Hi Eric! On Thu, Feb 24, 2011 at 11:27 PM, Eric Blake <eblake@xxxxxxxxxx> wrote: > On 02/24/2011 03:23 PM, Tadeus (Eus) Prastowo wrote: > > But, I am wondering how the tarball can be released when not all test > cases > > are okay ;-) > > They were all okay _for the person who did the release_, given the set > of software installed on one single machine. That does not mean the > testsuite is bug-free; sometimes, we don't find out about testsuite bugs > until after a release, when other people test with a wider combinations > of other software. > Ah, I see. > But that does not mean that a bug in the testsuite invalidates using > that build of autoconf. You can most likely safely ignore those > testsuite failures, especially given that Ralf has pointed out the > likely commit that fixed that particular testsuite bug. > Yup, it makes sense. > > That is the reason why I try to bootstrap from what I have in my git > clone. > > Bootstrap generally requires starting from an installed tarball, not > from a git clone. Once you have installed something, then you can work > from git and install a git snapshot. But you have to start somewhere, > and that's from a tarball. > Ah, I understand now. Thank you very much for the information :-) > -- > Eric Blake eblake@xxxxxxxxxx +1-801-349-2682 > Libvirt virtualization library http://libvirt.org > > -- Best regards, Eus (FSF member #4445) In this digital era, where computing technology is pervasive, your freedom depends on the software controlling those computing devices. Join free software movement today! It is free as in freedom, not as in free beer! Join: http://www.fsf.org/ _______________________________________________ Autoconf mailing list Autoconf@xxxxxxx http://lists.gnu.org/mailman/listinfo/autoconf
