| 7 Jan 2025 |
 hexa | building firefox rapid on an rk3588 soc with 16GB RAM, 50% zramswap and 16GB swap | 13:58:27 |
| hexa | ooms while linking the second build at 37,9 GB | 13:58:53 |
 nbp | surprising, maybe with a different linker? | 14:00:48 |
 K900 | Second build includes PGO | 14:02:11 |
| K900 | Could be more intermediate artifacts or something | 14:02:18 |
| K900 | Also 5 hours oof | 14:02:27 |
| hexa | we're using lld right now, and optimizing for my smol arm sbc is probably not a worthwhile goal 🙂 | 14:02:52 |
| hexa | yeah, and it could use a nicer fan | 14:03:13 |
| nbp | whait, these are hours and not minutes? | 14:04:55 |
| nbp | * wait, these are hours and not minutes? | 14:05:02 |
| K900 | Yeah | 14:05:06 |
| hexa | lmao | 14:05:11 |
| K900 | It's not a very fast machine | 14:05:12 |
| K900 | 4xA76 + 4xA55 | 14:05:20 |
| hexa | Redacted or Malformed Event | 14:05:21 |
| K900 | Big cores clocked at around 2.2-2.4 GHz, small cores at 1.8 GHz | 14:06:03 |
| nbp | I've been spoiled by thread ripers. | 14:06:21 |
| K900 | We used to have a community Ampere | 14:06:53 |
| K900 | But now we don't | 14:06:56 |
| nbp | I could suggest to provide the compilation using qemu virtualization. | 14:14:30 |
| hexa | but you wouldn't 😄 | 14:14:58 |
| K900 | You mean qemu-user? | 14:15:06 |
| hexa | I'm not going back to emulation | 14:15:10 |
| K900 | Yeah no, it's awful slow | 14:15:15 |
| K900 | My RK3588 generally beats my 7950X3D with qemu-user | 14:15:33 |
| K900 | So unless someone wants to invest time into an actually fast aa64-on-x86_64 JIT | 14:16:08 |
| K900 | We can't really do that | 14:16:13 |
| K900 | (would be cool if someone did, though) | 14:16:28 |
| hexa | https://www.mozilla.org/security/advisories/mfsa2025-01 rapid
https://www.mozilla.org/security/advisories/mfsa2025-02 esr | 16:25:43 |
| hexa | https://www.mozilla.org/en-US/firefox/134.0/releasenotes/
https://www.mozilla.org/en-US/firefox/128.60/releasenotes/ | 16:25:56 |
