Elog crashes with null Username, posted by Alan Grant on Sun Dec 18 08:51:47 2016
|
I haven't found any reporterd issues in the forum similar yet, but it appears there is a bug in Elog when logging into logbooks. If I leave Username and Password null and click Submit the daemon crashes. We've been having this problem off and on and after some verbose logging level 3 I was drawn to these recurring lines in the log:
16-Dec-2016 18:20:22 [172.23.113.4] {SER Reports} LOGIN user "" (attempt)
16-Dec-2016 23:15:52 [] Server listening on port 8080 ..
16-Dec-2016 23:18:05 [172.23.113.4] {Daily Request Log} LOGIN user "dmorrison" (attempt) |
Re: Elog crashes with null Username, posted by Stefan Ritt on Mon Dec 19 12:28:47 2016
|
Ups. This bug must have lingered there since the beginning of time. Funny that nobody noticed in the last ten years or so. I have fixed it in the current git revision.
| Alan Grant wrote: |
|
I haven't found any reporterd issues in the forum similar yet, but it appears there is a bug in Elog when logging into logbooks. If I leave Username and Password null and click Submit the daemon crashes. We've been having this problem off and on and after some verbose logging level 3 I was drawn to these recurring lines in the log:
16-Dec-2016 18:20:22 [172.23.113.4] {SER Reports} LOGIN user "" (attempt)
16-Dec-2016 23:15:52 [] Server listening on port 8080 ..
16-Dec-2016 23:18:05 [172.23.113.4] {Daily Request Log} LOGIN user "dmorrison" (attempt)
|
|
|
Re: Elog crashes with null Username, posted by Alan Grant on Mon Dec 19 19:42:05 2016
|
Thanks for the speedy fix.
FYI, our other Elog instances which are running ealrier versions do not exhibit this problem, as I confirmed that entering null ID/password returns "Invalid Username or password" as expected. This may be why it wasn't mentioned by anyone else before. Maybe something got inadvertently dropped in the newer version.
| Stefan Ritt wrote: |
|
Ups. This bug must have lingered there since the beginning of time. Funny that nobody noticed in the last ten years or so. I have fixed it in the current git revision.
| Alan Grant wrote: |
|
I haven't found any reporterd issues in the forum similar yet, but it appears there is a bug in Elog when logging into logbooks. If I leave Username and Password null and click Submit the daemon crashes. We've been having this problem off and on and after some verbose logging level 3 I was drawn to these recurring lines in the log:
16-Dec-2016 18:20:22 [172.23.113.4] {SER Reports} LOGIN user "" (attempt)
16-Dec-2016 23:15:52 [] Server listening on port 8080 ..
16-Dec-2016 23:18:05 [172.23.113.4] {Daily Request Log} LOGIN user "dmorrison" (attempt)
|
|
|
|
Elog source code giving errors when compiling about missing header files, posted by rahul bhandari on Wed Dec 21 20:15:19 2016
|
I downloaded the source code from the git repository which contained the new fix that was made for the missing username-crash error. I tried compiling the elogd.c file using a GCC compiler and it gave an error about missing header files. It first gave an error about netdb.h file and when I commented that header file call, it gave further errors for other header files. I do not really understand why it gives an error about that. |
|
Re: Elog source code giving errors when compiling about missing header files, posted by Stefan Ritt on Thu Dec 22 09:41:35 2016
|
You need the netdb.h header file to compile elog. I see that you compile under Windows. You have to make sure that your compiler installation comes with all header files. The netdb.h file is usually in /usr/include under Unix.
| rahul bhandari wrote: |
|
I downloaded the source code from the git repository which contained the new fix that was made for the missing username-crash error. I tried compiling the elogd.c file using a GCC compiler and it gave an error about missing header files. It first gave an error about netdb.h file and when I commented that header file call, it gave further errors for other header files. I do not really understand why it gives an error about that.
|
|
|
elogd crashes during Find and Mirror operations under MacOS, posted by Andreas Warburton on Sat Dec 24 10:40:00 2016
|
When installing ELOG 3.1.2 on my Mac Sierra (10.12.2) with the latest XCode 8.2.1, it seemed to install fine but for one warning when building elogd:
cc -O3 -funroll-loops -fomit-frame-pointer -W -Wall -Wno-deprecated-declarations -DHAVE_SSL -I /opt/local/include -o elogd
src/elogd.c crypt.o auth.o regex.o mxml.o -lssl
src/elogd.c:13840:13: warning: '__builtin___strlcpy_chk' will always overflow
destination buffer [-Wbuiltin-memcpy-chk-size]
strlcpy(user_email, user_list[i], NAME_LENGTH);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/secure/_string.h:105:3: note: expanded from macro 'strlcpy'
__builtin___strlcpy_chk (dest, src, len, __darwin_obsz (dest))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
My installation was done with "sudo make install", as the instructions state. I have also been using the prescribed "launchctl" approach.
