I had some trouble getting anything but UTC as my time zone. I searched the forums. Other people seem to solve this by using an appropriate link in /etc/localtime. Well, I had an appropriate link, to /usr/share/zoneinfo/MST7MDT, but "emerge world" yanked the rug out from under me! I had to go back to version 2006n of timezone-data to get the majority of the time zones back, for instance anything at all under US, or any [ECMP]ST ones or their [ECMP]ST{5E,6C,7M,8P}DT versions, or anything but a few Indiana or North_Dakota United States ones under America. What happened?
2006p installs all the relevant timezones for me post `emerge --info` like the bug report page told you to and then post the output as an attachment of: emerge timezone-data >& log
Created attachment 105350 [details] output from "emerge --info" as requested Do you want the 2006p, 2006o and 2006n emerges from the porthole "process" log, too?
i only really care about the latest version
Created attachment 105384 [details] output from "ls -R /usr/share/zoneinfo" before back-off to 2006n
Created attachment 105385 [details] porthole process log for emerge 2006o (replacing 2006p) Note: 423849 QA Notice: USE Flag 'elibc_FreeBSD' not in IUSE for sys-libs/timezone-dat a-2006o 423850 QA Notice: USE Flag 'elibc_glibc' not in IUSE for sys-libs/timezone-data- 2006o
Created attachment 105386 [details] output from "emerge timezone-data" for 2006p, replacing 2006n The emerge to 2006p took place after the other emerges. It has the same complaints regarding IUSE and libc.
Note that the emerge attached to comment #6 did result in a /usr/share/zoneinfo that has many more files than the original emerge world that left the truncated tree, including the MST7MDT one I actually use. Whether it has every file it should have, I don't know. I don't think there were any disk space problems during the failing emerges, and although I had some other problems with the 200 or so emerges (don't remember what those were, something with esound was one,) the only remaining one is the vnc problem with TLS mismatch and libc, something about Native POSIX Thread Library, I guess, which seems resistant to suggestions from the forums. I have the nptlonly USE flags set for glibc, it appears. I though both nptl and nptlonly were set for everything, but the raw /etc/make.conf and /etc/portage/* only show a glibc line that includes the string nptl. I only use the client to vnc anyway, so I could just remove it from my "world" list, I suppose. I put it in /etc/portage/package.mask to get through the "emerge world."
If I search for timezone-data missing I get no results. Is that because the index is updated only every several days, or is the search just pretty useless? I suppose there are more possibilities than just those two, but there may be less traffic for duplicate problems if we could find the elusive discussion that has already happened. Maybe the forums search is a little better, but I have trouble finding the relevant discussion amid dozens or hundreds of results that aren't remotely similar to what I'm trying to find. Once google gets it indexed, that works better, unless it delivers thousands instead of hundreds of irrelevant results. Also, although I keep adding a cc for myself, I'm not getting an e-mail copy of anything I send to bugzilla, except the very first submission. The replies from others seem to be arriving, though. I do want a copy of each, including what I send, so I'm not happy with that. Somewhere in comment #7, I missed a "t," so "I thought" became "I though" instead. Third line from the end of the middle paragraph, to be specific.
Sigh. I'm proofreading. Comment #4 was supposed to say that the MST7MDT file was one I manually copied in, first from some .../jre/... directory, which didn't work, maybe 16-bit internationalized character set or something, and later one that did work from a Red Hat 9 system I have. And, thanks for your interest, SpanKY. I submitted this because I thought maybe somebody else has run into the same problem, and we could get a definitive fix, but it looks like it's one of those "intermittent" problems, so, even if it's fairly common, it probably won't be easy to fix.
yours is the first report where someone said timezone-data was actually missing timezones ;)