Bug#785404: hurd: opendir(/dev/null) succeeds
Package: libc0.3
Version: 2.19-18+b1
richard@heceptor:~/junk$ cat t.c
#include dirent.h
#include assert.h
#include stdio.h
#include errno.h
#include string.h
int main() {
DIR *dp = opendir(/dev/null);
printf(%p\n, dp);
struct dirent *de = readdir(dp);
printf(%p %d %s\n, de, errno, strerror(errno));
return 0;
}
richard@heceptor:~/junk$ gcc -o t t.c ./t
0x804a0e0
(nil) 1073741844 Not a directory
ii hurd 1:0.6-4 hurd-i386GNU Hurd
ii libc0.3:hurd-i 2.19-18+b1 hurd-i386GNU C Library: Shared libraries
Based on Glibc history there used to be some defence against this, but
it was removed in 2006.
ttfn/rjk
--
To UNSUBSCRIBE, email to debian-glibc-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/55562e96.3080...@greenend.org.uk
Bug#954230: In a sticky directory root cannot write to files owned by normal users
Sounds like protected_regular from https://www.kernel.org/doc/Documentation/sysctl/fs.txt ttfn/rjk
Bug#357051: [PATCH] fix sorting/search example
Patch attached.
ttfn/rjk
--- manual/examples/search.c~ 2009-12-08 20:10:20.0 +
+++ manual/examples/search.c2011-02-05 13:32:59.0 +
@@ -36,8 +36,9 @@
/* This is the comparison function used for sorting and searching. */
int
-critter_cmp (const struct critter *c1, const struct critter *c2)
+critter_cmp (const void *c1v, const void *c2v)
{
+ const struct critter *c1 = c1v, *c2 = c2v;
return strcmp (c1-name, c2-name);
}
Bug#458070: tzdata forgot my timezone
Package: tzdata Version: 2007j-1etch1 Preparing to replace tzdata 2007b-1 (using .../tzdata_2007j-1etch1_all.deb) ... Unpacking replacement tzdata ... Setting up tzdata (2007j-1etch1) ... Running 'tzconfig' to set this system's timezone. Your current time zone is set to Unknown Do you want to change that? [n]: The system previously knew perfectly well what my timezone was and had done for over a decade. ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#458070: tzdata forgot my timezone
Aurelien Jarno writes: Aurelien Jarno wrote: Richard Kettlewell a écrit : Nothing has changed in this new version, except the contents of the timezones. What was your timezone settings *before* the upgrade? It looks like it was already broken before the upgrade, like for example a missing /etc/timezone. Any news about that? Short of a time machine it's rather hard to tell! However, I've not seen this behavior on any other machines. Perhaps there is a bug in something else entirely which caused the deletion? ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#506420: incorrect hwcap line in /etc/ld.so.conf.d/libc6-xen.conf?
Package: libc6-xen Version: 2.7-16 As shipped, /etc/ld.so.conf.d/libc6-xen.conf read as follows: # This directive teaches ldconfig to search in nosegneg subdirectories # and cache the DSOs there with extra bit 1 set in their hwcap match # fields. In Xen guest kernels, the vDSO tells the dynamic linker to # search in nosegneg subdirectories and to match this extra hwcap bit # in the ld.so.cache file. hwcap 1 nosegneg However I still got thousands of 4gb seg fixup messages and ldd revealed that the runtime linker was not using the Xen-friendly Libc. Changing the 1 to 0, based on a mailing list posting I found, fixed the problem. Unfortunately I've not found where any of this is documented (the ldconfig man page does not contain any useful pointers) so this is slightly guesswork on my part. Perhaps the man page could be improved. (I'm using etch's Xen and running testing and unstable VMs; the kernel in all cases is 2.6.18-6-xen-686.) ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#173486: can't build stable's glibc
Package: glibc Version: 2.2.5-11.2 I get segfaults in tests when trying to build glibc-2.2.5-11.2, on stable. A full transcript can be found at: http://www.greenend.org.uk/rjk/junk/glibc.build.gz The build doesn't stop as soon as a problem is found. However looking through the typescript I found what looks like a couple of segfaults at lines 23398 and 23995. make[4]: *** [/export/usr.src/glibc/glibc-2.2.5/i386-linux/obj/localedata/tst_strfmon.out] Error 139 make[4]: *** [/export/usr.src/glibc/glibc-2.2.5/i386-linux/obj/libio/tst-widetext.out] Error 139 I ran the build twice and the same problem occurred in the same place both times (I diff'd the typescript). ttfn/rjk ii gcc2.95.4-14 The GNU C compiler. ii gcc-2.95 2.95.4-11woody The GNU C compiler. ii gettext0.10.40-5 GNU Internationalization utilities ii make 3.79.1-14 The GNU version of the make utility. ii dpkg-dev 1.9.21 Package building tools for Debian ii debianutils1.16 Miscellaneous utilities specific to Debian. ii tar1.13.25-2 GNU tar ii bzip2 1.0.2-1A high-quality block-sorting file compressor ii texinfo4.1-2 Documentation system for on-line information ii texi2html 1.64-cvs200104 Convert Texinfo files to HTML ii file 3.37-3.1 Determines file type using magic numbers ii autoconf 2.54-2 automatic configure script builder ii libc6 2.2.5-11.2 GNU C Library: Shared libraries and Timezone ii libc6-dev 2.2.5-11.2 GNU C Library: Development Libraries and Hea ii binutils 2.13.90.0.10-1 The GNU assembler, linker and binary utiliti ii automake1.41.4-p6-5 A tool for generating GNU Standards-complian ii automake1.71.7.1-1A tool for generating GNU Standards-complian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#173486: 173486 is not fixed and therefore should not be closed.
