| 1 |
On 05/11/19 15:05, Mickaël Bucas wrote: |
| 2 |
> I remember reading an article about a man trying to reproduce binary |
| 3 |
> packages of a binary distribution and failing to do so, because there |
| 4 |
> are so many parts involved. I've read later that distributions have |
| 5 |
> done some work to have reproducible builds, but I'm not sure how |
| 6 |
> successful they are, even when all choices are predefined. |
| 7 |
|
| 8 |
It gets worse ... a major cause of two consecutive compiles on the same |
| 9 |
system not agreeing is that a lot of this contains date stamps etc. |
| 10 |
|
| 11 |
Reproducible builds are coming along, but they've got to analyze out or |
| 12 |
remove all the compile time info that ends up in the binary. They're |
| 13 |
coming because they're needed for security purposes. |
| 14 |
|
| 15 |
Cheers, |
| 16 |
Wol |
Archives