During last update I've got the following error: * Messages for package dev-java/sun-jdk-1.6.0.29: * Fetch failed for 'dev-java/sun-jdk-1.6.0.29', Log file: * '/var/tmp/portage/dev-java/sun-jdk-1.6.0.29/temp/build.log' * GNU info directory index is up-to-date. And only in build log read the following message: Due to Oracle no longer providing the distro-friendly DLJ bundles, the package has become fetch restricted again. Alternatives are switching to dev-java/icedtea-bin:6 or the source-based dev-java/icedtea:6 Please download jdk-6u29-linux-i586.bin from: http://www.oracle.com/technetwork/java/javase/downloads/jdk-6u29-download-513648.html and move it to /var/distfiles Normally the check on fetch restriction should be done before starting update and block it until source becomes available.
I'm not sure if there's something off with only this package, or you mean the whole portage behavior is undesirable ? (In reply to comment #0) > During last update I've got the following error: > > * Messages for package dev-java/sun-jdk-1.6.0.29: The messages should have been echoed just fine. But if you used e.g. --keep-going and ELOG_SYSTEM="echo" then you probably expected them to be printed again at the end of emerge, as the original ones might have scrolled away already. It's true that the package could use elog instead of einfo, which would make the difference on a typical elog configuration. > Normally the check on fetch restriction should be done before starting update > and block it until source becomes available. This suggests you mean the whole portage behavior as there should be nothing specific with sun-jdk here. I think normally portage downloads distfiles before each package (or in parallel with parallel-fetch) but does not resolve them before the whole list of scheduled packages starts. With "emerge -f" you only download the distfiles without building and thus get also the nofetch instructions, but it can take a while on slow connection due to other, fetchable distfiles. Seems there is indeed a room for improvement, to e.g. warn about fetch restricted packages missing distfiles and print the instructions immediately.
Currently on this system I don't use '--keep-going' option. It was update and not first fetch-restricted package I install. Earlier (although AFAIR it was a time of portage 1.x) portage executed pre emerge/fetch check: should it be possible to fetch all distfiles and if not (fetch restricted atoms) in exited with error ,essage and does nothing before the issue solution. Now (sys-apps/portage-2.1.10.11) this check seems to be skipped (although fetch restriction message is echoed). If necessary, I (or anyone else) could make a check with any other fetch-restricted package. I think, that pre-emerge check with interruption on fail should be returned.
(In reply to comment #2) > Now (sys-apps/portage-2.1.10.11) this check seems to be skipped (although fetch > restriction message is echoed). > If necessary, I (or anyone else) could make a check with any other > fetch-restricted package. > > I think, that pre-emerge check with interruption on fail should be returned. OK, this is pretty much bug 388093 then. *** This bug has been marked as a duplicate of bug 388093 ***
