clacke
@lispi314 @dalias @LunaDragofelis @mjg59 Aaah haha yes this is what abstractions get us. I didn't even know, it was pure luck that I invoked yet another monster.
Thank you for finding yet another example on how trying to optimize with complex methods buys us more side channels.
LisPi
@clacke @mjg59 @dalias @LunaDragofelis It can work, but it requires explicit consideration of the security & trust boundaries.
Much the same reason why hyperthreading can be considered safe if you've set the hypervisor to only schedule threads from the same VM on shared cores and that you consider individual VMs to be the trust boundaries. Then, even with all known exploitation methods available, no breach of the trust boundary happens.
clacke
@dalias @lispi314 @LunaDragofelis @mjg59 The way I read "bloating memory requirements", the complaint was not how much intrinsic memory use compilation takes, but how much incidental memory use there is.
My joke was how wasting memory is fine, as long as you waste it the exact same way and then optimize for it, in the same way that we always seem to avoid going back to root causes.
Rich Felker
@lispi314 @clacke @mjg59 @LunaDragofelis Dupe page merging is not going to help compile jobs anyway. A page containing pointers (all of them in a compiler) will never be duplicate.
GNU social JP is a social network, courtesy of GNU social JP管理人. It runs on GNU social, version 2.0.2-dev, available under the GNU Affero General Public License.
All GNU social JP content and data are available under the Creative Commons Attribution 3.0 license.