- May 08, 2021
-
-
Daniel Salzman authored
-
- Apr 23, 2021
-
-
- Apr 08, 2021
-
-
Daniel Salzman authored
-
-
- Feb 23, 2021
-
-
- Jan 29, 2021
-
-
Daniel Salzman authored
-
Daniel Salzman authored
-
- Sep 17, 2020
-
-
Daniel Salzman authored
-
- Sep 14, 2020
-
-
Daniel Salzman authored
-
- Sep 09, 2020
-
- Aug 27, 2020
-
-
Daniel Salzman authored
-
- Aug 26, 2020
-
-
Daniel Salzman authored
-
-
- Jul 01, 2020
-
-
- Jun 18, 2020
-
-
Daniel Salzman authored
Absence of '__attribute__(cleanup)' would lead to memory leaks.
-
- May 25, 2020
-
-
Daniel Salzman authored
The change didn't work as expected :-/ This reverts commit d71091d1.
-
- May 24, 2020
-
-
Daniel Salzman authored
-
- May 22, 2020
-
-
Daniel Salzman authored
-
- May 19, 2020
-
-
Daniel Salzman authored
It's needed for Debian Stretch and Ubuntu Xenial
-
- May 17, 2020
-
-
Daniel Salzman authored
-
- May 11, 2020
-
-
-
Daniel Salzman authored
-
Daniel Salzman authored
-
- Apr 21, 2020
-
-
Daniel Salzman authored
-
Daniel Salzman authored
-
Daniel Salzman authored
-
Daniel Salzman authored
-
-
-
-
- Feb 03, 2020
-
-
Vladimír Čunát authored
gnutls 3.6.12 adds support for this, given that you use sufficiently new nettle (not released yet). I tested this with Knot Resolver and https://rootcanary.org/test.html Note: in some switch statements I moved the labels inside #ifdef guards. Otherwise the fall-through wouldn't be correct in the weird case of not supporting ed25519 but supporting ed448 (shouldn't happen anyway).
-
- Oct 10, 2019
-
- Oct 04, 2019
-
-
- Oct 02, 2019
-
-
Daniel Salzman authored
-
- Sep 24, 2019
-
-
Daniel Salzman authored
-
-
- Apr 12, 2019
-
-
Daniel Salzman authored
-
- Apr 09, 2019
-
-
Daniel Salzman authored
-
- Mar 05, 2019
-
- Mar 04, 2019
-
-
Daniel Kahn Gillmor authored
It's not particularly useful to record -fdebug-prefix-map= and friends in the list of used CFLAGS arguments. Additionally, since these compiler arguments are typically used to make reproducible binary artifacts by stripping out information about the ephemeral build path used, embedding them in the resulting binary actually re-introduces the reproducibility problem. Note that gcc itself omits these values from the DWARF information that it generates: https://gcc.gnu.org/git/?p=gcc.git;a=blob;f=gcc/dwarf2out.c;h=1b17f2bc1d5eef3d3ef70abf92a4d09da285d2b8;hb=HEAD#l24396 You can see the reproducibility issue here: https://tests.reproducible-builds.org/debian/dbd/unstable/amd64/knot_2.7.6-2.diffoscope.html
-