[PATCH] D57014: [llvm-objcopy] Consistently use createStringError instead of make_error<StringError>. NFC.

[Martin Storsjö via Phabricator via llvm-commits]

mstorsjo added a comment.

In D57014#1365204 <https://reviews.llvm.org/D57014#1365204>, @jhenderson wrote:

> Make sure to mention in the commit comment that this also quotes certain symbol names.


Ah, right, yes - will do.

> I'm surprised that several of the error messages appear to be untested (only one test needed updating, but there are at least 3 different behaviour changes here). I take it testing them is not straightforward?

Correct, the two cases where I add quotes but there's no test change for are currently not possible to trigger, I think - both the input as set by Reader, and updates by Object currently should make them impossible to trigger. I guess asserts would be an option then as well. (At an earlier point in the evolution of the removeSections patch, some of them were triggerable, but I realized it's easy to do the right thing instead of erroring out on dangling associative sections). There's also lots of error messages in Reader that aren't tested, mainly for broken inputs.


Repository:
  rL LLVM

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D57014/new/

https://reviews.llvm.org/D57014