Demo Discussion
Forum Config Examples Contributions Vulnerabilities
  Discussion forum about ELOG, Page 351 of 807  Not logged in ELOG logo
ID Date Icon Author Author Email Category OS ELOG Version Subjectdown
  66128   Tue Dec 23 12:22:51 2008 Reply Stefan Rittstefan.ritt@psi.chQuestionWindows2.7.5-2140Re: Problems with execute delete

 

Stefan Kanitz wrote:

 

Stefan Ritt wrote:

 

Stefan Kanitz wrote:

i have problems with execute delete (execute new works fine) in following config (the attribute 'Counter' will not evaluated). Can anybody help me?

 

Attribute substitution on 'execute delete' was not implemented, since I thought that the ELOG message ID would be used as the primary key in an SQL database, so one would not need the attributes. I added howver this functrionality in SVN revision 2159, so the next release will contain this fix. 

 

This sounds very good! Thank you very much (Could you please add this functionality for execute edit too? Please :-))

 

For 'execute edit' this worked from the beginning. 

  66129   Tue Dec 23 12:31:39 2008 Reply Stefan Kanitzskmainz@web.deQuestionWindows2.7.5-2140Re: Problems with execute delete

Stefan Ritt wrote:

 

Stefan Kanitz wrote:

 

Stefan Ritt wrote:

 

Stefan Kanitz wrote:

i have problems with execute delete (execute new works fine) in following config (the attribute 'Counter' will not evaluated). Can anybody help me?

 

Attribute substitution on 'execute delete' was not implemented, since I thought that the ELOG message ID would be used as the primary key in an SQL database, so one would not need the attributes. I added howver this functrionality in SVN revision 2159, so the next release will contain this fix. 

 

This sounds very good! Thank you very much (Could you please add this functionality for execute edit too? Please :-))

 

For 'execute edit' this worked from the beginning. 

 Thank you!

  65819   Wed Apr 9 23:03:53 2008 Reply Stefan Rittstefan.ritt@psi.chBug reportLinux2.7.3-2072Re: Problems with elog client

Yoshio Imai wrote:

Since our upgrade to elog 2.7.3, it is not possible any more to edit an existing elog entry using the elog client with -e <id> option. The only console output is "Error transmitting message". Submitting an entry via the client is not problem.

Running the server with -v option does not yield any output at the time of the edit attempt. Running the client with -v option also doesn't help, because whatever the other options, only the help page is printed out and nothing else done. Btw, there is now a conflict between -s for "use SSL" and -s for the "subdir" option.

 

Any ideas?

The problem had to do with the integration of native SSL into elog/elogd. I tried to fix it in revision 2099. Please give it a try. I have not tried yet the SSL mode, but I changed the -s for "subdir" to -d not to collide with SSL.

  65822   Thu Apr 10 14:21:13 2008 Reply Yoshio ImaiBug reportLinux2.7.3-2099Re: Problems with elog client
I have tried the new revision, recompiling both client and server.
Unfortunately, the overall situation has not changed. We can still create new entries using the elog client, but editing an existing entry still does not work. However, thanks to your patch, the -v option now works again for the client side:
> elog -s -h <elog server> -p 443 -l current -u <user> <password> -a <Attributes> "TEST ENTRY"
Message successfully transmitted, ID=1

> elog -s -h <elog server> -p 443 -l current -u <user> <password> -e 1 "EDIT THIS ENTRY" -v
Successfully connected to host <elog server>, port 443
Request sent to host:
GET /current/1?cmd=download HTTP/1.0
User-Agent: ELOG
Cookie: unm=<user>;upwd=<password>;


Response received:
ET
Error transmitting message
Again, there is no output on the server side even when running with -v in console mode. There is the expected output when an entry is successfully created using the elog client. One aspect about this output is strange, though: the "Return" statement
==== Return ================================
HTTP/1.1 302 Found
contains the wrong "Location" header. It is not
Location: https://elog-server.domainname/logbook/ID
but rather
Location: https://client-host.domainname/logbook/ID
where "client-host" is the hostname of the computer where the elog client was called. But, since the logbook entry is correctly created, this is perhaps not a problem at all.



