Comiling elogd 2.0.4 under Solaris obviously fails with

gcc  -o elog elog.c -lsocket -lnsl
gcc  -o elogd elogd.c -lsocket -lnsl
elogd.c: In function `ss_file_find':
elogd.c:1372: `alphasort' undeclared (first use in this function)
elogd.c:1372: (Each undeclared identifier is reported only once
elogd.c:1372: for each function it appears in.)
make: *** [elogd] Error 1

because alphasort is an ucb'ism which does not exist under SYSV

... not very portable :-(

> Comiling elogd 2.0.4 under Solaris obviously fails with
> 
> gcc  -o elog elog.c -lsocket -lnsl
> gcc  -o elogd elogd.c -lsocket -lnsl
> elogd.c: In function `ss_file_find':
> elogd.c:1372: `alphasort' undeclared (first use in this function)
> elogd.c:1372: (Each undeclared identifier is reported only once
> elogd.c:1372: for each function it appears in.)
> make: *** [elogd] Error 1
> 
> because alphasort is an ucb'ism which does not exist under SYSV
> 
> ... not very portable :-(

I was not aware of this. Actually, I don't need 'alphasort' so I removed it 
from the source. The update is at

http://midas.psi.ch/cgi-bin/cvsweb/elog/elogd.c

Please give it a try.

> > Comiling elogd 2.0.4 under Solaris obviously fails with
> > 
> > gcc  -o elog elog.c -lsocket -lnsl
> > gcc  -o elogd elogd.c -lsocket -lnsl
> > elogd.c: In function `ss_file_find':
> > elogd.c:1372: `alphasort' undeclared (first use in this function)
> > elogd.c:1372: (Each undeclared identifier is reported only once
> > elogd.c:1372: for each function it appears in.)
> > make: *** [elogd] Error 1
> > 
> > because alphasort is an ucb'ism which does not exist under SYSV
> > 
> > ... not very portable :-(
> 
> I was not aware of this. Actually, I don't need 'alphasort' so I removed it 
> from the source. The update is at
> 
> http://midas.psi.ch/cgi-bin/cvsweb/elog/elogd.c
> 
> Please give it a try.

ok - i just continued with 2.0.5, but apparently it still has an UCB'ism
(scandir) that does not work under SYSV:

$ make
gcc -g -o elogd elogd.c -lsocket -lnsl
Undefined                       first referenced
 symbol                             in file
scandir                             /tmp/ccy0Z3iV.o
ld: fatal: Symbol referencing errors. No output written to elogd
collect2: ld returned 1 exit status
make: *** [elogd] Error 1
$ egrep scandir elogd.c
  n = scandir(path, &namelist, NULL, NULL);
$ 

> > > Comiling elogd 2.0.4 under Solaris obviously fails with
> > > 
> > > gcc  -o elog elog.c -lsocket -lnsl
> > > gcc  -o elogd elogd.c -lsocket -lnsl
> > > elogd.c: In function `ss_file_find':
> > > elogd.c:1372: `alphasort' undeclared (first use in this function)
> > > elogd.c:1372: (Each undeclared identifier is reported only once
> > > elogd.c:1372: for each function it appears in.)
> > > make: *** [elogd] Error 1
> > > 
> > > because alphasort is an ucb'ism which does not exist under SYSV
> > > 
> > > ... not very portable :-(
> > 
> > I was not aware of this. Actually, I don't need 'alphasort' so I removed 
it 
> > from the source. The update is at
> > 
> > http://midas.psi.ch/cgi-bin/cvsweb/elog/elogd.c
> > 
> > Please give it a try.
> 
> ok - i just continued with 2.0.5, but apparently it still has an UCB'ism
> (scandir) that does not work under SYSV:
> 
> $ make
> gcc -g -o elogd elogd.c -lsocket -lnsl
> Undefined                       first referenced
>  symbol                             in file
> scandir                             /tmp/ccy0Z3iV.o
> ld: fatal: Symbol referencing errors. No output written to elogd
> collect2: ld returned 1 exit status
> make: *** [elogd] Error 1
> $ egrep scandir elogd.c
>   n = scandir(path, &namelist, NULL, NULL);
> $ 
I am having the same experience with ELOG 2.0.5 and Solaris 8. Any work 
around woud be appreciated.

