aboutsummaryrefslogtreecommitdiffstats
AgeCommit message (Collapse)AuthorFilesLines
2022-11-25Avoid creating test files in the current directoryDavid Robillard2-1/+22
2022-11-25Merge similar testsDavid Robillard3-143/+119
2022-11-25Always handle the return value of serd_env_set_base_uri()David Robillard3-16/+22
2022-11-25Always handle the return value of eat_byte_safe()David Robillard3-61/+56
2022-11-24Gracefully handle bad characters in Turtle blank node syntaxDavid Robillard6-3/+20
2022-11-24Gracefully handle bad characters in Turtle datatype syntaxDavid Robillard4-3/+14
2022-11-24Explicitly ignore the posix_fadvise() return valueDavid Robillard2-2/+3
I can't figure out a sensible way to test this, and even if it did fail, at most a warning message could be printed because this is just a performance hint anyway.
2022-11-24Check serd_writer_write_statement() return valueDavid Robillard1-2/+2
2022-11-24Avoid redundant comparisonDavid Robillard1-1/+1
2022-11-24Avoid mutation in testDavid Robillard1-18/+22
2022-11-24Simplify error handling logicDavid Robillard1-6/+5
2022-11-24Fix improper use of chunk sinkDavid Robillard1-4/+4
2022-11-15Fix potential Python errors on test suite failureDavid Robillard1-5/+5
2022-11-15Fix implementation-defined integer conversionDavid Robillard1-1/+1
2022-11-14Arrange CI rows more logicallyDavid Robillard1-91/+57
This is organized in a way that hopefully makes more sense when viewed in a report. It also puts "basic" cases first, to ensure that tedious configuration issues are caught as early as possible when running rows sequentially for local testing.
2022-11-14Add Windows path separator support to serd_node_new_file_uri()David Robillard3-10/+69
2022-10-24Add meson subproject files to .gitignoreDavid Robillard1-0/+1
This is generated in the source directory by meson when the project is checked out of git by the meson wrap system. Ignoring it here ignores issues with tools that assume a clean source directory, like the REUSE test.
2022-10-24Fix MSVC warning in header testDavid Robillard1-0/+4
2022-10-23Fix hang when skipping an error at EOF when lax parsingDavid Robillard2-1/+2
2022-10-23Fix crash when trying to read chunks without startingDavid Robillard3-3/+11
2022-10-23Fix whitespaceDavid Robillard1-2/+2
2022-10-23Add trailing commas to enum declarationsDavid Robillard2-7/+7
This has been supported since C99 and C++11, and makes for less noisy diffs and fewer conflicts when things are added.
2022-10-23Fix "suspicious" string comparisonDavid Robillard2-2/+1
This is beyond silly.
2022-10-23Strengthen clang-tidy configurationDavid Robillard1-2/+0
2022-10-23Use uppercase integer literal suffixesDavid Robillard8-60/+58
2022-10-23Simplify clang-tidy configurationDavid Robillard3-16/+15
2022-10-23Test header for warnings more strictlyDavid Robillard5-3/+50
2022-10-07Override pkg-config dependency within mesonDavid Robillard4-3/+12
2022-09-08Serd 0.30.16v0.30.16David Robillard4-5/+5
2022-09-08Use 0BSD for trivial "public domain intent" thingsDavid Robillard22-142/+33
Fedora takes issue with CC0. Although it doesn't really matter for this stuff (dual licensed anyway, questionably "software", certainly not patentable, and so on), this is simpler and more consistent with the ISC license used for the actual software anyway.
2022-09-08Only run REUSE test in strict modeDavid Robillard1-12/+14
2022-09-01Fix licensing metadataDavid Robillard4-4/+4
The intent here is that relatively generic configuration files, build definitions, or things considered "simple facts" are released to the Commons, but more substantive and original works, including documentation, are licensed the same as the software.
2022-09-01Update NEWS fileDavid Robillard1-1/+2
2022-08-31Adopt REUSE machine-readable licensing standardDavid Robillard65-494/+335
2022-08-22Use standard ISC license textDavid Robillard1-7/+7
I'm not sure where the disclaimer variant that starts with "THIS" came from (although it is better that way), but this one that starts with "THE" is the standard text which is recognized by tools like `licensee` (and therefore Github) as a perfect match. There are no other changes other than whitespace.
2022-08-22Only run autoship test in strict modeDavid Robillard2-3/+9
2022-08-09Allow programs to be used from subprojectDavid Robillard2-1/+4
2022-08-09Fix array size warningDavid Robillard1-3/+7
2022-07-20Make NEWS file readable by dpkg-parsechangelogDavid Robillard1-20/+20
2022-07-17Use consistent pkg-config descriptionDavid Robillard1-1/+1
2022-07-16Remove fill from logoDavid Robillard1-17/+9
This makes it look reasonably good on both dark and light backgrounds.
2022-07-16Simplify command line usage printingDavid Robillard1-16/+19
2022-07-16Remove redundant check for command line argument countDavid Robillard1-4/+1
2022-07-16Write test outputs to a temporary directoryDavid Robillard1-12/+10
While occasionally useful, I almost always end up reproducing the issue live to investigate something anyway. Not keeping the many tests results around results in less clutter, and hopefully makes the test suites faster in environments with bad I/O like Docker.
2022-07-16Define _POSIX_C_SOURCE globally in the build systemDavid Robillard4-8/+13
Using inconsistent defines like this that affect the standard library implementation can cause issues. So, doing this consistently for the whole library is a better approach, although it unfortunately makes the code more difficult to compile manually.
2022-07-16Use a lookup table to determine UTF-8 encoding lengthDavid Robillard1-20/+38
2022-07-16Build with C11 on CIDavid Robillard1-13/+13
Since this is not the default (because that would break the default build for some older systems), distributions and users are likely covering the C99 case heavily, so this is the best way to ensure that both are covered well.
2022-07-16Fix spelling error in serdi man pageDavid Robillard4-5/+12
2022-07-14Set strict C warnings globallyDavid Robillard1-1/+1
This is only used when serd is not a subproject.
2022-07-13Suppress clang-tidy identifier length warning in headerDavid Robillard1-0/+1