Skip to content

Commit e069c98

Browse files
committed
docs(rl-next): add notes for curl-based s3
1 parent 3224636 commit e069c98

File tree

1 file changed

+26
-0
lines changed

1 file changed

+26
-0
lines changed
Lines changed: 26 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,26 @@
1+
---
2+
synopsis: "Improved S3 binary cache support via HTTP"
3+
prs: [13823, 14026, 14120, 14131, 14135, 14144, 14170, 14190, 14198, 14206, 14209, 14222, 14223, 13752]
4+
issues: [13084, 12671, 11748, 12403]
5+
---
6+
7+
S3 binary cache operations now happen via HTTP, leveraging `libcurl`'s native
8+
AWS SigV4 authentication instead of the AWS C++ SDK, providing significant
9+
improvements:
10+
11+
- **Reduced memory usage**: Eliminates memory buffering issues that caused
12+
segfaults with large files
13+
- **Fixed upload reliability**: Resolves AWS SDK chunking errors
14+
(`InvalidChunkSizeError`)
15+
- **Lighter dependencies**: Uses lightweight `aws-crt-cpp` instead of full
16+
`aws-cpp-sdk`, reducing build complexity
17+
18+
The new implementation requires curl >= 7.75.0 and `aws-crt-cpp` for credential
19+
management.
20+
21+
All existing S3 URL formats and parameters remain supported, with the notable
22+
exception of multi-part uploads, which are no longer supported.
23+
24+
Note that this change also means Nix now supports S3 binary cache stores even
25+
if build without `aws-crt-cpp`, but only for public buckets which do not
26+
require auth.

0 commit comments

Comments
 (0)