On Wed, Jan 14, 2015 at 11:25:24AM +1100, Kubilay Kocak wrote: > On 14/01/2015 6:58 AM, Sean Bruno wrote: > > > > > For now, I'm going to start looking into Makefile patches to setup the > > correct variables so that waf doesn't assume we want number of CPU's > > detected. > > > > Might this be a good candidate for USE_PYTHON= waf feature for some > consistency and minimised duplication in waf-based Python Ports?
Recently I had to look into the code for waf to fix a build issue (see bug 195953 for the gory details). If such a feature is going to be built, maybe this PR can be taken into account. What I observed is that python2.7-config is a Python sccript, while python3.4m-config is a shell-script. And waf assumes the config script is a Python script. The version of waf bundled with that port is quite old (1.6.4). The source code of that version is extremely terse with few if any docstrings and almost no blank lines. It produces an awesome amount of errors and warnings when you run it through flake8. The most recent version is a lot better in that regard. So I wonder if the different versions are compatible enough to make a generic USE_PYTHON=waf feature possible. Roland -- R.F.Smith http://rsmith.home.xs4all.nl/ [plain text _non-HTML_ PGP/GnuPG encrypted/signed email much appreciated] pgp: 5753 3324 1661 B0FE 8D93 FCED 40F6 D5DC A38A 33E0 (keyID: A38A33E0)
pgpwaxV98iihT.pgp
Description: PGP signature