As I pointed out several times in the past, the performance hit of enabling SHA1DC globally is not acceptable. This patch series not only demonstrates that clearly in the perf test it adds (it is the last patch in the current series, and its commit message has some numbers), it also shows an early glimpse of what I will be proposing to fix the situation. Obviously, this patch series is not complete yet: - the first patch is an obvious bug fix that I sent out separately, but it is needed to unbreak almost all tests on Windows. - the most important part will be the patch turning core.enableSHA1DC into a tristate: "externalOnly" or "smart" or "auto" or something indicating that it switches on collision detection only for commands that accept objects from an outside source into the local repository, such as fetch, fast-import, etc Johannes Schindelin (7): sha1dc: safeguard against outside definitions of BIGENDIAN Makefile: optionally compile with both SHA1DC and SHA1_OPENSSL config: add the core.enablesha1dc setting t0013: do not skip the entire file wholesale without DC_SHA1 t0013: test DC_AND_OPENSSL_SHA1, too mingw: enable DC_AND_OPENSSL_SHA1 by default p0013: new test to compare SHA1DC vs OpenSSL Makefile | 13 +++++++++++++ config.c | 8 ++++++++ config.mak.uname | 1 + hash.h | 2 +- sha1dc/sha1.c | 23 ++++++++++++++++++++++- sha1dc/sha1.h | 16 ++++++++++++++++ t/helper/test-sha1.c | 10 ++++++++++ t/perf/p0013-sha1dc.sh | 23 +++++++++++++++++++++++ t/t0013-sha1dc.sh | 18 ++++++++++++------ 9 files changed, 106 insertions(+), 8 deletions(-) create mode 100644 t/perf/p0013-sha1dc.sh base-commit: 063fe858b89ef8ee27965115fd6b1ed12e42e793 Published-As: https://github.com/dscho/git/releases/tag/sha1dc-knob-v1 Fetch-It-Via: git fetch https://github.com/dscho/git sha1dc-knob-v1 -- 2.12.1.windows.1