Description: The Time Zone tab doesn't shows correct characters when using pt_BR as the system default language. How reproducible: It's to reproduce anytime, following the step described below. Steps to Reproduce: 1. Change the system default language to Portuguese Brazil 2. Edit /etc/sysconfig/i18n to LANG="pt_BR" instead of "pt_BR.UTF-8" (needed to solve problems with man and to access created with special chars, specially windows fat32 files); 3. Open redhat-config-date and look at Time Zone tab. Actual results: Some garbage where accents and other special chars where expect Expected results: Names like América showing properly.
Can you attach a screenshot that shows what you are seeing?
Created attachment 94380 [details] I'm able to reproduce the problem as well
I'm a little confused. The 'file' command says that the pt_BR.po file is UTF-8, yet it appears that the encoding is not truly UTF-8 if we're seeing these weird characters in the display. pt_PT seems to be ok, but pt_BR is messed up. pgampe, can you have the translation team look at this?
Hi Brent, I went through the .po files today and in the timezones directory was a mix of various encodings. I have converted them all to UTF-8, and added a target to the Makefile should you need to do this in the future. pt_BR was in ISO-8859-1 but is now in UTF-8 Checking in pt_BR.po; /usr/local/CVS/redhat-config-date/po/timezones/pt_BR.po,v <-- pt_BR.po new revision: 1.3; previous revision: 1.2
Hi Brent, Paul, I've detected the same weird characteres with GTali, on RH 9 In another machine, I'm using severn (in fact, it's the same machine with another set of hard disks :)) and this bug is solved in GTali, but appears in redhat-config-date.
Hi Paul. Well, I've built a new version with the updated files but the problem still exists. When I run 'file' on po/pt_BR.po and po/timezones/pt_BR.po, here's what I see: [bfox@bfox po]$ file pt_BR.po pt_BR.po: UTF-8 Unicode English text [bfox@bfox po]$ file timezones/pt_BR.po timezones/pt_BR.po: UTF-8 Unicode English text So, both appear to be UTF-8 encoded, which is good. Then I run 'make' in the po directory and it generates pt_BR.mo, which contains the translations for both files. However, when I run 'msgunfmt pt_BR.mo | less', I can see weird characters that clearly aren't UTF-8. Any idea what's happening? For example (I hope mozilla doesn't mangle these cut-and-pasted strings): msgid "Africa/Banjul" msgstr "ÃÂfrica/Banjul" msgid "America/Antigua" msgstr "América/AntÃÂgua" msgid "Asia/Aden" msgstr "ÃÂsia/ÃÂden" msgid "Australia/Queensland" msgstr "Austrália/Queensland"
Good catch Brent, sorry I should have checked that. Looks like the file was originally in utf-8 but labeled as ISO-8859-1. So I had done a double encoding conversion. I have gone through and fixed the header and encoding and run the test you described above. All looks good to me now.
Thanks for your help. redhat-config-date-1.5.21-1 should contain the corrected translations. Resolving as Rawhide.