* Gary V. Vaughan wrote on Sun, Jun 24, 2007 at 04:02:57PM CEST: > On Jun 24, 2007, at 7:27 AM, Ralf Wildenhues wrote: > > >Please add necessary casts to this patch so compilation with C++ > >compiler is not broken again. > > Sounds like we need a test that tries to compile libltdl with C++.
No, it's much simpler: you can just keep a second build tree around with ../libtool/configure -C CC=g++ and get in the habit of compiling that. And eventually running 'make check' too. I don't do it for every review because I forget to, also because I don't have two build trees but also one for cross compile to MinGW, one that uses pmake, one that uses the Intel, Sun compiler suites, one for ccache. And of course (on other systems) several for other values of $host. If I run each and every one for every review, then that takes an extra day. But you know that. > Before I submit a patch, I simply ensure that there are no regressions > in the testsuite on a couple of system types (usually darwin and > Linux). I don't use C++ (or even know it particularly well), and > will definitely forget to run everything through a C++ compiler by > hand after each patch I write. OK. Then let's do it when you've finished the major libltdl work (which I hope is soon?). > >These hunks are all superfluous, except of course the one functional > >change to the tryall_dlopen call, for which one must look closely > >to not overlook it. Please separate functional changes from cleanup > >changes in patch postings, that would make the former be much more > >easily spottable. > > That's because you asked me to run the file through 's, ,^I,' !! Did I? Oh, sorry. Cheers, Ralf
