Description of problem: guiles fails to build because of failures in the test suit. The failures say: UNRESOLVED: i18n.test: text collation (Czech): string-locale<? for 'ch' FAIL: i18n.test: number->locale-string: French: integer - arguments: (expected-value "123 456" actual-value "123\xa0456") FAIL: i18n.test: number->locale-string: French: negative integer - arguments: (expected-value "-1 234 567" actual-value "-1\xa0234\xa0567") FAIL: i18n.test: number->locale-string: French: fraction - arguments: (expected-value "1 234,567" actual-value "1\xa0234,567") FAIL: i18n.test: number->locale-string: French: fraction, 1 digit - arguments: (expected-value "1 234,6" actual-value "1\xa0234,6") FAIL: i18n.test: format ~h: French: 12345.678 - arguments: (expected-value "12 345,678" actual-value "12\xa0345,678") FAIL: i18n.test: monetary-amount->locale-string: French: integer - arguments: (expected-value "123 456,00 +EUR" actual-value "123\xa0456,00 +EUR") FAIL: i18n.test: monetary-amount->locale-string: French: fraction - arguments: (expected-value "1 234,57 EUR " actual-value "1\xa0234,57 EUR ") Version-Release number of selected component (if applicable): 2.0.14-3 How reproducible: always on koji
This has been blocking module builds for a while.
FYI, these failures appear to be the result of an upstream glibc update to the handling of numerics for the fr_FR locale. The thousands separator character is now defined to be a non-breaking space rather than a plain space. guile's i18n.tests, however, are demanding a plain space. Other glibc locales were also updated in a similar manner. See https://sourceware.org/git/?p=glibc.git;a=commit;h=70a6707fa15e63591d991761be025e26e8d02bb6
Thanks for the hint, Merlin. The i18n test should be now fixed to expect non-breaking spaces separating thousands.