On Aug 16, 2016, at 3:04 PM, Peter Danecek wrote: > could anyone explain me what exactly happened to one of my builds? Looks like > the selfupdate step fails, but why? > A race condition while updating/signing the tar? Was the base.tar touched > recently? > > https://build.macports.org/builders/ports-10.6_i386_legacy-watcher/builds/334/steps/selfupdate/logs/stdio > > There are more builds which fail for the same reason and I’d assume this is > not related to the port itself. However, this way the related up-to-date > binaries were not generated.
I've seen these types of errors before. There must be some way, in the current implementation of the script that generates and signs the base.tar file, that they don't get updated at the same time. I intend to rewrite the script in such a way that ensures an atomic update. Until then, you can try to force the build again an hour later, by which time there should be a new file. _______________________________________________ macports-dev mailing list [email protected] https://lists.macosforge.org/mailman/listinfo/macports-dev
