
| 107 | + |
| 108 | + Properly unlock locked mutexes on thread cleanup. |
| 109 | + |
| 110 | + Mutexes were being improperly unlocked on thread cleanup. This bug was |
| 111 | + introduced in 050a8954395. |
| 112 | + |
| 113 | + We must keep a reference from the mutex to the thread, because if the fiber |
| 114 | + is collected before the mutex is, then we cannot unlink it from the thread in |
| 115 | + `mutex_free`. If it's not unlinked from the the thread when it's freed, it |
| 116 | + causes bugs in `rb_thread_unlock_all_locking_mutexes`. |
| 117 | + |
| 118 | + We now mark the fiber when a mutex is locked, and the thread is marked |
| 119 | + as well. However, a fiber can still be freed in the same GC cycle as the |
| 120 | + mutex, so the reference to the thread is still needed. |
| 121 | + |
| 122 | + The reason we need to mark the fiber is that `mutex_owned_p()` has an ABA |
| 123 | + issue where if the fiber is collected while it's locked, a new fiber could be |
| 124 | + allocated at the same memory address and we could get false positives. |
| 125 | + |
| 126 | + Fixes [Bug #21342] |
| 127 | + |
| 128 | + Co-authored-by: John Hawthorn < [email protected]> |
| 129 | + |
| 130 | +commit 88a96147a21900a2c22d3ae84d05cd3f6ee795c7 |
| 131 | + Author: Takashi Kokubun < [email protected]> |
| 132 | + AuthorDate: 2025-09-30 06:31:05 +0900 |
| 133 | + Commit: Takashi Kokubun < [email protected]> |
| 134 | + CommitDate: 2025-09-30 06:31:05 +0900 |
| 135 | + |
| 136 | + merge revision(s) ef2b26cc3eaed06c5c9d4ef2c6d8669ff357afa4: [Backport #21611] |
| 137 | + |
| 138 | + `struct iseq_catch_table` is packed |
| 139 | + |
| 140 | +commit e900e9fc9ace5e36de5170c83df0f56167e4c8a9 |
| 141 | + Author: Samuel Williams < [email protected]> |
| 142 | + AuthorDate: 2025-09-18 12:24:02 +0900 |
| 143 | + Commit: GitHub < [email protected]> |
| 144 | + CommitDate: 2025-09-18 12:24:02 +0900 |
| 145 | + |
| 146 | + Use `ec->interrupt_mask` to prevent interrupts. [Backport #21610] |
| 147 | + |
| 148 | + Disallow pending interrupts to be checked during `FiberScheduler#unblock`. |
| 149 | + |
| 150 | + Ractors can send signals at any time, so the previous debug assertion can fail if a Ractor sends a signal. |
| 151 | + |
| 152 | + Co-authored-by: Luke Gruber < [email protected]> |
| 153 | + |
| 154 | +commit 6882012473b3bd3a9eb53ad7f7754144c0e63ef0 |
| 155 | + Author: Jean Boussier < [email protected]> |
| 156 | + AuthorDate: 2025-09-17 06:17:51 +0900 |
| 157 | + Commit: GitHub < [email protected]> |
| 158 | + CommitDate: 2025-09-17 06:17:51 +0900 |
| 159 | + |
| 160 | + [3.4] compile.c: Handle anonymous variables in `outer_variable_cmp` (#13493) |
| 161 | + |
| 162 | +commit b5346f2c5be00f886c9ad2f4b807deb728265a0a |
| 163 | + Author: Nick Dower < [email protected]> |
| 164 | + AuthorDate: 2025-05-16 22:44:45 +0900 |
| 165 | + Commit: Takashi Kokubun < [email protected]> |
| 166 | + CommitDate: 2025-09-17 05:36:08 +0900 |
| 167 | + |
| 168 | + Add a test case for `it` in a regex. |
| 169 | + |
| 170 | + Co-authored-by: Alan Wu < [email protected]> |
| 171 | + |
| 172 | +commit 1f6d845011cf8900c0e9c6fe5e4243a63a5ecae3 |
| 173 | + Author: Nick Dower < [email protected]> |
| 174 | + AuthorDate: 2025-05-16 12:18:28 +0900 |
| 175 | + Commit: Takashi Kokubun < [email protected]> |
| 176 | + CommitDate: 2025-09-17 05:36:08 +0900 |
| 177 | + |
| 178 | + [Bug #21313] Handle `it` in rescue and ensure blocks. |
| 179 | + |
| 180 | + The following is crashing for me: |
| 181 | + |
| 182 | + ```shell |
| 183 | + ruby --yjit --yjit-call-threshold=1 -e '1.tap { raise rescue p it }' |
| 184 | + |
| 185 | + ruby: YJIT has panicked. More info to follow... |
| 186 | + thread '<unnamed>' panicked at ./yjit/src/codegen.rs:2402:14: |
| 187 | + ... |
| 188 | + ``` |
| 189 | + |
| 190 | + It seems `it` sometimes points to the wrong value: |
| 191 | + |
| 192 | + ```shell |
| 193 | + ruby -e '1.tap { raise rescue p it }' |
| 194 | + false |
| 195 | + |
| 196 | + ruby -e '1.tap { begin; raise; ensure; p it; end } rescue nil' |
| 197 | + false |
| 198 | + ``` |
| 199 | + |
| 200 | + But only when `$!` is set: |
| 201 | + |
| 202 | + ```shell |
| 203 | + ruby -e '1.tap { begin; nil; ensure; p it; end }' |
| 204 | + 1 |
| 205 | + |
| 206 | + ruby -e '1.tap { begin; nil; rescue; ensure; p it; end }' |
| 207 | + 1 |
| 208 | + |
| 209 | + ruby -e '1.tap { begin; raise; rescue; ensure; p it; end }' |
| 210 | + 1 |
| 211 | + ``` |
| 212 | + |
3 | 213 | commit dbd83256b1cec76c69756ecb8758b9e1079833de |
4 | 214 | Author: Takashi Kokubun < [email protected]> |
5 | 215 | AuthorDate: 2025-09-16 07:38:23 +0900 |
|
0 commit comments