| Age | Commit message (Collapse) | Author | Files | Lines | |
|---|---|---|---|---|---|
| 2022-08-12 | Remove release news prior to the unified "lv2" release | David Robillard | 11 | -434/+11 | |
| Some of this metadata had issues, and presenting both the old individual extension releases and the new unified releases tends to confuse people. So, simply remove it; a summary of changes that old aren't really useful to anyone now anyway. | |||||
| 2022-08-12 | Fix documentation build | David Robillard | 1 | -0/+2 | |
| 2022-08-12 | Fix 1.18.6 release link | David Robillard | 1 | -1/+1 | |
| 2022-08-09 | Fix syntax check on Windows | David Robillard | 1 | -3/+6 | |
| 2022-08-09 | Avoid running syntax check in cross builds | David Robillard | 1 | -1/+1 | |
| This doesn't work because lv2_check_syntax.py uses serdi, and so would need to call it with a wrapper. That's possible (and is how the test scripts work), but this check doesn't need to run everywhere, so just disable it in cross builds to avoid the issue. | |||||
| 2022-08-08 | eg-sampler: Add resampling via libsamplerate | falkTX | 3 | -7/+52 | |
| 2022-08-08 | eg-sampler: Convert multi-channel audio files to mono | falkTX | 1 | -4/+25 | |
| 2022-08-08 | eg-sampler: Fix potentially corrupt notification events | David Robillard | 2 | -12/+13 | |
| 2022-08-08 | eg-sampler: Install click.wav to plugin bundle | David Robillard | 1 | -1/+1 | |
| 2022-08-07 | midigate: Fix output timing | Timothy Krause | 2 | -3/+5 | |
| Previously, the state was updated before writing the output (up to the current time), so output "in the past" would be incorrectly written based on the current state. This fixes that by switching things around, so the output is first written up to the current time, then the state is updated (and the loop continues). This takes advantage of the fact that write_output() effectively does nothing with a zero length. | |||||
| 2022-08-07 | Avoid running checks or building documentation if Python is too old | David Robillard | 1 | -1/+10 | |
| 2022-08-07 | Only run flake8/pylint/black tests with strict option enabled | David Robillard | 1 | -21/+23 | |
| Older versions of these tools can cause issues, but they are only useful for development anyway. So, just disable them by default to avoid build issues for users on ancient systems. | |||||
| 2022-08-05 | Remove unused Doxygen configuration | David Robillard | 2 | -174/+0 | |
| Older versions of Doxygen warn about these. So, remove them to make the build less noisy since they are all either unused or set to their defaults anyway. | |||||
| 2022-08-05 | Fix inconsistent namespace index order | David Robillard | 1 | -1/+1 | |
| This was "automatically" sorted on some systems and/or Python versions, but not others. | |||||
| 2022-08-05 | Fix documentation build with Python 3.7 | David Robillard | 3 | -4/+6 | |
| 2022-08-04 | Fix inconsistent plugin class labels | David Robillard | 3 | -4/+13 | |
| 2022-08-04 | Fix lv2.h missing from installation | David Robillard | 2 | -1/+6 | |
| 2022-07-21 | Fix documentation build with meson 0.56.2 | David Robillard | 2 | -1/+3 | |
| 2022-07-21 | Remove configuration not supported by Doxygen 1.9.1 | David Robillard | 1 | -57/+0 | |
| These all have the default value anyway. | |||||
| 2022-07-19 | Fix pylint warning | David Robillard | 1 | -2/+2 | |
| 2022-07-19 | Log syntax error origins when building index | David Robillard | 1 | -1/+5 | |
| 2022-07-19 | Fix invalid syntax | David Robillard | 1 | -0/+1 | |
| 2022-07-19 | lv2specgen: Fix installed data paths | David Robillard | 3 | -4/+11 | |
| 2022-07-17 | LV2 1.18.6v1.18.6 | David Robillard | 2 | -2/+5 | |
| 2022-07-17 | lv2specgen: Fix style and validation when installed | David Robillard | 2 | -18/+84 | |
| 2022-07-17 | lv2specgen: Remove warnings on external terms | David Robillard | 1 | -3/+0 | |
| This conveniently pointed out potential issues in the LV2 documentation, but makes the tool very noisy to use on anything else. Referring to some external thing really shouldn't be treated like a warning in linked data. Preventing mistakes (where full URIs are in the output instead of the expected prefixed names) will have to be done in some other, better way (probably based on an explicitly given set of prefixes that should be documented). | |||||
| 2022-07-17 | lv2specgen: Fix invalid output for terms about the spec itself | David Robillard | 1 | -1/+5 | |
