summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
8 daysMPv 7.1.(14.53615 & 13.53713), no uriparser depHEADmasterErik Kundiman
2025-04-23Megapahit viewer 7.1.13.53691 ebuildErik Kundiman
Keyword is masked this time.
2025-04-23Megapahit viewer 7.1.13.53573 ebuildErik Kundiman
with NDOF on, and keyword not masked because 7.1.13.53452 wouldn't build any more now that Portage has upgraded Pipewire to 1.4.2. The 7.1.13.53452 ebuild is removed as well for that reason.
2025-04-107.1.13 has been upgraded to revision 53452Erik Kundiman
The optimisation flag environment custom setting is removed so that it won't override user's in /etc/portage/make.conf. The C++ standard environment custom setting is removed because Portage already has a default for this ebuild, which is gnu++20, that I don't think the custom one would cause anything significantly different, if there's any.
2025-04-07Repo setup & viewer version 7.1.13.53431 ebuildErik Kundiman
unpacker is needed for uncompressing zst files. Downloading 3rd-party files as part of the upstream CMake configuration wasn't allowed, the files need to be registered as part of the emerge process, but the unpacking destination path didn't seem customisable, so the hack is to just move the unpacked files to where they are needed to be in upstream building, rather than patching upstream files to work with Portage's default. The keyword is not masked because 7.1.13 is expected to be the default (stable) version soon.. even though the Gentoo port still crashes sometimes (might just be leftover cache from wrong CMake configuration build type, explained next). This ebuild is not a live Git using one, so the upstream configuration process can't figure out the revision number, so it needs to be set on the environment. Setting the CMake build type in mycmakeargs didn't have any effect, emerge set it to RelWithDebInfo instead (maybe that's why it would often crash, and the way it's supposed to be done is to set it to the cmake eclass CMAKE_BUILD_TYPE variable instead. Patching Collada DOM also seemed to have to be done as part of the emerge process (the one in the upstream configuration process didn't seem to have any effect). No USEs yet for now, and the WebRTC build being used is Crow Misia's for now, as it's been proven to be the safer bet not to prevent CEF from working.