[PATCH] D22589: [ELF] - Introduce output section description class for holding section command

[Rui Ueyama via llvm-commits]

You may be able to learn something from the previous code, but as I said
copying it into the new code base require us to start over. Parsing is not
a difficult part, and we have no problem as to how to parse it. George,
Eugene and I are making a good progress, and we don't need such radical
change. Importing the old code not a good trade off at all.

You seems to think that the AST part is done in the previous code. I think
that's not correct. Since it was not plugged in to anywhere, it's not
proved to be in a useful shape (because it can dump the same script as it
read, it's certain that the parser is capable of creating a parse tree, but
it may lack "A" in AST.) On the other hand, in the new linker script
processor, we are careful to not add too much pausing code beforehand, so
everything we can parse is proved to be able to be processed efficiently
and correctly. You are suggesting we replace proved code with unproved one.

So overall I'm very sure that importing the previous code is not a good
idea. We can learn from it when we want it, but that's different from
copying code.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20160721/df1a0178/attachment.html>