aboutsummaryrefslogtreecommitdiff
path: root/src/analyze.cpp
AgeCommit message (Collapse)Author
2018-01-31fix assertion fail when using global var number literalAndrew Kelley
closes #697
2018-01-31*WIP* error setsAndrew Kelley
2018-01-30add compile error for duplicate struct, enum, union fieldsAndrew Kelley
closes #730
2018-01-29allow packed containers in extern functionsAndrew Kelley
2018-01-27fix crash when align 1 field before self referential...Andrew Kelley
...align 8 field as slice return type closes #723
2018-01-26fix compiler crash on function with invalid return typeAndrew Kelley
closes #722
2018-01-25syntax: functions require return type. remove `->`Andrew Kelley
The purpose of this is: * Only one way to do things * Changing a function with void return type to return a possible error becomes a 1 character change, subtly encouraging people to use errors. See #632 Here are some imperfect sed commands for performing this update: remove arrow: ``` sed -i 's/\(\bfn\b.*\)-> /\1/g' $(find . -name "*.zig") ``` add void: ``` sed -i 's/\(\bfn\b.*\))\s*{/\1) void {/g' $(find ../ -name "*.zig") ``` Some cleanup may be necessary, but this should do the bulk of the work.
2018-01-22add new kind of test: generating .h files. and moreAndrew Kelley
* docgen supports obj_err code kind for demonstrating errors without explicit test cases * add documentation for `extern enum`. See #367 * remove coldcc keyword and add @setIsCold. See #661 * add compile errors for non-extern struct, enum, unions in function signatures * add .h file generation for extern struct, enum, unions
2018-01-22fix crash on union-enums with only 1 fieldAndrew Kelley
closes #713
2018-01-18fix null debug info for 0-length array typeAndrew Kelley
closes #702
2018-01-15fix compiler crash related to @alignOfAndrew Kelley
2018-01-15clean up error return tracingAndrew Kelley
* error return tracing is disabled in release-fast mode * add @errorReturnTrace * zig build API changes build return type from `void` to `%void` * allow `void`, `noreturn`, and `u8` from main. closes #535
2018-01-14error return trace pointer prefixes other paramsAndrew Kelley
instead of being last. This increases the chances that it can remain in the same register between calls.
2018-01-13typecheck the panic functionAndrew Kelley
2018-01-12*WIP* proof of concept error return tracesAndrew Kelley
2018-01-11functions which can return errors have secret stack trace paramAndrew Kelley
See #651
2018-01-11the same string literal codegens to the same constantAndrew Kelley
this makes it so that you can send the same string literal as a comptime slice and get the same type
2018-01-09remove %% prefix operatorAndrew Kelley
See #632 closes #545 closes #510 this makes #651 higher priority
2018-01-07replace `a %% b` with `a catch b`Andrew Kelley
See #632 better fits the convention of using keywords for control flow
2018-01-06Darwin -> MacOSX, added Zen. See #438Andrea Orru
2017-12-26self-hosted: build against zig_llvm and embedded LLDAndrew Kelley
Now the self-hosted compiler re-uses the same C++ code for interfacing with LLVM as the C++ code. It also links against the same LLD library files.
2017-12-22explicitly return from blocksAndrew Kelley
instead of last statement being expression value closes #629
2017-12-20add labeled loops, labeled break, labeled continue. remove gotoAndrew Kelley
closes #346 closes #630 regression: translate-c can no longer translate switch statements. after #629 we can ressurect and modify the code to utilize arbitrarily returning from blocks.
2017-12-19fix assert when wrapping zero bit type in nullableAndrew Kelley
closes #659
2017-12-19export keyword works againAndrew Kelley
2017-12-19wip bring back export keywordAndrew Kelley
2017-12-18wip export rewriteAndrew Kelley
2017-12-10rename @EnumTagType to @TagType in type namesAndrew Kelley
2017-12-06add higher level arg-parsing API + misc. changesAndrew Kelley
* add @noInlineCall - see #640 This fixes a crash in --release-safe and --release-fast modes where the optimizer inlines everything into _start and clobbers the command line argument data. If we were able to verify that the user's code never reads command line args, we could leave off this "no inline" attribute. * add i29 and u29 primitive types. u29 is the type of alignment, so it makes sense to be a primitive. probably in the future we'll make any `i` or `u` followed by digits into a primitive. * add `aligned` functions to Allocator interface * add `os.argsAlloc` and `os.argsFree` so that you can get a `[]const []u8`, do whatever arg parsing you want, and then free it. For now this uses the other API under the hood, but it could be reimplemented to do a single allocation. * add tests to make sure command line argument parsing works.