1. Is the above warning something that needs attention/correcting?
2. My elogd crashes now every time that I attempt any kind of "Find" operation.
3. I synchronize my MacOS version of elog with a linux-based version on my university's web server. I have installed 3.1.2 on the linux server, and it is running there without problems. When I now try to synchronize my 3.1.2 ELOG on MacOS laptop with the same version on my linux server, it appears to do the first mirroring operation correctly, but elogd crashes when I try a second synchronization at a later time. (For now, I have "Mirror simulate = 1" to prevent inadvertent damage.)
Can anyone help me with the above points?
Thank you,
Andreas
|
|
Re: elogd crashes during Find and Mirror operations under MacOS, posted by Andreas Luedeke on Tue Jan 3 09:20:23 2017
|
I have no experience with ELOG on OSX, therefore I can only give very vague and indirect support (sorry).
I suggest that you first try an installation without actual logbook data (just move it to a save place for a while). Does a "find" still crash elogd?
If elogd is still crashing, you should prune your configuration until it stops crashing. Submit the smallest config here that still crashes the logbook.
Since the logbook starts crashing after a sync, I would rather guess that it is somehow related to specific logbook entries. It would be helpful if you could isolate the offending entries.
Another approach would be to compile elogd with "make debug" and run it in the debugger. Once it crashes the debugger will show where it happened.
Happy New Year! Andreas
| Andreas Warburton wrote: |
|
When installing ELOG 3.1.2 on my Mac Sierra (10.12.2) with the latest XCode 8.2.1, it seemed to install fine but for one warning when building elogd:
cc -O3 -funroll-loops -fomit-frame-pointer -W -Wall -Wno-deprecated-declarations -DHAVE_SSL -I /opt/local/include -o elogd
src/elogd.c crypt.o auth.o regex.o mxml.o -lssl
src/elogd.c:13840:13: warning: '__builtin___strlcpy_chk' will always overflow
destination buffer [-Wbuiltin-memcpy-chk-size]
strlcpy(user_email, user_list[i], NAME_LENGTH);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/secure/_string.h:105:3: note: expanded from macro 'strlcpy'
__builtin___strlcpy_chk (dest, src, len, __darwin_obsz (dest))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
My installation was done with "sudo make install", as the instructions state. I have also been using the prescribed "launchctl" approach.
1. Is the above warning something that needs attention/correcting?
2. My elogd crashes now every time that I attempt any kind of "Find" operation.
3. I synchronize my MacOS version of elog with a linux-based version on my university's web server. I have installed 3.1.2 on the linux server, and it is running there without problems. When I now try to synchronize my 3.1.2 ELOG on MacOS laptop with the same version on my linux server, it appears to do the first mirroring operation correctly, but elogd crashes when I try a second synchronization at a later time. (For now, I have "Mirror simulate = 1" to prevent inadvertent damage.)
Can anyone help me with the above points?
Thank you,
Andreas
|
|
|
Re: elogd crashes during Find and Mirror operations under MacOS, posted by Stefan Ritt on Tue Jan 3 09:44:18 2017
|
The current GIT revision (c017d07) fixes the problem with the strlcpy() function (macOS 10.12.2, XCode 8.2). Can you try if that fixes your other problems, too?
Plese remember to use the recursive clone to get that version:
git clone https://bitbucket.org/ritt/elog --recursive
Stefan
| Andreas Warburton wrote: |
|
When installing ELOG 3.1.2 on my Mac Sierra (10.12.2) with the latest XCode 8.2.1, it seemed to install fine but for one warning when building elogd:
cc -O3 -funroll-loops -fomit-frame-pointer -W -Wall -Wno-deprecated-declarations -DHAVE_SSL -I /opt/local/include -o elogd
src/elogd.c crypt.o auth.o regex.o mxml.o -lssl
src/elogd.c:13840:13: warning: '__builtin___strlcpy_chk' will always overflow
destination buffer [-Wbuiltin-memcpy-chk-size]
strlcpy(user_email, user_list[i], NAME_LENGTH);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/secure/_string.h:105:3: note: expanded from macro 'strlcpy'
__builtin___strlcpy_chk (dest, src, len, __darwin_obsz (dest))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
My installation was done with "sudo make install", as the instructions state. I have also been using the prescribed "launchctl" approach.
1. Is the above warning something that needs attention/correcting?
2. My elogd crashes now every time that I attempt any kind of "Find" operation.
3. I synchronize my MacOS version of elog with a linux-based version on my university's web server. I have installed 3.1.2 on the linux server, and it is running there without problems. When I now try to synchronize my 3.1.2 ELOG on MacOS laptop with the same version on my linux server, it appears to do the first mirroring operation correctly, but elogd crashes when I try a second synchronization at a later time. (For now, I have "Mirror simulate = 1" to prevent inadvertent damage.)
Can anyone help me with the above points?
Thank you,
Andreas
|
|
|