On Fri, May 21, 2010 at 22:45, Brandon Casey <brandon.casey.ctr@xxxxxxxxxxxxxxx> wrote: > > Hi Tor, I've been out of town for a few weeks... No problem, thanks for reviewing the log. > On 04/23/2010 09:59 AM, Tor Arntsen wrote: >> Below is a diff between the output from a run on Linux and the Tru64 one >> (instead of a full log), just to illustrate which ones failed. > > I skimmed through the log and it looks like the failures are mostly > related to tar and iconv. Possibly your tar is too old, or does not > support some option that git is using. The iconv failures related > to ISO8859-1 are a little strange since that encoding is very > common. Some implementations of iconv can produce a listing of the > encodings that are supported. You could check for support for ISO > 8859-1. I have re-run everything after installing GNU tar, and that fixed up the tar-related problems. There's still an iconv issue, and I'm going to look at it at some point - for now however I suspect that it might be a local problem caused by mixing native (/usr) and local (/usr/local) versions of iconv. > Many of the other differences are just because some tests were skipped > on the Tru64 run that were not skipped on Linux. The tests were skipped > because some required binary was missing, like gpg or cvs. > > Looks like you were able to run most of the test suite which is good. Yep, it looks pretty good (after the re-run), except for the iconv problem. I haven't posted any update of my Tru64 Makefile patch yet, I'm waiting to see if Gary Vaughan's portability patches are accepted - they included a Tru64 update to the Makefile, essentially a subset of my patch, enough to make Tru64 build with ./configure. If his patch gets included I can send a new patch which will add the necessary stuff to make it build also without ./configure. -Tor -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html