summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorEevee (Alex Munroe) <eevee.git@veekun.com>2014-11-12 16:51:13 -0800
committerEevee (Alex Munroe) <eevee.git@veekun.com>2014-11-12 16:51:13 -0800
commitd4c74c5ddf65277bd5ce77b44c7f921b79413f6d (patch)
treec3f90d3d787166b35f665eda1e77eab1bec0d4fb
parent6689b62877361e362654f0e3b9e9ba9a29b29665 (diff)
downloadpyscss-d4c74c5ddf65277bd5ce77b44c7f921b79413f6d.tar.gz
Comment update.
-rw-r--r--scss/extension/core.py12
1 files changed, 4 insertions, 8 deletions
diff --git a/scss/extension/core.py b/scss/extension/core.py
index 227b559..3e8f2e8 100644
--- a/scss/extension/core.py
+++ b/scss/extension/core.py
@@ -26,14 +26,10 @@ class CoreExtension(Extension):
"""Implementation of the core Sass import mechanism, which just looks
for files on disk.
"""
- # TODO virtually all of this is the same as the django stuff, except
- # for the bit that actually looks for and tries to open the file.
- # would be much easier if you could just stick an object in the search
- # path that implements the pathlib API. the only problem is what to do
- # when one path is a child of another, so the same file has two names,
- # but tbh i'm not actually sure that's something worth protecting
- # against...? like, the only cost is that we'll parse twice (or, later
- # on, not respect single-import), and the fix is to just Not Do That
+ # TODO this is all not terribly well-specified by Sass. at worst,
+ # it's unclear how far "upwards" we should be allowed to go. but i'm
+ # also a little fuzzy on e.g. how relative imports work from within a
+ # file that's not actually in the search path.
# TODO i think with the new origin semantics, i've made it possible to
# import relative to the current file even if the current file isn't
# anywhere in the search path. is that right?