Re: Compile issues on Fedora withe current elog source, posted by Stefan Ritt on Wed Sep 12 11:10:18 2018
|
This warning is triggered by the use of the sprintf() funciton, which can write beyond the boundary of the destination string. I replaced it by snprintf(),
which should make the compiler happy. Code is committed. Unfortuantely I have no FC27 here, so if there is still some waring with the current code, please
post the full warning list here. |
|
Re: Compile issues on Fedora withe current elog source, posted by Allen Tuttle on Wed Sep 12 15:20:11 2018
|
I don't see the update on elog/download/tar/.
Do you keep builds elsewhere, perhaps on get or sourceforge?
Stefan |
|
Re: Compile issues on Fedora withe current elog source, posted by Allen Tuttle on Wed Sep 12 15:25:07 2018
|
I would like to try the new code, I don't see it on elog/download/tar/.
Do you keep updates elsewhere like on get or sourceforge?
Stefan |
|
Re: Compile issues on Fedora withe current elog source, posted by Allen Tuttle on Wed Sep 12 15:57:31 2018
|
A quick test suggests I can't just replace sprintf with snprintf.
Original error:
src/elogd.c:4503:4: note: ‘sprintf’ output between 1 and 542 bytes into a destination of size 256 [-Wformat-overflow=] |
|
Re: Compile issues on Fedora withe current elog source, posted by Andreas Luedeke on Wed Sep 12 20:52:22 2018
|
Stefan said "code is commited": that means you can download the latest version from https://bitbucket.org/ritt/elog, as described in https://elog.psi.ch/elog/download.html,
and then compile it. Clear now?
Stefan |
|
Re: Compile issues on Fedora withe current elog source, posted by Allen Tuttle on Thu Sep 13 06:12:59 2018
|
Latest compile effort on Fedora 28 using the bitbucket source downloaded at 21:12 PDT:
make
gcc -O3 -funroll-loops -fomit-frame-pointer -W -Wall -Wno-deprecated-declarations -Imxml -DHAVE_SSL -c -o mxml.o mxml/mxml.c |
Logfile not registering entry numbers?, posted by Sergio Navarrete on Fri Oct 19 08:38:06 2018
|
I have configured a logbook with the logfile on, but when a user replies to an entry the line logged goes
Date Time [User@IP] {Logbook} NEW entry #0
How can I make the #0 be the real entry number for the reply? |
|
Re: Logfile not registering entry numbers?, posted by Andreas Luedeke on Fri Oct 19 13:08:30 2018
|
It looks like you've found a bug in ELOG. I've checked my elog.log and see that all NEW entry lines show "#0".
I've looked into the code: the message is written before the new entry is submitted, and only then the entry ID is defined.
For
new entries one would need to make the logging print line later - but that would blow up the code.
The message IDs are correct for saving |