Package: locales-all
Version: 2.3.6-18
When testing locales-all with piuparts, we get the following error:
0m54.7s ERROR: Package purging left files on system:
/usr/lib/locale
/usr/lib/locale/locale-archive
The full piuparts log file is at
http://piuparts.cs.helsinki.fi/bugged/locales-all_2
la, 2006-04-29 kello 20:56 +0200, Rainer Zocholl kirjoitti:
> Install libc6 fails because(?) no sed was installed but seems to
> be used in libc6.postinst.
sed is essential, so always guaranteed to be there:
[EMAIL PROTECTED] dpkg --status sed
Package: sed
Essential: yes
--
"Quack, damn you!" –
While pondering the feedback on locales-all (which I hope to get back to
soon), I've played a bit with creating a timezones-all package.
Rationale: On many small systems, such as routers and stuff, with little
disk space, there is no need for any time zone except UTC. The time zone
database is abo
ke, 2005-09-28 kello 14:33 +0200, Florent Bayle kirjoitti:
> After running dpkg-reconfigure locales and choosing [EMAIL PROTECTED] as
> default system environment locale, all locales remains set to "POSIX" :
>
> jupiter:/home/florent# sudo dpkg-reconfigure locales
> Generating locales (this might
su, 2005-09-25 kello 19:00 +0300, Lars Wirzenius kirjoitti:
> Generating the three locales needed on my mail/web/etc server, which is
> used by several people, takes about 15 seconds. Locales with more
> complicated character sets, especially many Asian languages, will make
> this r
There has occasionally been talk of a "locales-all" package, which would
contain compiled binary forms of all locales. This would use up a bit
more bandwidth and disk space in order to save CPU time: generating the
locales can take a long time, if you need many and don't have a fast
processor.
I'v
Package: libc6
Version: 2.3.2.ds1-22
Severity: normal
File: /sbin/ldconfig
The ldconfig manual page says it creates symlinks, but does not mention
anything about deleting them. It does remove them, however, and Policy
requires shared library packages to run ldconfig in their postrms for
this reaso
Hi, gotom, and others.
For a while, I will be able to work full time on Debian things, and
would like to help with maintain glibc for Debian. I don't aim at
becoming a co-maintainer as such, but working on the bugs reported
against glibc and sending comments and patches is what I had in mind.
How
> [EMAIL PROTECTED]:~ > TZ=IDIOT date
> Wed Jan 19 19:54:58 IDIOT 2000
The above is a concise summary of the problem: if TZ is invalid, date
(tzset, really) sets the time zone to UTC, but uses the invalid name,
which confuses users.
In 2002, Andrew Suffield sent a patch to this bug report that cl
su, 2005-06-19 kello 17:25 +1000, Nathan Hand kirjoitti:
> Package: nscd
> Version: 2.3.5-1ubuntu7
Does this also happen with a Debian version? Debian's versions of 2.3.5
are in experimental, though.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contac
> The getgrname(3) man page says:
>
> The getgrnam() function returns a pointer to a structure containing the
> group information from /etc/group for the entry that matches the group
> name name.
The glibc info document says this:
-- Function: struct group * getgrnam (const char *NAME)
tag 177940 + fixed-in-experimental
thanks
I used the attached program (based on the test in autoconf) to test that
this bug occurs with libc6 2.3.2.ds1-22 (the version in unstable) but
not with version 2.5.2-2 (the version in experimental). Thus, I tag it
fixed-in-experimental.
#include
#include
The attached dpatch file enabled me to compile version 2.3.5-1 from
experimental with /bin/sh pointing at dash. It's a very simple change by
now, I just copied the remaining unfixed part from Eric Wong's latest
mail in this bug report.
dash.dpatch
Description: application/shellscript
to, 2005-06-16 kello 12:10 +0200, Baurzhan Ismagulov kirjoitti:
> On Thu, Jun 16, 2005 at 12:47:13PM +0300, Lars Wirzenius wrote:
> > The -std=c99 option means that you want strict compliance to the 1999
> > version of the C standard. That standard does not define struct timespec
&
to, 2005-06-16 kello 10:47 +0200, Baurzhan Ismagulov kirjoitti:
> #include
>
> int main(void)
> {
> struct timespec a;
> nanosleep(&a, &a);
> return 0;
> }
>
> Compilation with "gcc -Wall -g -std=c99" produces the following errors:
>
> c.c: In Funktion »main«:
> c.c:5: error:
ti, 2005-05-24 kello 12:11 +0200, Rolf Leggewie kirjoitti:
> for some reason I do not fully understand locale-gen fails on my system.