2017-12-05add implicit cast from enum to unionAndrew Kelley
when the enum is the tag type of the union and is comptime known to be of a void field of the union See #642
2017-12-04fix abi alignment of union-enums not counting tag typeAndrew Kelley
add more tests for unions See #618
2017-12-03rework enums and unions and their relationship to each otherAndrew Kelley
* @enumTagName renamed to @tagName and it works on enums and union-enums * Remove the EnumTag type. Now there is only enum and union, and the tag type of a union is always an enum. * unions support specifying the tag enum type, and they support inferring an enum tag type. * Enums no longer support field types but they do support setting the tag values. Likewise union-enums when inferring an enum tag type support setting the tag values. * It is now an error for enums and unions to have 0 fields. * switch statements support union-enums closes #618
2017-12-02ability to set tag values of enumsAndrew Kelley
also remove support for enums with 0 values closes #305
2017-12-02casting between integer and enum only works via tag typeAndrew Kelley
See #305
2017-12-01packed structs can have enums with explicit tag typesAndrew Kelley
See #305
2017-11-30ability to specify tag type of enumsAndrew Kelley
see #305
2017-11-29fix capturing value of switch with all unreachable prongsAndrew Kelley
closes #635
2017-11-24fix assertion failed when invalid type encounteredAndrew Kelley
2017-11-24rename "parsec" to "translate-c"Andrew Kelley
2017-11-16union secret field is the tag index instead of distinct type indexAndrew Kelley
See #144
2017-11-15unions have a secret field for the typeAndrew Kelley
See #144
2017-11-15basic union supportAndrew Kelley
See #144
2017-11-10fix test failuresAndrew Kelley
put all the codegen for fn prototypes to the same place
2017-11-08fix enum sizes too largeAndrew Kelley
closes #598
2017-11-06add @memberType and @memberName builtin functionsAndrew Kelley
see #383 there is a plan to unify most of the reflection into 2 builtin functions, as outlined in the above issue, but this gives us needed features for now, and we can iterate on the design in future commits
2017-11-03Windows libc & static libc are located in the same dir which is already ↵dimenus
covered by msvc_lib_dir
2017-11-02fix incorrect debug info for empty structsAndrew Kelley
closes #579 now all tests pass for llvm master branch
2017-11-01windows: use the same libc search within a compilation unitAndrew Kelley
2017-11-01WIN32: Linking with the CRT at runtime. (#570)Dimenus
Disclaimer: Forgive me if my format sucks, I've never submitted a PR before! Fixes: #517 I added a few things to allow zig to link with the CRT properly both statically and dynamically. In Visual Studio 2017, Microsoft changed how the c-runtime is factored again. With this change, they also added a COM interface to allow you to query the respective Visual Studio instance for two of them. This does that and also falls back on a registry query for 2015 support. If you're using a Visual Studio instance older than 2015, you'll have to use the existing options available with the zig compiler. Changes are listed below along with a general description of the changes. all_types.cpp: The separate variables for msvc/kern32 have been removed and all win32 libc directory paths have been combined into a ZigList since we're querying more than two directories and differentiating one from another doesn't matter to lld. analyze.cpp: The existing functions were extended to support querying libc libs & libc headers at runtime. codegen.cpp/hpp: Microsoft uses the new 'Universal C Runtime' name now. Doesn't matter from a functionality standpoint. I left the compiler switches as is to not introduce any breaking changes. link.cpp: We're linking 4 libs and generating another in order to support the UCRT. Dynamic: msvcrt/d, vcruntime/d, ucrt/d, legacy_stdio_definitions.lib Static: libcmt/d, libvcruntime/d libucrt/d, legacy_stdio_definitions.lib main.cpp: Update function call names. os.cpp/hpp: COM/Registry interface for querying Windows UCRT/SDK. Sources: [Windows CRT](https://docs.microsoft.com/en-us/cpp/c-runtime-library/crt-library-features) [VS 2015 Breaking Changes](https://msdn.microsoft.com/en-us/library/bb531344.aspx)
2017-10-27use llvm named structs for const values when possibleAndrew Kelley
normally we want to use llvm types for constants. but union constants (which are found inside enums) when they are initialized with the non-most-aligned-member must be unnamed structs. these bubble up to all aggregate types. if a constant of an aggregate type contains, recursively, a union constant with a non-most-aligned-member initialized, the aggregate typed constant must be unnamed too. this fixes some of the asserts that were coming in from llvm master branch.
generated by cgit v1.2.3 (git 2.39.1) at 2025-12-30 09:32:04 +0000