Red Hat Bugzilla – Bug 227260
ISO-8859 manual pages mis-coded
Last modified: 2007-11-30 17:11:56 EST
Description of problem:
During the transition to UTF-8 some manual pages have been damaged.
Version-Release number of selected component (if applicable): man-pages-2.39-5
Steps to Reproduce:
1. Switch to an English locale (I used en_GB.UTF-8)
2. from a shell type e.g. exactly 'man iso_8859-15' (NB underscore and dash)
3. Examine the EURO SIGN
The obscure universal currency symbol ¤ is shown
The Euro currency symbol € should be shown
These manual pages were casualties of an otherwise mostly successful battle to
get rid of encoding problems in manual pages.
Each of these pages was converted to UTF-8 but with the assumption that their
original character set was ISO-8859-1 (or maybe Windows 1252). In fact they were
each in their respective native encoding. A tool like 'iconv' can fix this,
firstly undo the conversion to UTF-8, then redo it but choosing the correct
source encoding e.g.
iconv -f UTF-8 -t ISO-8859-1 < iso_8859-9.7 > tmpfile
iconv -t UTF-8 -f ISO-8859-9 < tmpfile > iso_8859-9.7
Meanwhile the ASCII versions of these pages of course don't do a very good job
of illustrating characters which explicitly aren't in ASCII. Since we can't
illustrate them properly we should just remove this column, there is a text
description of each character included so it's still useful as pure ASCII.
Fixed in man-peges-2.43-6.fc7 and man-pages-2_39-7_fc6.