reopen 173486 quit The bug still exists, even if you've decided not to do anything about it! ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#173486: can't build stable's glibc
Package: glibc Version: 2.2.5-11.2 I get segfaults in tests when trying to build glibc-2.2.5-11.2, on stable. A full transcript can be found at: http://www.greenend.org.uk/rjk/junk/glibc.build.gz The build doesn't stop as soon as a problem is found. However looking through the typescript I found what looks like a couple of segfaults at lines 23398 and 23995. make[4]: *** [/export/usr.src/glibc/glibc-2.2.5/i386-linux/obj/localedata/tst_strfmon.out] Error 139 make[4]: *** [/export/usr.src/glibc/glibc-2.2.5/i386-linux/obj/libio/tst-widetext.out] Error 139 I ran the build twice and the same problem occurred in the same place both times (I diff'd the typescript). ttfn/rjk ii gcc2.95.4-14 The GNU C compiler. ii gcc-2.95 2.95.4-11woody The GNU C compiler. ii gettext0.10.40-5 GNU Internationalization utilities ii make 3.79.1-14 The GNU version of the make utility. ii dpkg-dev 1.9.21 Package building tools for Debian ii debianutils1.16 Miscellaneous utilities specific to Debian. ii tar1.13.25-2 GNU tar ii bzip2 1.0.2-1A high-quality block-sorting file compressor ii texinfo4.1-2 Documentation system for on-line information ii texi2html 1.64-cvs200104 Convert Texinfo files to HTML ii file 3.37-3.1 Determines file type using magic numbers ii autoconf 2.54-2 automatic configure script builder ii libc6 2.2.5-11.2 GNU C Library: Shared libraries and Timezone ii libc6-dev 2.2.5-11.2 GNU C Library: Development Libraries and Hea ii binutils 2.13.90.0.10-1 The GNU assembler, linker and binary utiliti ii automake1.41.4-p6-5 A tool for generating GNU Standards-complian ii automake1.71.7.1-1A tool for generating GNU Standards-complian
Bug#173486: 173486 is not fixed and therefore should not be closed.
reopen 173486 quit The bug still exists, even if you've decided not to do anything about it! ttfn/rjk
Bug#234813: libc6 postinst overmounted my /tmp
Package: libc6 Version: 2.3.2.ds1-11 I upgraded libc6 and found that /etc/init.d/mountkernfs had remounted my /dev/shm, not noticing that it was already mounted. The new mount hid the old filesystem. /dev/shm is actually a link to /tmp on my system, which is why the detection didn't work. This broke all sorts of things, e.g. new X connections (as /tmp/.X11-unix was no longer visible.) Fortunately umount did the right thing. Perhaps mountkernfs could be changed to use some more robust way of detecting whether /dev/shm is already mounted - for instance, by looking at the first column of 'df /dev/shm'. ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#64865: error messages from glibc-doc setup
Package: glibc-doc Version: 2.1.3-10 Setting up glibc-doc (2.1.3-10) ... sh: /usr/share/info/libc.info.gz: No such file or directory install-info: read gzip -d /usr/share/info/libc.info.gz |: 256 Registering GNU C Library Documentation...done. lyonesse# echo $? 0 ttfn/rjk
Bug#64865: additional information
It seems not all the files were present. This may reflect a bug in dpkg. However, shouldn't the postinst have exited nonzero? ttfn/rjk
Bug#234813: libc6 postinst overmounted my /tmp
Package: libc6 Version: 2.3.2.ds1-11 I upgraded libc6 and found that /etc/init.d/mountkernfs had remounted my /dev/shm, not noticing that it was already mounted. The new mount hid the old filesystem. /dev/shm is actually a link to /tmp on my system, which is why the detection didn't work. This broke all sorts of things, e.g. new X connections (as /tmp/.X11-unix was no longer visible.) Fortunately umount did the right thing. Perhaps mountkernfs could be changed to use some more robust way of detecting whether /dev/shm is already mounted - for instance, by looking at the first column of 'df /dev/shm'. ttfn/rjk
Bug#111767: syslog.h bug persists
GOTO Masanori writes: Neil Spring wrote: 2.2.5-14.3 still has the sys/syslog.h SYSLOG_NAMES bug. it'd be great if someone can apply: - char*c_name; + const char *c_name; or report it upstream. I realize there are bigger problems to deal with; thanks for your hard work. This report and proposed patch was rejected by upstream. Upstream explained that if we fix with this kind of patch for removing -Wwrite-strings warnings, some user programs (which don't use const char) start compilation warnings. It is never safe to write to a string literal so there is no advantage to them in treating them as non-const. I think those programs should be fixed instead of imposing this bug on correct programs. SYSLOG_NAMES is used for obsolete purpose, Please can you say what we should use instead? and it's not well used these days, so their argument seems one of accountable opinion for me. Well, I think there're various opinions, so you could discuss with it to upstream. I close this bug. ttfn/rjk -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#357051: libc.info s9.4 qsort example is wrong
Package: glibc-doc
Version: 2.3.2.ds1-22
The example for qsort has, among other things:
int
critter_cmp (const struct critter *c1, const struct critter *c2)
{
return strcmp (c1-name, c2-name);
}
/* ... */
result = bsearch (target, muppets, count, sizeof (struct critter),
critter_cmp);
/* ... */
qsort (muppets, count, sizeof (struct critter), critter_cmp);
This is incorrect. The comparison function must be declared as
int critter_cmp (const void *c1, const void *c2)
and should use casts to convert to the proper type.
ttfn/rjk
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