> The only thing unusual in this respect is that /usr is mounted over NFS
> (usually ro, but set to rw before calling locale-gen). no-root-squash
> is set for t
Package: libc6
Version: 2.3.2.ds1-22
Severity: minor
The English output of "localedef --help" contains this:
--posixBe strictly POSIX conform
The last word should surely be "conforming".
-- System Information:
Debian Release: 3.1
APT prefers unstable
APT policy: (500,
8 character set
+and add it to the default locale archive with the name
+.BR fi_FI.UTF-8 :
+.PP
+.RS
+localedef \-f UTF\-8 \-i fi_FI fi_FI.UTF\-8
+.RE
+.PP
+The same, but generate files into the current directory (note that the
+last argument must then contain a slash):
+.PP
+.RS
+localedef \-f UTF\-
It seems to me that gotom's analysis is correct: gethostbyname
interprets an all-numeric-no-dots argument string as an IP address; see
(libc.info.gz)Abstract Host Addresses for details. What this ultimately
means is that gethostbyname cannot resolve an all-numeric hostname into
a fully qualified do
Since it seems that netgroups in /etc/hosts.equiv haven't worked since
1997, and since NIS is anyway a security risk, and further since ssh can
do these things in better ways, it seems to me that fixing the bug by
adjusting the documentation to mention that "[EMAIL PROTECTED]" is not, in fact,
supp
I couldn't reproduce this either, using a trivial test case, and the
submitter says he can't either. Could this bug be closed?
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Attached is my attempt at writing a clear, yet succinct example for
reading directory listings using opendir/readdir/closedir. I think mine
is easier to follow (a more straightforward structure), but I'd be happy
enough with the patch by H. S. Teoh.
Once sarge is released and new libc packages can
In case it is of any use, here is the test case file
(stdio-common/bug-printf-0ll.c) I wrote while trying to fix the bug.
Gotom said on IRC that he has an actual fix for the bug, and my fix
wasn't a fix (need to learn to distinguish "unsigned long" from
"unsigned long long" even in the middle of th
Attached please find a very simple patch for
debian/local/manpages/iconv.1 to convert unescaped hyphens to escaped
ones. This means that man won't create troublesome Unicode characters
when it operates in a Unicode locale.
I patched iconv.1 instead of iconv.pod because I don't see in perlpod(1)
a
Package: glibc-doc
Version: 2.3.2.ds1-18
Severity: normal
The glibc documentation for mallopt does not describe its return value.
Nor does it document all parameter values (M_MXFAST is missing).
>From the source:
mallopt(int parameter_number, int parameter_value)
Sets tunable parameters The f
Package: glibc-doc
Version: 2.3.2.ds1-18
Severity: normal
file:///usr/share/doc/glibc-doc/html/libc_3.html#SEC59
libc.info(Resizing the Data Segment) (I hope that is the correct way to
refer to a node in an info file).
This section claims that the prototype for the sbrk function is:
in
I noticed that mtrace.1 (in debian/local/manpages) has the same problem
with \fT instead of \fR causing the rest of the paragraph be bold rather
than only the single word.
Of course, if these pages are automatically generated, fixing the
generator would be a good idea.
--
To UNSUBSCRIBE, ema
Package: libc6-dev
Version: 2.3.2.ds1-18
Severity: minor
The manual page for gencat has a "\fT" instead of a "\fR" which causes
all of the first paragraph to be boldface rather than just the word
"gencat". I think it makes sense to mark "catgets" as boldface as well,
and it might be good to add a
28 matches
Mail list logo
