On Jul 19, 2016, at 10:49 AM, Watson Ladd wrote: > We've got it building, the problems outstanding are one easy upstream patch > and its tendency to fetch things during build/vendored zlib. > > Personally I'm not volunteering to maintain distribution patches to fix the > fetch and zlib issues. It seems upstream won't budge on zlib. (So what? > Everyone vendors SQLite) and I don't know what they think about avoiding > fetch in build. > > If it compiles, installs, uninstalls, and works, I consider that problem > solved.
It does look like the use of the local zlib files is baked pretty pervasively into the build system so let's table that issue for now. The issue that the configure script fetches zlib, nanopass.ss and Mf-stex is easy to address without patching. The configure script only fetches and extracts those tarballs if their contents doesn't already exist. So you can simply have MacPorts fetch and extract those tarballs during the fetch and extract phases using the normal MacPorts directives for that. This will have the advantage that activities happen in the expected phases; distfiles can be mirrored across our mirror network so that in the unlikely but not unprecedented event that github is down, the port can still be built, and in any case users will get the files from the closest mirror which might be faster than github; and checksums can be used to verify that the downloaded files haven't been tampered with. If you don't want to mess with this, that's fine; I can add it to the port later. _______________________________________________ macports-dev mailing list [email protected] https://lists.macosforge.org/mailman/listinfo/macports-dev
