Bug#742942: Bug#749438: gparted stuck at searching /dev/sdc partitions
On Mon, May 26, 2014 at 06:52:11PM -0400, Phillip Susi wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA512 Ok, I have been able to reproduce this and diagnose the problem. Your 3 1/2 inch floppy link file contains the 1/2 character, which in code page 437 ( the original IBM PC character set ) is 0xBD. In UTF-8, this is an incomplete multi byte sequence. Dosfsck therefore, is emitting invalid UTF-8 character sequences by blindly spitting out the characters as they appear on disk. It should be translating them to the current system character set ( normally UTF-8 ) using the iconv library, and assuming the characters on disk are CP-437 unless otherwise directed. I am therefore cloning this bug and assigning it to dosfstools. In addition, when using iconv to convert this character to valid UTF-8, it maps it to U+255C instead of U+00BD, so I'm cloning the bug there as well. That is wrong. 0xBD in CP437 or CP850 page code is '╜' which corresponds to character U+255C as iconv says. 0xBD is the '½' character (U+00BD) in the CP1252 page code usually used by Windows. I therefore think this bug is invalid, please close it if you agree. -- Aurelien Jarno GPG: 4096R/1DDD8C9B aurel...@aurel32.net http://www.aurel32.net -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742942: Bug#749438: gparted stuck at searching /dev/sdc partitions
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 5/27/2014 4:27 AM, Aurelien Jarno wrote: That is wrong. 0xBD in CP437 or CP850 page code is '╜' which corresponds to character U+255C as iconv says. Not according to the table listed on http://en.wikipedia.org/wiki/Code_page_437. According to it, 0xBD is the 1/2 character. Also that is what 0xBD is shown as under emacs, as well as in Windows notepad. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJThJuMAAoJEI5FoCIzSKrw9isIAIQq7K7rKDy04kiiS+z9/4pj 9Z+FPZnIP6MfuO1FW10QPPE3Mp8uV6y7954/ijMLW6U/q5cUlxisPoN1YIwa/au3 2USSzg84NKSmeTf5AOceg9ZcH+ZSYm9SCd0NJZNMsZRmLleZ6GQf6VW7U2vKcV+l lCL2V0wwhLxZPqvdhMh+DO/sdwzX9izZ8DFG1t/dulTTJ6KYJnBQ32WEM1+RKBZj NLo6gTxZKDbSZ/7u1Z1aln/t9I4sbk7wat00htW66QhTjvU5IVgI9KRVpZcNvyx6 m8tSXkdruEhqbQSKYue6hvS+zbR1zfEvPVyuVcK0Fx8G6e3qT0+VGVbrp0bsbi0= =UHZB -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742942: Bug#749438: gparted stuck at searching /dev/sdc partitions
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 5/27/2014 10:05 AM, Phillip Susi wrote: On 5/27/2014 4:27 AM, Aurelien Jarno wrote: That is wrong. 0xBD in CP437 or CP850 page code is '╜' which corresponds to character U+255C as iconv says. Not according to the table listed on http://en.wikipedia.org/wiki/Code_page_437. According to it, 0xBD is the 1/2 character. Also that is what 0xBD is shown as under emacs, as well as in Windows notepad. It seems that iconv thinks that U+00BD maps to 0xAB in CP-437, which seems to be incorrect according to that wikipedia page, as well as this one from Microsoft: http://msdn.microsoft.com/en-us/library/aa245259%28v=vs.60%29.aspx And http://en.wikipedia.org/wiki/ISO/IEC_8859-1. 8859 is supposed to be a superset of cp-437 and also lists the 1/2 character as 0xBD. Now this is where things get really weird. On Windows, if I open notepad and type alt-0189 to enter the ½ character, it saves it as 0xBD. If I type that file from a command prompt however, it shows up as the '╜'. If I cat the file from a mingw32 bash shell, it shows up as ½. If I open it in Word, it asks me how to interpret it. If I select Windows Western European, it shows up as ½, and I think this corresponds to windows CP-1252, which also has that character as 0xBD according to wikipedia. If I tell Word to use MS-DOS ( OEM Default ), then it comes up as ╜ instead. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJThNiGAAoJEI5FoCIzSKrwI14IAKwErB9PLwpw2FJgDHOYtJRS iW0QkhFopWMgF3CLRFQXNslNp3BJHfQh2PQOIVtSmOOtRwQPUcPL1VYmtBTfVR+/ VjzyZae16Jnbtso6TzsuhpKdBMqVqvdo5UH2cYvSx58Pnpft7yfkqAkOMWNJs0vc MlJm0skffXWeB+kmZ2GWqrz19YW/CXOxQK+nOmJ8Wa7ajkNZlbvvlTdx/sdb6ywZ HeD5u+SsVWfnPLrBe4/+xG2hO4gInTQnnbIXOU1bqNKBGbgoojS5t/hp0Vm/w7lA SQ7pKiUkx1npg3K4cUrmDYfdb9aGjhJaluZuOMVfBjel5XqN8IWt5/YiedG1yv0= =EfPb -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#742942: Bug#749438: gparted stuck at searching /dev/sdc partitions
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I'm not sure why this conversation got dragged over to the wrong bug #, but I'm putting it back and closing it. I realized finally that I was just looking at the table wrong and the value I was looking at was in fact, the UTF-16 value for the character rather than the value in that CP. It is very odd that this character is not the same value in CP-437 and CP-1252/ISO-8859-1, but that does appear to be the case. -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQEcBAEBAgAGBQJThOXKAAoJEI5FoCIzSKrwMFIIAJOz5e/zqcafS02hjDQNhjoA I5rPGN1XxXQh22jpX/sO37mw2ZZDuAEyQa3DMatIZ9OlB0T1RQ7p0Z8OCLw/MgJQ Bx/SLVL94AH/0CDuqT93oHqBCkPxEpdaQ5/ooiWTVm39QjAFvgU6wPZRwRRwtg7A TZqgqy4zZePBEMrusGygAv4sW9VBslth4bLSFfywphUSYzVAGXnwRjflL7vIavHv JYF6R0Fgy69d+nSlEKLs1OEF/GSluCk+grz7nF1Gf9vwmQG+QcIyVOYXRWVt26jj Ic4mT609AV6illyGe6cjeYhYEmJ67wgQs368TzMCw6iIbV7yjYaS1ArL5BwnSj8= =5tG5 -END PGP SIGNATURE- -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
