utf8proc 1.2 release

[stevengj]

Now that we have nicer documentation (thanks to #29), as far as I can tell we are ready for a 1.2 release as soon as the version number is updated.

(I don't think it's necessary to wait until we can get the HTML docs into the release tarball; there are equivalent docs in utf8proc.h, after all, and in any case the lack of formatted docs in the tarball is not a regression.)

@jiahao, @tkelman, @nalimilan, @StefanKarpinski, your thoughts?

[tkelman]

Fine by me. Worth doing a sanity check (maybe you have already) that things are still fine when current utf8proc master is linked into Julia?

[stevengj]

I just tried it, and all the Julia tests pass on my machine with the utf8proc master.

[jakebolewski]

+1 it would be great to fix the unicode character width issues

[stevengj]

Hmm, one final tweak: looking at this, it seems like we need UTF8PROC_VERSION_MAJOR etc. #defines in utf8proc.h so that we can check the API version at compile-time. This will allow callers such as Julia to use e.g. utf8proc_charwidth for utf8proc-1.2 or later, and fall back to wcwidth for systems with older utf8proc versions.

Update: this is now added.

[tkelman]

Is calling runtime utf8proc_version not sufficient?

[nalimilan]

#defines can always be useful.

I'd just like to add support for make install DESTDIR before the new release. Patch coming.

[stevengj]

@tkelman, the runtime version is not sufficient if you need to decide at compile-time whether you can call a function like utf8proc_charwidth or must fall back to wcwidth. For example:

#if UTF8PROC_VERSION_MAJOR > 1 || (UTF8PROC_VERSION_MAJOR == 1 && UTF8PROC_VERSION_MINOR >= 2)
...use utf8proc_charwidth...
#else
...use wcwidth...
#endif

[stevengj]

@nalimilan, it is more standard to do make prefix=DESTDIR install (this is the GNU standard promulgated by autotools and is widely imitated), and this is already supported.

[tkelman]

the runtime version is not sufficient if you want to decide at compile-time whether you can call a function like utf8proc_charwidth or fall back to wcwidth.

Yeah, good point.

it is more standard to do make prefix=DESTDIR install

Distribution buildbots all use DESTDIR. It's "strongly recommended" practice - http://www.gnu.org/prep/standards/html_node/DESTDIR.html

[stevengj]

Thanks, I wasn't familiar with staged installs, so I didn't know about the DESTDIR/prefix dichotomy. (I normally use automake and get all these things for free, but automake seems like overkill here). Looks like we'll also get it for free if we switch to cmake.

[nalimilan]

OK, see #31

[stevengj]

Anything else to check before we are good to go? (I'll fix up the NEWS.md file to cross-ref the issues or PRs as in Julia.)

[nalimilan]

Fine with me.

[tkelman]

Looks like you haven't tagged yet, I realized that the Julia build on Windows does use the makefile here. We probably don't want to do the soname symlinking there either.

[tkelman]

Just checked deps/Makefile again - it doesn't use make install, it does make libmojibake.a and copies the static library and header into Julia's usr. So probably no big deal.

[stevengj]

Yes, the Julia build only makes the .a library, because it ends up getting linked into libjulia (by default).

[stevengj]

Okay, release tag is pushed and website is updated.

[stevengj]

(Unfortunately, I just noticed that there were some stray mojibake references remaining in the bench/ subdirectory. Not a critical issue since bench/bench.c is mainly for use by utf8proc developers, who should be using git anyway, but probably calls for a 1.2.1 release at some point. Let's wait a couple weeks, though—at least until JuliaLang/julia#10659 is merged and gives utf8proc 1.2 wider testing, to see if any other problems turn up).