> > > > Comiling elogd 2.0.4 under Solaris obviously fails with
> > > > 
> > > > gcc  -o elog elog.c -lsocket -lnsl
> > > > gcc  -o elogd elogd.c -lsocket -lnsl
> > > > elogd.c: In function `ss_file_find':
> > > > elogd.c:1372: `alphasort' undeclared (first use in this function)
> > > > elogd.c:1372: (Each undeclared identifier is reported only once
> > > > elogd.c:1372: for each function it appears in.)
> > > > make: *** [elogd] Error 1
> > > > 
> > > > because alphasort is an ucb'ism which does not exist under SYSV
> > > > 
> > > > ... not very portable :-(
> > > 
> > > I was not aware of this. Actually, I don't need 'alphasort' so I 
removed 
> it 
> > > from the source. The update is at
> > > 
> > > http://midas.psi.ch/cgi-bin/cvsweb/elog/elogd.c
> > > 
> > > Please give it a try.
> > 
> > ok - i just continued with 2.0.5, but apparently it still has an UCB'ism
> > (scandir) that does not work under SYSV:
> > 
> > $ make
> > gcc -g -o elogd elogd.c -lsocket -lnsl
> > Undefined                       first referenced
> >  symbol                             in file
> > scandir                             /tmp/ccy0Z3iV.o
> > ld: fatal: Symbol referencing errors. No output written to elogd
> > collect2: ld returned 1 exit status
> > make: *** [elogd] Error 1
> > $ egrep scandir elogd.c
> >   n = scandir(path, &namelist, NULL, NULL);
> > $ 
> I am having the same experience with ELOG 2.0.5 and Solaris 8. Any work 
> around woud be appreciated.

I removed also scandir() in the current version (right now only in CVS). The 
official release of 2.1.0 is planned for next week.

Hi,

I added many features to ELOG recently, but before making a new release, I 
would like some beta testing. In addition to the current modes of 
operation, a logbook can be configured to require a user to be logged in to 
submit new messages. If not logged it, one can only read messages. To 
obtain a new account, one can do a self-registering. In this process, one 
can also enter the email address which is then filled in automatically on 
new logbook entries.

So please go to the forum logbook, create a new account for yourself and 
try things out. If there is any problem, please let me know.

Many thanks in advance, Stefan Ritt.

Hello,
   If you have the text box turned off so you only enter attributes, is it 
possible to have a couple of attributes that have small scroll through text 
boxs of a couple of lines rather than just one line?   Like not as big as 
the regular text box but something small to be able to post a couple of 
lines in and if it gets bigger then you scroll down.  For instance if you 
are posting a problem and a solution just have one small text box for the 
problem and one for the solution.

Thanks,
Mo

> Hello,
>    If you have the text box turned off so you only enter attributes, is it 
> possible to have a couple of attributes that have small scroll through text 
> boxs of a couple of lines rather than just one line?   Like not as big as 
> the regular text box but something small to be able to post a couple of 
> lines in and if it gets bigger then you scroll down.  For instance if you 
> are posting a problem and a solution just have one small text box for the 
> problem and one for the solution.

Having multi-line attributes would be pretty difficult since it would break 
the way the database works (one attribute at a line). So for now, you have to 
use the text box and the reply functionality to deal with problems and fixes 
(like for this entry).

Hello,
    In the newest version of eLog, I was having trouble saving the 
configurations once I change them through the web interface.  I can change 
them and save them fine if I do it in notepad; but that way you have to 
restart the elogd server.  I think this is a bug on the new version unless 
I am doing something wrong.  Also I was wondering if you were going to add 
any of the wish list items on version 2.0.6 (Conditional attributes seems 
to have the most votes and would be an extremely helpful feature!! :-).  
Thank You.

Mo.

>     In the newest version of eLog, I was having trouble saving the 
> configurations once I change them through the web interface.  

I found a problem in the elogd.c code, which might cause that proplem 
sometimes. The fix is available from

http://midas.psi.ch/cgi-bin/cvsweb/elog/elogd.c

If you need a windows executable, please let me know.

> I can change 
> them and save them fine if I do it in notepad; but that way you have to 
> restart the elogd server.

