| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* docker experiment in travis
* let's run 'em all
* okay run as travis
* fine just use a number
* remove xenial since we test that elsewhere
* suspicion
* review feedback
* more comments
* use workdir
* heavy sigh
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Fixes #4459 -- switch all our builds to xenial on travis
* Attempt to fix tests
* Try disabling sudo?
* typo
* try this?
* shellcheck
|
|
|
| |
of course, if this works it might just be luck
|
| |
|
|
|
|
|
|
| |
* Don't clone wycheproof if we're doing a downstream test
* you and your rules
|
|
|
|
|
| |
rsa, and keywrap (#4310)
* Refs #3331 -- added initial wycheproof integration, starting with x25519 tests
|
|
|
|
|
|
|
|
| |
* try compiling with asm for our custom openssl
* we also need to update the cache dir
* try actually compiling it
|
| |
|
| |
|
|
|
|
|
|
| |
* build and test libre on travis
* remove libressl jenkinsfile data
|
| |
|
| |
|
|
|
|
|
|
| |
* we pinned coverage for xml generation...let's try everywhere
* add so many comments
|
|
|
|
|
|
| |
* Get our travis from pypy
* doh
|
|
|
|
|
|
|
|
|
|
| |
* update pythons and pypy, a little prep for pypy3 in CI again
* slightly renamed
* oops
* pypy renames all around
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* compile all our mac builds against 1.1.0 in travis
* Just install
* try this why not
* ...
* of course
* upgrade the existing openssl
|
|
|
|
|
|
|
|
| |
* Revert "Attempted workaround for intermittent coverage issues (#3661)"
This reverts commit 84a634c1e7645970382a3a271dfae9ef95685e7a.
* pin coverage to 4.3.4
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Switch from 12.04 to 14.04 on travis
* test?
* fix for pyenv
* sigh
* try removing this entirely
* one more try
* empty commit to retrigger travis
|
|
|
|
|
|
|
|
|
|
| |
* Fixed #3334 -- added Python 3.6 support
* install py36
* empty commit to retrigger travis
* this is an impressively dumb typo
|
| |
|
|
|
|
|
|
|
|
| |
* Simplify OpenSSL bits in travis
* more simplify
* missed one
|
|
|
|
|
|
| |
* test against 1.1.0a instead of 1.1.0
* change the dir to foil travis caching
|
|
|
|
|
|
|
|
|
|
|
|
| |
* update our travis configs to run against 10.10, 10.11, 10.12
Drops 10.9 from Travis. The xcode8 image is also currently not 10.12,
but is planned to become that soon see:
https://blog.travis-ci.com/2016-09-15-new-default-osx-image-coming/
* add output of sw_vers for mac builders on travis
* reorder
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* make pre5 work
* add a blank line to make the diff happier
* 1.1.0-pre6 working
* support the changes since 1.1.0-pre6
* fixes
* add 1.1.0 to travis
* expose the symbol
* better testing for numericstring
* handle libre...
* actually use the 1.1.0 we compile
* cache the ossl-110 dir on travis
* add some newlines
* changelog entry for 1.1.0 support
* note that we test on 1.1.0
* proper skip on this test
* reorder
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* empty commit
* only run this one build
* try pinning this
* why wasn't this installed?
* revert this
* english, how does it work?
* roll back these changes
|
|
|
| |
Refs #2982
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Test with both old pypy and new pypy
* use an old as heck pypy
* revert
* dont try to coverage pypy2.6.1
* ALL THE PYPYS
|
| |
|
|
|
|
|
|
| |
* Remove unused py26 on OSX support
* Updated the accuracy of a comment
* Word-wrapped some other comments
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We do this to prevent a version collision between the custom one we're
installing and whatever the system has. OpenSSL 1.0.0 through 1.0.2 all
declare an SHLIB version of 1.0.0, so if Python has been linked against
1.0.1 and cryptography against 1.0.0, but then you try to load Python
it will consider either 1.0.1 or 1.0.0 to satisfy the library version.
This is, of course, nonsense since 1.0.1 has substantially more symbols
than 1.0.0. The result is that if you do a LD_LIBRARY_PATH that points
at the "real" 1.0.0 then Python will fail to load because there are
missing symbols. We can avoid this entire nonsense by changing the major
version. The dynamic linker will happily load both versions and nothing
will kerplode.
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
This will probably slow down the test infrastructure as we're running
far more CommonCrypto tests than before. We should figure out what to
do there before merging this.
|
| |
|
| |
|
| |
|