s-expressions are too limited a code representation #52

Closed
cemerick opened this Issue Sep 28, 2012 · 2 comments

Comments

Projects
None yet
2 participants
Contributor

cemerick commented Sep 28, 2012

This manifests itself most immediately right now in that e.g. code processed using cljx loses all whitespace, and therefore stack traces and such report incorrect line numbers. I can imagine more significant issues around using LineNumberingPushbackReader (e.g. not being able to apply rules to tagged reader literals, #= forms, and probably more), but the whitespace issue is what's biting me personally at the moment.

Something like a "lossless" reader would be more appropriate, with sjacket being the best candidate so far. (It's quite young itself, but is currently used by reply and therefore leiningen FWIW.)

@cemerick cemerick added a commit to cemerick/cljx that referenced this issue Feb 11, 2013

@cemerick cemerick keep emitted forms on the same lines as they were found in source
.cljx files (hacky fix for jonase/kibit#52)
079f1c8
Collaborator

danielcompton commented Nov 9, 2014

@cemerick I understand the lossy nature of reading a transformed cljx file but I don't understand how this would be causing issues when using kibit. What's the scenario where this is a problem? By the way cljx files where added in #102 but aren't in a release yet.

Contributor

cemerick commented Nov 10, 2014

cljx hasn't used kibit in some time, starting with v0.3.0. This can be closed if it is not germane or useful to anyone else.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment