SDK tools pipeline evolution #40
Conversation
[skip ci]
…chdenny/azure-sdk-tools into users/mitchdenny/pipelines-evolution
|
cc @scbedd |
|
Looks like you moved things under a tools folder for developers working in this repo do you think it is still better to group things by language? |
This repo is different than all the others. In the others we have a lot of code that all follows the same overall pattern, but in this one we've got lots of small tools/utilities, each one optimized for the particular scenario it is trying to solve. Rather than try to group by language and enforce a common set of build patterns by that language grouping I thought it would be easier just to have sub-directories under tools to allow for some flexibility - although some things will be reusable. |
Ok we can try that but I do think it might be nice to be able to share some things across the langauges, for example common targets/props in .NET, etc. Lets see what patterns end up emerging. |
|
Closing this one down as it is now way out of date. |
* core library template. * az_core * let * minor * AZ_CSTR * az_str and az_static_assert. * az_cstr * length * stream * az_http * test * test * more functions * additional options * data state * another try * more * strange * number draft * az_json_number * number + string * keywords * az_json_token * data types. * data types * Slices and index ranges. * JSON token * reorder * readme * Thoughts about reactive parser. * something new * errors * digit, string * state * readme files * issues * comments * notes * it works * readme.md * a lot * pointers * another test * No need for JSON state. * No need for az_sign.h * Stack overflow check. * az_const_str_eq * minor. * ++*p => *p += 1 * C99 unions, stack push/pop. * more tests. * Update documentation. * Symols. * az_is_digit => isdigit * Changes. * Simplify number parsing. * test * Less nested `if`. * AZ_MAKE_ERROR / AZ_MAKE_RESULT * RETURN_IF_NOT_OK. * az_const_str * az_dec_number * Test * AZ_SPAN * Span of bytes. * tests * run tests * enable_testing() * Fix test * Address comments. * az_json_read. * az_span * Reader * const bug * Tabs -> spaces (Azure#37) * Remove the last tab occurrence (appeared after the most recent merge) (Azure#38) * Tabs -> spaces * internal PR (Azure#39) * az_option_byte * refactoring for az_json_read_keyword_rest * az_json_read_string_rest * az_span_reader improved. * az_const_span_sub * clang-format. * CLANG * address comments * formatting. * Add core error for any argument-related errors (Azure#40) * Comments. * Update clang-format and reformat core files (Azure#42) Update clang-format and reformat core files * Remove az_option * Revert "Remove az_option" This reverts commit bad54a61702ef5e77c16a631d9339862b4878a27. * EOF * Add prefix/suffix header files; Switch to 100 char width style (Azure#43) * Add prefix/suffix header files * Switch code formatting back to 100 char max width * EOL@EOF * Fix forgotten instance to include epilog * Shorten the name for AZ_CORE_ERROR_ARGUMENT, as seen in another Sergey's branch. (Azure#44) * _az_cfg * AZ_ERROR_FLAG * static assert * counter * inline static_assert * dummy * another one * dummy function. * defaults * static inline * static * main * GCC * UNIX
This PR shows some early exploration into restructuring the SDK tools repo and associated pipelines. I've created a
pipelinesfolder underenglike in theazure-sdk-for-net. I also restructured the source files under atoolsfolder with a sub-path for each discrete tool (still a little more to be done).For now I just have one CI pipeline that builds everything. I collapsed doc-warden down into a single job build for performance reasons and overall the pipeline for pixel-server, doc-warden and the roslyn analyzers only runs for 3-4 minutes. At this point I don't think we'll get much extra by optimizing it further since all these jobs run in parallel.
I wanted to share this now so we could discuss direction here.