That should not be necessary, since elogd reads in the configuration file at 
each access, so it can be modified externally without problem (although the 
internal configuration should be the preferred way...)

> Also I was wondering if you were going to add 
> any of the wish list items on version 2.0.6 (Conditional attributes seems 
> to have the most votes and would be an extremely helpful feature!! :-).  

I will have some spare time towards the end of August, so I hope to work on 
these requests next month.

Just compiled elog-2.0.5 under Solaris 8 and 9  without any warnings/errors
but the elogd binary just core dumps:

# /opt/ELOG/bin/elogd -c /opt/ELOG/elogd.cfg 
Indexing logbook "demo"...
Bus Error (core dumped)

.. tracing elogd shows:

[ lines snipped ]
2964:   brk(0x005B8A68)                                 = 0
2964:   ioctl(1, TCGETA, 0xFFBEE6BC)                    = 0
Indexing logbook "demo"...
2964:   write(1, " I n d e x i n g   l o g".., 27)      = 27
2964:       Incurred fault #5, FLTACCESS  %pc = 0x000164C0
2964:         siginfo: SIGBUS BUS_ADRALN addr=0x2F757372
2964:       Received signal #10, SIGBUS [default]
2964:         siginfo: SIGBUS BUS_ADRALN addr=0x2F757372
2964:           *** process killed ***
#

... any glue ?

> Just compiled elog-2.0.5 under Solaris 8 and 9  without any warnings/errors
> but the elogd binary just core dumps:
> 
> # /opt/ELOG/bin/elogd -c /opt/ELOG/elogd.cfg 
> Indexing logbook "demo"...
> Bus Error (core dumped)
> 
> .. tracing elogd shows:
> 
> [ lines snipped ]
> 2964:   brk(0x005B8A68)                                 = 0
> 2964:   ioctl(1, TCGETA, 0xFFBEE6BC)                    = 0
> Indexing logbook "demo"...
> 2964:   write(1, " I n d e x i n g   l o g".., 27)      = 27
> 2964:       Incurred fault #5, FLTACCESS  %pc = 0x000164C0
> 2964:         siginfo: SIGBUS BUS_ADRALN addr=0x2F757372
> 2964:       Received signal #10, SIGBUS [default]
> 2964:         siginfo: SIGBUS BUS_ADRALN addr=0x2F757372
> 2964:           *** process killed ***
> #
> 
> ... any glue ?

Hmmm - stupid me used a broken patch  used to get rid of alphasort under 2.0.4
this entry and it's predecessor should be DELETED - sorry for the confusion :-(

Version 2.0.5 of elog has been released today. Among various bugfixes, the 
display speed for threaded lists has been improved significantly. Display 
all messages of this forum and press the reload button to see the 
difference.

Is there a way to change the From address used when sending notifications 
from ELOG?   I was hoping there might be another attribute that would go 
along with the smtp host section.

Thanks,
Eric

> Is there a way to change the From address used when sending notifications 
> from ELOG?   I was hoping there might be another attribute that would go 
> along with the smtp host section.
> 
> Thanks,
> Eric

Yes there is an option

Use Email from = xxxx

see documentation.

Hello,

I noticed a strange behaviour with Elog 2.0.4 (i386 RPM) in Linux (RH 7.2). 
I started Elog out-of-the-box with portnumber 888 ('sbin/elogd -p 888').

It runs just fine, but appears to listen to port 8080, the default! It 
looks like the -p option on the command-line is ignored.

Anyone seen similar problems? 

Joeri

> Anyone seen similar problems? 
Probably not if you read the config file, 'cause I didn't. Shame on me...

But what this shows (Stefan: correct me if I'm wrong) is that if you set 
the port number in the [global] section of the config file, the command-line
option '-p' is ignored. FYI...

Joeri

> > Anyone seen similar problems? 
> Probably not if you read the config file, 'cause I didn't. Shame on me...
> 
> But what this shows (Stefan: correct me if I'm wrong) is that if you set 
> the port number in the [global] section of the config file, the command-line
> option '-p' is ignored. FYI...
> 
> Joeri

I changed that behaviour, so from 2.0.5 on the command line port setting has 
precedence over the configuration file (as it should be).