ed/refactor
1
0
Fork 0
mirror of https://github.com/Ed94/refactor.git synced 2024-09-20 04:37:33 -07:00
Code Issues Packages Projects Releases Wiki Activity
231c893c6b
refactor/Readme.md
Ed_ 231c893c6b Rework of project implementation 
For include and multi-file support. I still need to debug it,
Test will be adjusted as well; I want to get all the files not just zpl refactored using a powershell script.

I dropped the idea of semantically identifiying macros. While it may be possible, I don't see the utility vs the regular idendentifier distinction.
I want to keep the refactoring as simple as possible, where it just takes one pass to go through a file without any context to other files.

So far the ignores behave as a good guard filter for unwanted refactors and the only true weak area was the includes (which should be aleviated with the coming support for it.
2023-03-17 02:09:19 -04:00

2.5 KiB
Raw Blame History

refactor

Refactor c/c++ files (and problably others) with ease.

Parameters :

  • -num : Used if more than one source file is provided (if used, number of destination files provided MUST MATCH).
  • -src : Source file to refactor
  • -dst : Destination file after the refactor (omit to use the same as source)
  • -spec : Specification containing rules to use for the refactor.

Syntax :

  • not Omit word or namespace.
  • include Preprocessor include related identifiers.
  • word Fixed sized identifier.
  • namespace Variable sized identifiers, mainly intended to redefine c-namespace of an identifier.
  • , is used to delimit arguments to word or namespace.
  • L-Value is the signature to modify.
  • R-Value is the substitute ( only available if rule does not use not keyword )

The only keyword here excluisve to c/c++ is the include as it does search specifically for #include <L-Value>.
However, the rest of the categorical keywords (word, namespace), can really be used for any langauge.

There is no semantic awareness this is truely just a simple find and replace, but with some filters specifiable, and
words/namespaces only being restricted to the rules for C/C++ identifiers (alphanumeric or underscores only)

The main benefit for using this over other stuff is its faster and more ergonomic for large refactors on libraries that
you may want to have automated in a script.

There are other programs more robust for doing that sort of thing but I was not able to find something this simple.

Note

  • Building for debug provides some nice output with context on a per-line basis.
  • Release will only show errors for asserts (that will kill the refactor early).
  • If the refactor crashes, the files previously written to will retain their changes. Make sure to have the code backed up on a VCS or in some other way.
  • This was compiled using meson with ninja and clang on windows 11. The ZPL library used however should work fine on the other major os platforms and compiler venders.
  • The scripts used for building and otherwise are in the scripts directory and are all in powershell (with exception to the meson.build). Techncially there should be a powershell package available on other platorms but worst case it should be pretty easily to port these scripts to w/e shell script you'd perfer.

TODO:

  • Possibly come up with a better name.
  • Test to see how much needs to be ported for other platforms (if at all)
  • Provide binaries in the release page for github. (debug and release builds)