| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Can be optimized a bit by checking stat.st_mtime.
Currently, every reload looks like this:
14:37:49.551073 open("/etc/abrt/abrt.conf", O_RDONLY) = 13
14:37:49.551229 fstat(13, {st_mode=S_IFREG|0644, st_size=869, ...}) = 0
14:37:49.551393 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f541037f000
14:37:49.551505 read(13, "# With this option set to \"yes\",\n# only crashes in signed packages will be analyzed.\n# the list of "..., 4096) = 869
14:37:49.551709 read(13, "", 4096) = 0
14:37:49.551812 close(13) = 0
14:37:49.552051 open("/etc/abrt/gpg_keys", O_RDONLY) = 13
14:37:49.552183 fstat(13, {st_mode=S_IFREG|0644, st_size=36, ...}) = 0
14:37:49.552338 mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7f541037f000
14:37:49.552448 read(13, "/etc/pki/rpm-gpg/RPM-GPG-KEY-fedora\n", 4096) = 36
14:37:49.552591 read(13, "", 4096) = 0
14:37:49.552692 close(13) = 0
14:37:49.552792 munmap(0x7f541037f000, 4096) = 0
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
|
| |
This loses libstdc++, libm and libgcc_s libraries from abrtd
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
| |
This makes abrt-action-save-package-data twice as small, amd loses libstdc++.so.6
dependency
|
|
|
|
|
|
|
|
| |
This change will not compile, because in C++, void* cannot be
automatically cast to char*. Next change renames Settings.cpp to *.c
and fixes this.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In preparation for
https://fedorahosted.org/abrt/ticket/189
"abrtd should reload configuration when some of it's config files changes"
I would like to stop running abrt-dump-oops as a child of abrtd.
We experimented with it: we made abrt-ccpp hook (de)installation
a separate service, and in my opinion it ended up a better
solution than making it a part of daemon. Among other things,
it made abrt and abrt-addon-ccpp packages independent.
This change does the same for abrt-dump-oops.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
|
|
|
|
|
|
|
| |
Update sha1 code to a smaller version.
Verified that ccpp hashing produces the same results as the old code.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
When, say, reported_to is changed, inotify doesn't see it
two dir levels up. abrt-gui needs a bit of help from wizard
by sending SIGCHLD, and treating it as "please rescan" request.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
|
|
|
| |
The change is somewhat big because I can no longer (ab)use tooltip to store
event name. So I created a bit more generic method of remembering
which events are shown and selected.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
|
|
| |
Took it from firewall cfg GUI. One useful effect is that
event log and event result msg above it are more clearly separated.
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|\ |
|
| |
| |
| |
| | |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This makes descriptions resizable, so long descriptions will no longer
make windows insanely wide. (Can't do that with checkboxes. Tried everything)
We need longer descriptions for https://fedorahosted.org/abrt/ticket/202
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
| |
| |
| | |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
| |
| |
| |
| | |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|/
|
|
|
|
|
| |
- the first part analyzes the coredump and writes the required
build-ids into a file
- the second app reads the file and tries to find the debuginfo
packages for build-ids
|
|
|
|
| |
Signed-off-by: Nikola Pajkovsky <npajkovs@redhat.com>
|
| |
|
| |
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
elements of dump directory.
This patch changes it to generic handling, where abrt-cli doesn't know
anything about particular elements (like backtrace).
Changes in detail:
* crash_item->flags how has CD_FLAG_UNIXTIME bit.
* format_crash_item(item) returns malloced formatted string
(currently only formatted time for items with CD_FLAG_UNIXTIME)
or NULL, if item->content is to be used as-is
* crash_item->flags how has CD_FLAG_LIST bit, it is set on a
small number of elements which are to be shown by abrt-cli -l.
* abrt-cli -l doesn't use fixed names, it looks at CD_FLAG_LIST.
* abrt-cli -i doesn't use fixed names, it prints all one-line elements:
Directory : /var/spool/abrt/ccpp-1298264192-2705
analyzer : CCpp
architecture : x86_64
cmdline : metacity
comment : Abrt testing, please disregard
component : metacity
coredump : /var/spool/abrt/ccpp-1298264192-2705/coredump
count : 1
crash_function : pa_atomic_load
executable : /usr/bin/metacity
global_uuid : 1e8b716d2094fb22dccdb5321ac0cf6f14eb6689
hostname : dhcp-25-227.brq.redhat.com
kernel : 2.6.34.7-61.fc13.x86_64
os_release : Fedora release 13 (Goddard)
package : metacity-2.30.0-3.fc13
rating : 4
reason : Process /usr/bin/metacity was killed by signal 11 (SIGSEGV)
time : Mon 21 Feb 2011 05:56:32 AM CET
uid : 500
uuid : 453085d0f703b96ddc3a5172dd7d5a29479f5b3f
* abrt-cli -i --backtrace is removed. Instead, abrt-cli -i --full
is to be used. It shows all elements, one-line and multi-line ones.
* abrt-cli code now has an example how to list crast_data_t
in the sorted-by-key order.
Further ideas:
abrt-cli needs -o elem1,elem2,elem3 option which tells which elements to
show, and/or -O elem1,elem2,elem3 option which tells elements to NOT
show. Rationale: abrt-cli -i --full output is way too long because of
smaps, need a way to suppress unneeded elements display
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Denys Vlasenko <dvlasenk@redhat.com>
|
|
|
|
| |
Signed-off-by: Nikola Pajkovsky <npajkovs@redhat.com>
|
| |
|
|\ |
|
| |\ |
|
| | |
| | |
| | |
| | | |
Signed-off-by: Nikola Pajkovsky <npajkovs@redhat.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Nikola Pajkovsky <npajkovs@redhat.com>
|
| | |
| | |
| | |
| | | |
Signed-off-by: Nikola Pajkovsky <npajkovs@redhat.com>
|
| |/
| |
| |
| |
| | |
- also added default event configuration for rhtsupport
- and tweaked a spec file a little
|