Anthony J. Bentley wrote:
David Uhden Collado <[email protected]> wrote:I removed --locked from MODCARGO_BUILD_ARGS and from the test invocation. MODCARGO already vendors the dependency set and builds offline, so there was no good reason to add it explicitly.I also changed ICU from RUN_DEPENDS to LIB_DEPENDS and added the dynamically loaded ICU libraries to WANTLIB, with a comment noting that they are loaded through dlopen. The EDIT_CFG_ICU*_SONAME settings remain necessary because upstream opens those libraries explicitly at runtime. I removed ONLY_FOR_ARCHS as well. amd64 and arm64 were simply the two architectures I had tested, which was not sufficient justification for restricting the port. There does not appear to be an upstream requirement limiting Edit to those architectures. The updated port builds successfully, and the complete test suite also completes successfully:The contents of the tarball you sent are identical to the previous one. Since you agree with the changes, I went ahead and made them in my copy. I also removed the custom do-test target since the default tests work fine. The attached version is ok bentley@
Hi, Sorry, I made a mistake while trying to fix this quickly and ended up sending a tarball containing the same version of the port as before. The changes I described in my previous email were in fact implemented; I simply attached the wrong tarball. I've attached the version I intended to send with the previous email. Sorry for the confusion and inconvenience. Thanks, David.
msedit-2.0.0.tar.gz
Description: application/gzip
