-
Couldn't load subscription status.
- Fork 472
Fixes regressions introduced in previous merges #455
Conversation
|
Thank you @mgreter. I was just about to install everything on my new system to fix this. 👍 |
|
There still needs to be few fixes to be made. The idea really is:
Both of the above assertions aren't complied by libsass at the moment. For that matter: First, we need to store Secondly, in "/*# sourceMappingURL=" + resolve_relative_path(source_map_file, source_path, cwd) + " */";I guess after that the API will be quite flexible and finally browsers will be able to locate correct sources with source maps generated by libsass. Thoughts? |
Note that this does *not* collapse x/../y sections into y. This is by design. If /foo on your system is a symlink to /bar/baz, then /foo/../quux is actually /bar/quux, not /quux as a naive ../-removal would give you.
Adds canonical file resolving where appropriate. It might should be noted that output_path does not mean that libsass will write the result to that location. It is merely used to create file paths relative to the output file (which defaults to input_path with a 'css' extension).
Accounts for parent references (`..`) in `stripped_base`.
|
@am11: Libsass already resolves sources relative to the source map file. I have added some additonal commits which should do what you have asked for (sourceMappingURL relative to output_path). @akhleung: With these changes paths are internally always represented with forward slashes. The |
Add a source entry for stdin. This mimics exactly the behavior of the Google Closure Compiler on stdin data.
|
@mgreter, thanks! All these changes are extremely useful. 👍 Regarding the source-map initializer issue, please see the |
Points to working directory (ending with a slash) if no input file path has been given (i.e. on 'compile_string').
|
Thanks @mgreter! 😃 👍 |
|
Sorry, I just haven't had time to look at it yet. Will try to get it merged soon. |
|
@mgreter, yes; all our path-related concerns regarding Hopefully we get the mapping issue fixed as well. I don't have right insights about the sass-compiler's internals and how it reports sources. If you guys can take a look or point me to the locations in code so I can give it a try, it will truly make a big impact and help many affected out there.. Thanks! |
|
Looked over the code and it looks good ... I'll test/merge it in the morning when I have access to my work machine. |
Fixes regressions introduced in previous merges
|
Merge o'clock has begun. On Fri, Aug 8, 2014 at 9:09 PM, Aaron Leung [email protected]
|
|
Indeed 😃 I wish if someday node-sass will get blessed with such a thing: "Merge o' clock"... node-sass has been insanely slow lately in adapting changes from libsass, accepting PRs, releases and all those usual things which happen to a repository.. |
|
Unfortunately, it looks like node-sass's maintainer is no longer able to devote time to the project: |
|
Can anyone else take the throne and keep the wheel spinning? |
Fixes a few issues introduced in recently merged pull requests.
#441:
resolve_relative_pathwould always return lowercase paths on windows, which will not work if one would deploy the files to a linux server (or any other case sensitive filesystem). Implemented a better and IMO correct way to handle this. Also introduced a preprocessor directiveFS_CASE_SENSITIVEso that the behaviour could be overruled on compile time.#440:
sourceMappingURLis always added even ifsource_map_filewas not set to anything at all.