CommandHelper Builds

View last successful build

Branch # Status Changes Finished
master #3515- Tests passed: 634, ignored: 7
  • Re-wrap SOE in closure's execute also (8f3478 by ladycailin)
4 months ago
master #3514- Tests passed: 634, ignored: 7
  • Re-wrap SOE and use CRESOE, so we get more useful errors and can generally support this in the exception system. This might need to be added to the execute method as well. (39e5f2 by ladycailin)
4 months ago
master #3513- Tests passed: 634, ignored: 7
  • Fix site-deploy Now, in addition to fixing the actual issue (in a way that that won't happen again), we now also print out usable error messages. For whatever reason, Logger was not working, so instead of fixing that, it has simply been removed, and made to work with the existing paradigm anyways. This should make it easier to diagnose errors within the documentation itself in the future. (48e260 by ladycailin)
4 months ago
master #3512- Tests passed: 634, ignored: 7
  • Properly pass column information through the exception system Previously, the causedBy section of the exceptions were not properly printed in the default unhandled exception handler. Now, the column information that is passed through is properly printed (though I think the column information itself is still not accurate, but this is a problem in the compiler, not the exception system). Also added "at" to the beginning of each stack trace line, which brings the format more in line with other languages stacktrace prints. (06fe30 by ladycailin)
4 months ago
master #3511- Tests passed: 634, ignored: 7
  • Add ValueType semantics This allows objects to mark themselves as value types. Currently, all primitives are marked as such. In the future, the compiler may be able to further optimize these types, particularly if true compilation is added. For now, this is unused, and the compiler treats everything as pass by reference, but this is not ideal. At least going forward now, the ValueType class is established, and other classes can go ahead and start marking themselves as such. Eventually, this mechanism can probably be exposed to user classes, but for now it is internal only. (b24b6d by ladycailin)
4 months ago
master #3510- Tests passed: 633, ignored: 7
  • Checkstyle (dbeef8 by ladycailin)
4 months ago
master #3509- Tests passed: 633, ignored: 7; exit code 1 (new)
  • Fix compiler bug when dealing with bad breaks Previously, when using the break(1) format, if the integer wasn't atomic, then this would cause a java stack trace. Now it causes a normal compile error. (bae587 by ladycailin)
4 months ago
master #3508- Tests passed: 633, ignored: 7
  • Add a more helpful error message (2c40de by ladycailin)
4 months ago
master #3507- Tests passed: 633, ignored: 7
  • Catch Error too This allows us to always (at least attempt) to print an error message. While OOME will still leave us in a precarious state, we should at least attempt to print out a helpful error message. (5128f8 by ladycailin)
4 months ago
master #3506- Tests passed: 633, ignored: 7
  • Construct to mixed2018 (#511) This set of changes completes the goals of the original construct-to-mixed branch. Function unfortunately worked with Constructs, rather than the Mixed interface. This was a violation of the Dependency Inversion Principal, which has prevented several critical improvements to the language. The problem is that in Java, there is not multiple inheritance, so while normal classes could easily extend Construct and then be added into the ecosystem, for other types of classes, where it's impossible to extend Construct (particularly because they are already extending another class, which is true in at least the cases of enums and exceptions), these objects simply could not be added into the ecosystem. Now, everything accepts and for the most part deals with Mixed, which is an interface. This has the unfortunate side effect of breaking binary compatibility with extensions, but this change has been in the works for at least three years. As part of this change, implementing enums was used as a test case. That has now also been done, and barring a few bugs, seems to mostly work. This allows access of enums as first class objects, rather than as arrays of strings as previously implemented through reflect_pull. The PR opens lots of possibilities in the future, and so is a welcome change to the base, but there are a few short term goals after this PR is merged: * Move exceptions into first class status. Currently they are implemented as arrays, but this is a shoehorn and not ideal. * Write up documentation for Enums. There is some documentation already available, and this may be mostly correct, but it should at least be read through to ensure it is correct, and then linked to the learning trail. It's also possible to iterate enums with foreach, and so this should be added as an example in various places. * There are still a few bugs with enums, particularly it is not possible to assign an enum class to a variable. ClassType @s = ArraySor... (496418 by ladycailin)
5 months ago
master #3500- Tests passed: 627, ignored: 7
  • Fix docs for uuid (676db5 by ladycailin)
5 months ago
master #3499- Tests passed: 627, ignored: 7
  • Add NativeObjectArchitecture design doc This is in conflict with the NewObjects design doc, and these conflicts need to be thought through and one or the other changed. My current thinking is to use the style defined in the NativeObjectArchitecture doc, but I need to think more on this. (6cb2e6 by ladycailin)
5 months ago
master #3498- Tests passed: 627, ignored: 7
  • Refine cross casting docs (9fc3c9 by ladycailin)
5 months ago
master #3497- Tests passed: 627, ignored: 7
  • Refine ObjectUsage docs (020ebe by ladycailin)
5 months ago
master #3496- Tests passed: 627, ignored: 7
  • Create documentation for Object Usage. (1ff50a by ladycailin)
5 months ago

Available as:  RSS feed JSON feed