On Thu, Jan 02, 2020 at 12:24:44AM +0100, Charlene Wendling wrote: > On Wed, 1 Jan 2020 23:33:23 +0200 > Leo wrote: > > > On Wed, Jan 01, 2020 at 11:30:03PM +0200, Leo wrote: > > > On Wed, Jan 01, 2020 at 04:22:27PM +0100, Charlene Wendling wrote: > > > > Hi, > > > > > > > > On Tue, 31 Dec 2019 20:31:47 +0200 > > > > Leo wrote: > > > > > > > > > > > > > > On Mon, Dec 30, 2019 at 05:21:30PM +0200, Leo wrote: > > > > > > On Mon, Dec 30, 2019 at 10:24:32PM +1100, Jonathan Gray wrote: > > > > > > > On Mon, Dec 30, 2019 at 06:05:23AM +0200, Leo wrote: > > > > > > > > please don't punch me for removing subpackaging > > > > > > > > > > > > > > Send patches inline and not as attachments. > > > > > > > > > > > > > > Do not do things like remove ONLY_FOR_ARCHS or multi > > > > > > > packages in diffs to update a port. Especially with no > > > > > > > justification. > > > > > > > > > > > > > > Making whitespace changes also makes it harder to follow. > > > > > > > > > > > > > > As this is a few days out from the the first release in two > > > > > > > years where they switched to Tesseract increasing the > > > > > > > hardware requirements and making a bunch of other changes > > > > > > > it would be better to wait for a patch release assuming it > > > > > > > is still useable. > > > > > > > > > > > > ONLY_FOR_ARCHS should list all archs where > > > > > > OpenGL is supported. Where OpenGL isn't > > > > > > supported RedEclipse server should still be > > > > > > runnable. I did whitespace changes to make > > > > > > it more readable for myself, there was too > > > > > > much padding. > > > > > > > > > > > > Besides that I was confused in line: > > > > > > COMPILER = base-clang ports-gcc base-gcc > > > > > > > > > > here we go: > > > > > > > > HOMEPAGE should be moved to https. > > > > > > > > FYI, i tried to build it on powerpc, it indeed fails [0], > > > > similarly to geo/qgis. > > > > > > > > Charlène. > > > > > > > > > > > > [0] https://bin.charlenew.xyz/redeclipse.log > > > > > > > > > > That's a broken compiler... > > It's more about the vintage architecture than the compiler in that > case. I was reacting mostly to the ONLY_FOR_ARCHS removal. >
yes, but looking at log it's impossible to see what exactly caused that error, that's why I blamed GCC > > q66 from RedEclipse asks you to throw -Os into > > CFLAGS and CXXFLAGS > > Same results sadly. q66 can mail me if they want more details, so we > don't bring too much noise in that list, but trying to fix it on macppc, > an old platform EOL'd more than a decade ago is probably a waste of time > for them. > > ok, I was thinking about replacing ports-gcc with ports-clang in that line, but can we mark not for archs at those platforms where Red Eclipse really doesn't build and run instead of marking it as only for i386 and amd64? anyway, q66 said it's a waste of time because OpenBSD doesn't have general powerpc support