Disable md5 publishing #164
Conversation
|
Should we perhaps also stop publishing (not start publishing) sha-1)? I.e., perhaps sha-1 can/should be used to populate only when getting it from Maven but not when using p2 to publish a new repository given that both sha-256 and sha-512 are now published and only one of the latter two will ever be used at runtime such that a sha-1will never be consumed so is just useless overhead... |
|
Yes if we eventually add sha-1 (the PR is still pending) it could be disabled like md5 fro publishing. |
...tory/src/org/eclipse/equinox/internal/p2/artifact/processors/checksum/ChecksumUtilities.java
Outdated
Show resolved
Hide resolved
...tory/src/org/eclipse/equinox/internal/p2/artifact/processors/checksum/ChecksumUtilities.java
Outdated
Show resolved
Hide resolved
laeubi
force-pushed
the
disable_md5_publishing
branch
from
3ff049c
to
aca6f29
Compare
...tory/src/org/eclipse/equinox/internal/p2/artifact/processors/checksum/ChecksumUtilities.java
Outdated
Show resolved
Hide resolved
laeubi
force-pushed
the
disable_md5_publishing
branch
4 times, most recently
from
c64d3ac
to
d3d890d
Compare
...quinox.p2.repository/src/org/eclipse/equinox/p2/repository/artifact/IArtifactDescriptor.java
Outdated
Show resolved
Hide resolved
|
Lets first get the option to disable a checksum in: |
laeubi
force-pushed
the
disable_md5_publishing
branch
from
d3d890d
to
7005fb8
Compare
|
@mickaelistria please update your review, @akurtakov thats how far I was able to get with the tests, I'll keep this PR just in case someone likes to build on top of it. |
akurtakov
force-pushed
the
disable_md5_publishing
branch
from
7005fb8
to
fd5f93c
Compare
|
Just so I get things correct, what prevents this patch from getting merged is failing tests? |
Yes tests need to be adjust (or deleted or ...) because they depend on md5 calculated (what is no longer the case) |
|
Would you please reset this PR on master? aka removing my merge commit. |
laeubi
force-pushed
the
disable_md5_publishing
branch
from
e13f768
to
241c5c2
Compare
Done! |
This is disabling the md5 checksum, we publish sha-256 > 4 years and warn about md5 > 1 year now so it seems valid to stop publishing it now at all.
akurtakov
force-pushed
the
disable_md5_publishing
branch
from
241c5c2
to
22e00d0
Compare
|
Pushed as e60c5f1 |
Currently all registered checksums are computed and published, but in some cases it might be usefully to be able to verify a checksum but we don't want to publish it.
This also includes disabling the md5 checksum, we publish sha-256 > 4 years and warn about md5 > 1 year now so it seems valid to stop publishing it now at all.