Another thing: I have noticed that when using conditional preset texts, the "Preset Text =" statements have to be declared for each logbook of a logbook group; it is not sufficient to only declare it in the "[global]" section of the logbook group. This means that we only need to declare the conditions once, e.g.
Options Category = .... , Shift {1} , ....
{1} Options Subject = Shift start {a}, Shift end {b}, ....

but the preset text declarations
{1&a} Preset text = shift_checklist.txt
{1&b} Preset text = shift_summary.txt
have to be declared in every individual logbook section for the logbooks of this group. Is this expected behaviour, or a bug?

Cheers

Yoshio
  65823   Thu Apr 10 15:39:26 2008 Reply Stefan Rittstefan.ritt@psi.chBug reportLinux2.7.3-2099Re: Problems with elog client

Yoshio Imai wrote:
> elog -s -h elog -p 443 -l current -u <user> <password> -e 1 "EDIT THIS ENTRY" -v
Successfully connected to host elog, port 443


As I wrote in my previous message, the "SSL part has not been tested" (=means: does not work). I just fixed this today, so try revision 2100.
  65824   Thu Apr 10 15:56:12 2008 Agree Yoshio ImaiBug reportLinux2.7.3-2100Re: Problems with elog client

Stefan Ritt wrote:
As I wrote in my previous message, the "SSL part has not been tested" (=means: does not work).


Sorry for the misunderstanding. I was probably confused because the creation of new entries already worked, even with SSL.
Now, the editing also works.

Thank you very much!

Yoshio
  65831   Tue Apr 15 20:34:49 2008 Reply Stefan Rittstefan.ritt@psi.chBug reportLinux2.7.3-2099Re: Problems with elog client

Yoshio Imai wrote:
Another thing: I have noticed that when using conditional preset texts, the "Preset Text =" statements have to be declared for each logbook of a logbook group; it is not sufficient to only declare it in the "[global]" section of the logbook group. This means that we only need to declare the conditions once, e.g.
Options Category = .... , Shift {1} , ....
{1} Options Subject = Shift start {a}, Shift end {b}, ....

but the preset text declarations
{1&a} Preset text = shift_checklist.txt
{1&b} Preset text = shift_summary.txt
have to be declared in every individual logbook section for the logbooks of this group. Is this expected behaviour, or a bug?


I cannot reproduce your problem. I just tried following file:
[global]
port = 8000
password file = passwd
Options Category = Other , Shift {1} , Routine
Attributes = Author, Category, Subject
{1} Options Subject = Shift start {a}, Shift end {b}
{1&a} Preset text = shift_checklist.txt
{1&b} Preset text = shift_summary.txt

[demo1]
Comment = General linux tips & tricks

[demo2]
Comment = General linux tips & tricks

and it worked as expected. Can you try it? If it works for you as well, then try to morph this file towards your config file, and figure out when the unexpected behavior starts.
  65834   Thu Apr 17 19:11:36 2008 Reply Yoshio ImaiBug reportLinux2.7.3-2099Re: Problems with elog client
I have tried your sample config file, and this one also works for me.
The problem arises when using top groups. I have produced a minimal config file including a top group, of the form
[global]
logbook tabs = 1
SSL = 1
Theme = default

Top group testgroup = demo1, demo2
Show top groups = 1


[global testgroup]
Attributes      = Author, Subject, Type
Options Subject = ThisOne{1}, Shift info {2}, ThatOne{3}

{1} Options Type = OneType1, OneType2, OneType3
{2} Options Type = Shift start {a}, Shift end {b}, Info
{3} Options Type = TwoType1, TwoType2, TwoType3

Required Attributes = Author, Subject, Type

{2&a} Preset Text = shift_checklist.txt
{2&b} Preset Text = shift_summary.txt


[demo1]
Comment = DemoLog 1

[demo2]
Comment = DemoLog 2

Although the conditional attributes work for the logbooks demo1 and demo2, the preset text does not work. It works when the Preset Text= statement is (also) places in the individual logbook sections. Can you check if it behaves the same on your side?

Yoshio
ELOG V3.1.5-3fb85fa6