log

age author description
2019-09-09 Chris Cannam Avoid simple static allocation of mutex, as it could lead to mutex being destroyed before last adapter that needs to use it (since adapters are usually also static) c++11-mutex
2019-09-09 Chris Cannam Don't even ask to unload library if another plugin is loaded from it c++11-mutex
2019-08-16 Chris Cannam Toward making PluginAdapterBase::Impl actually thread-safe! c++11-mutex
2019-08-14 Chris Cannam Ensure VAMPSDK_DIR is defined before being used
2019-05-16 Chris Cannam Looks like choco currently has a broken version of wget; use an older one
2019-02-07 Chris Cannam Added tag vamp-plugin-sdk-v2.8 for changeset be688783aa37
2019-02-07 Chris Cannam Correct release date vamp-plugin-sdk-v2.8
2019-02-07 Chris Cannam More build files
2019-02-07 Chris Cannam Tiny whitespace change
2019-02-01 Chris Cannam Update versions throughout to 2.8
2019-02-01 Chris Cannam Update CHANGELOG (release date is provisional)
2018-06-11 Chris Cannam Header for getenv
2018-06-11 Chris Cannam Use VAMP_PATH_32 environment variable instead of VAMP_PATH, when running in a 32-bit process within 64-bit Windows (WoW64)
2018-06-08 Chris Cannam Wide-char environment variable lookup
2018-05-11 Chris Cannam Doc note
2018-02-08 Chris Cannam Merge from branch vamp-kiss-naming
2018-02-07 Chris Cannam Libsndfile path vamp-kiss-naming
2018-02-06 Chris Cannam Another attempt to run the installer. Not hopeful about this one. vamp-kiss-naming
2018-02-06 Chris Cannam Have a go at bringing in libsndfile vamp-kiss-naming
2018-02-06 Chris Cannam Bit confused by this now vamp-kiss-naming
2018-02-06 Chris Cannam Ah, duplicate key problem - try this arrangement instead vamp-kiss-naming
2018-02-06 Chris Cannam Experiment with AppVeyor build vamp-kiss-naming
2018-01-30 Chris Cannam Merge vamp-kiss-naming
2018-01-30 Chris Cannam Avoid buffer underflow
2018-01-30 Chris Cannam Extensively rename things in the KissFFT headers to use a Vamp prefix. The motivation is not to change anything about the Vamp SDK library builds, but to avoid confusion in case any other code (for example that pulls in the Vamp SDK as part of a wider project definition) accidentally includes these headers instead of, or as well as, some other copy of KissFFT. vamp-kiss-naming
2017-12-11 Chris Cannam The PluginRateExtractor hack was exploiting undefined behaviour. Let's just give in and make that API official
2017-12-11 Chris Cannam Rework canonicalisation so as to avoid theoretical possibility of integer overflow. (This is unit tested in svcore, not here - tests pass)
2017-10-10 Chris Cannam Update before install; seems to be necessary (based on other projects) to avoid recent Ruby-related error
2017-05-25 Chris Cannam More (off-by-default) debug output
2017-03-06 Chris Cannam Added tag vamp-plugin-sdk-v2.7.1 for changeset 2dbb3f920abc
2017-03-06 Chris Cannam Update CHANGELOG and versions vamp-plugin-sdk-v2.7.1
2017-03-06 Chris Cannam Added tag vamp-plugin-sdk-v2.7.1 for changeset 25e023bad200
2017-03-05 Chris Cannam Choose another method (which should actually work) to override linkage for KissFFT
2017-02-24 Chris Cannam Added tag vamp-plugin-sdk-v2.7 for changeset 124752918a1b
2017-02-24 Chris Cannam Add Docker build scripts vamp-plugin-sdk-v2.7
2017-02-24 Chris Cannam Update CHANGELOG
2017-02-24 Chris Cannam Update installation notes for MSVC
2017-02-24 Chris Cannam Travis build file
2017-02-24 Chris Cannam Enforce c++98
2017-02-23 Chris Cannam Update MSVC build projects - two solutions, one for plugin SDK and one for host SDK,
2017-02-23 Chris Cannam Ignore these
2017-02-23 Chris Cannam Copyright dates
2017-02-23 Chris Cannam Remove this, it isn't useful
2017-02-23 Chris Cannam Might as well add these, since pretty much everyone seems to need them for Homebrew-installed libsndfile
2017-02-09 Chris Cannam Update a couple of comments to match the specifications elsewhere
2017-01-31 Chris Cannam Typo fix
2017-01-09 Chris Cannam Fix inadequate #if guard for non-gcc builds
2016-11-18 Chris Cannam Update regression test expected cases. One of the "obscure" RealTime fixes made recently actually fixed a substantial error (of 1 sample frame, sometimes 2) in printing timestamps resulting from input-domain adapter processing