| 2022-07-17 | Use the same set of external vocabulary prefixes everywhere | David Robillard | 3 | -49/+21 | |
| The RDFa in specification pages doesn't actually use prefixed names for the documented resources. So, simplify things by trimming the namespaces down to a consistent set that is actually used in the LV2 universe. | |||||
| 2022-07-17 | Suppress new warnings in clang and clang-tidy 14 | David Robillard | 3 | -0/+8 | |
| 2022-07-17 | Fix outdated doc comments | David Robillard | 1 | -12/+4 | |
| 2022-07-17 | Rearrange source tree to be directly usable by dependants | David Robillard | 146 | -1202/+465 | |
| This allows the LV2 source distribution to be used as an include path for compilers and an LV2_PATH for applications, at the expense of self-contained bundles. That's a nice idea, but it made LV2 itself weird and annoying to depend on. This rearranges things so that directories in the source tree correspond more closely to installation directories. To make this possible, the "aux" directory in the documentation output has been changed to "style", to avoid the reserved name "aux" on Windows. | |||||
| 2022-07-17 | Switch to meson build system | David Robillard | 85 | -1593/+3554 | |
| 2022-07-17 | Remove unnecessary parentheses | David Robillard | 2 | -7/+6 | |
| Unfortunately, GCC and clang-tidy have conflicting warnings here. Side with compiler warnings, being generally more relevant than a tool only used by developers. | |||||
| 2022-07-17 | Add missing release metadata | David Robillard | 2 | -0/+20 | |
| These should have been added for the 1.18.4 release. This seems a bit silly for minor changes like this that span the entire project, but the documentation tools currently require it. Simplifying this would be nice, perhaps by using NEWS files as the canonical source instead, which would be a nicer and more typical developer experience. | |||||
| 2022-07-17 | Consistently use the prefix "dcterms" everywhere | David Robillard | 12 | -23/+23 | |
| 2022-07-17 | Format all schemas consistently with serdi | David Robillard | 6 | -1007/+1101 | |
| 2022-07-17 | Replace canonical dcs ontology with a minimal version for LV2 | David Robillard | 2 | -238/+41 | |
| Since most applications pay to keep these ontologies in memory, it is best to keep them small. This replaces the "canonical" dcs ontology with a minimal and restricted version specifically written for the subset of the vocabulary that LV2 uses (like the xsd ontology). | |||||
| 2022-07-17 | Remove archaic properties from foaf vocabulary | David Robillard | 2 | -53/+2 | |
| 2022-07-17 | Clean up and modernize Python support code | David Robillard | 4 | -88/+102 | |
| Switches to safer and more modern idioms, and addresses many issues raised by flake8 and pylint. | |||||
| 2022-07-17 | Format all Python code with black | David Robillard | 3 | -93/+103 | |
| 2022-07-17 | Fix whitespace and add missing clang-format configuration | David Robillard | 4 | -4/+40 | |
| 2022-07-17 | Add dark mode style for documentation | David Robillard | 4 | -67/+904 | |
| 2022-07-17 | Remove symlink from repository | David Robillard | 2 | -2/+1 | |
| 2022-07-17 | Clean up documentation CSS | David Robillard | 2 | -36/+28 | |
| 2022-07-17 | Move documentation foreground/background properties to the document | David Robillard | 1 | -1/+4 | |
| This has better compatibility with tools that manipulate the colour scheme, like Dark Reader. | |||||
| 2022-07-17 | Use "python3" in shebangs | David Robillard | 3 | -3/+3 | |
| Python 2 hit EOL long ago, and some systems don't have a "python" at all. | |||||
| 2022-07-17 | Avoid -Warray-bounds with GCC 10 for arm32 | David Robillard | 1 | -6/+6 | |
| I'm not a fan of this style since it makes it possible to mess up the cast (whereas &a.atom always has the right type), but the compiler gets confused about the valid memory bounds otherwise, and it's equivalent anyway. | |||||
| 2022-07-17 | Consistently use unsigned literals for enum constants | David Robillard | 1 | -5/+5 | |
| 2022-07-17 | Make test logging portable to MinGW and old MS runtimes | David Robillard | 2 | -4/+3 | |
| 2022-07-17 | Make potentially lossy type conversions explicit | David Robillard | 4 | -6/+6 | |
 |
index : lv2.git | |
| LV2 | d@drobilla.net |
| aboutsummaryrefslogtreecommitdiffstats |