ConvertUTF.cpp is shipped with a non "University of Illinois Open Source License" compatible license

[sylvestre]

Bugzilla Link	32962
Version	trunk
OS	Linux
CC	@dwblaikie,@DimitryAndric,@jlebar,@MatzeB

Extended Description

Detected by lintian, the Debian static analyzer:

lib/Support/ConvertUTF.cpp include material under a non-free
license from Unicode Inc. Therefore, it is not possible to ship this in
main or contrib.

This license does not grant any permission to modify the files (thus
failing DFSG#3). Moreover, the license grant to attempt to restrict use
to "products supporting the Unicode Standard" (thus failing DFSG#6).

In this case a solution is to use libicu and to remove this code by
repacking.

Refer to https://bugs.debian.org/823100 for details.

And https://en.wikipedia.org/wiki/Debian_Free_Software_Guidelines for Debian information.

[DimitryAndric]

Interesting... This file seems to have been added a long time ago, here:

https://reviews.llvm.org/rL68208

Originally it was in cfe/trunk/lib/Basic, then it got moved around and edited a bit, but the original license was always in there.

[DimitryAndric]

Also note that there also appear to be older IBM versions of this file, with a completely different license, written by the original author, Mark E. Davis. I'm not sure how the Unicode org managed to change the license. :)

[sylvestre]

Other occurrences of this issue:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=900596
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=823100
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=916945

[sylvestre]

Looks like there is a license compatible version in breakpad:
https://raw.githubusercontent.com/google/breakpad/master/src/common/convert_UTF.h

I am trying to fix that here:
https://reviews.llvm.org/D64740

[sylvestre]

well done :)

[bkmgit]

@jyknight Discussion at AmokHuginnsson/replxx#12 indicates these can be replaced by https://github.com/nemtrif/utfcpp

[jyknight]

The code does not need to be replaced for license reasons. The files were relicensed to a proper license by Unicode.org some 15 years ago, so we fixed this issue for llvm by applying the "new" license terms to our copy. I would suggest anyone else who is in a similar situation should do the same. (Of course, if you want to use a different library for functional reasons, that's another matter.)