Demo Discussion
Forum Config Examples Contributions Vulnerabilities
  Discussion forum about ELOG, Page 193 of 801  Not logged in ELOG logo
 Find |  Login |  Help 
 Full | Summary | Threaded | Hide attachments  Show only new entries    6408 Entries 
Goto page Previous  1, 2, 3 ... 192, 193, 194 ... 799, 800, 801   Next  
ID Date Icon Authordown Author Email Category OS ELOG Version Subject
  1268   Thu Jul 14 17:29:42 2005 Reply Stefan Rittstefan.ritt@psi.chCommentAll Re: A new ELOG user wants to register on "127.0.0.1"

Emiliano Gabrielli wrote:
A new ELOG user wants to register on "127.0.0.1"


the scenario is:
- elog on localhost
- stunnel on the external interface

I dont want elog to listen on external interface, so.. why do not use the URL cfg attribute for this issue ?


You can specify the interface to liste on with the 
"-n <interface>"
parameter of elogd.
  1269   Thu Jul 14 17:34:12 2005 Reply Stefan Rittstefan.ritt@psi.chBug reportAll Re: short/long_name should point the same user

Emiliano Gabrielli wrote:
In my case I changed it by error, people inserted entries and now I restored the correct one .. so now I have to unlock the attribute and change every Author by hand as admin
Smile


Well, that teaches you not to do this error again Wink
  1272   Thu Jul 14 20:24:07 2005 Reply Stefan Rittstefan.ritt@psi.chBug reportAll Re: short/long_name should point the same user

Emiliano Gabrielli wrote:
uhm.. I think the confusion intrinsict in elog between long and short name is something to be solved ..
an attribute of type "userlist" fills always with the long_name .. but if I would to insert it as short ?

the users shown in the users admin dropdown menu is short .. why? .. may be I didnt understood the way this issue works .. Crying


The "short name" is the equivalent to the unix login name. Under /etc/passwd, you have a login (short) name and a "full" (long) name. The first may not contain blanks, must be unique, while the second is more like a "real" name. This concept has been adapted in elog. While many people use cryptic or abbreviated login names, it's still nice know the real name, like if you get an email notification from someone. The userlist fills with the long_name because people refer to other people in the logbook usually with the real name (sometimes they even don't know the people's login name). The admin dropdown menu uses the short names because you look at the user database more from an administration point of view. Like if you edit /etc/passwd, you first look at the login name, not the full one. Maybe what one could add is to make the full name in the admin page a dropdown list as well, so the admin can either select the short or the long name. Another item for the wishlist Crying
  1275   Sat Jul 16 12:30:33 2005 Reply Stefan Rittstefan.ritt@psi.chRequestAll Re: < img > in Display Attribute

Emiliano Gabrielli wrote:
IMHO the solution may be in allowing html code only in the cfg, not in the value of the field..
i.e., elog has to first allow html in the cfg attribute and then replace every $attribute with their values... this values as to be translated in html entities in order to solve the problem

in other wordsthe value of the attribute as to httl_encoded when displayed.. and the cfg line not

or I am loosing some point ?


I don't like that solution. Sometimes I want to write some HTML tag myself into an attribute (like "<b>") and have this interpreted, meaning having the following word in bold, and sometimes I want to talk about HTML tags, like to tell you to use <b> for bold, and have it not interpreted. The same with the ELCode tags.

So I added an escape character. If you write <b> or [b], then it gets interpreted. If you write \<b> or \[b] with the "\" in front of the tag, it does not get interpreted, thus the '\' works like an escape character. I edited your old entry with the <img> in the subject, and as you see it is now not interpreted.

The modification is in CVS and documented in the ELCode help.
  1281   Wed Jul 20 20:13:26 2005 Reply Stefan Rittstefan.ritt@psi.chQuestionWindows2.5.4-3Re: Cannot stat() config file

Stefan Rudat wrote:
Cannot stat() config file; No such file or directory.


I do not know what exactly is going on. This error occurs during runtime when elog tries to check if the config file (usually elogd.cfg) has changed. It checks if the file is there on startup, but apparently it is there, otherwise elog would not start. Then it checks about once per second if the file access time has changed. If so, it rereads the file. This way changes to the file become valid immediately, without restarting elog. The fact that elog has access to the file on startup, but not later, is somehow strange. Maybe you lock this file with an editor, or it is on a network drive which gets disconnected, or the permission changed after the startup. Something in that direction.
  1282   Wed Jul 20 20:54:01 2005 Reply Stefan Rittstefan.ritt@psi.chQuestionLinux2.5.9-2Re: Omitting some fields in e-mail
> E.g, if you have the following:
> Attributes = TA, Endret, UA, Pri, Status, Oppdragsgiver, EpostTil, EpostIKT, Tittel
> 
> Then you get the heading:
> Logbook             : 
> TA                  : 
> Endret              : 
> UA                  : 
> Pri                 : 
> Status              : 
> Oppdragsgiver       : 
> EpostTil            : 
> EpostIKT            : 
> Tittel              : 
> 
> (removed the attribute data),
> 
> But I would like to send out only:
> UA                  : 
> Pri                 : 
> Status              : 
> Oppdragsgiver       : 
> EpostTil            : 
> Tittel              : 
> 
> Any way to do this in the current version?
> 
> And perhaps change the order too... I'm asking for much perhaps?

I added the option "Email attributes = <list>" for that. So you can specify

Email attributes = UA, Pri, Status, Oppdragsgiver, EpostTil, Tittel

which also sets the order of the attributes. The cange is in CVS.
  1283   Wed Jul 20 21:03:29 2005 Reply Stefan Rittstefan.ritt@psi.chBug fixLinux2.5.9Re: elog utility for submission used wrong 'Host:' in POST header

Emiliano Gabrielli wrote:
This is not completally true IMHO .. better, it is, but it is not the only problem.

Elog seems to speak HTML/1.0, where "host:" is not implemented ... Since ELOG does not support Vhosts I think the right beaviour is to remove the "Host:" header at all ...


So I remove the Host: ... line at all, I hope this is ok with everybody.
  1285   Wed Jul 20 21:43:56 2005 Reply Stefan Rittstefan.ritt@psi.chBug reportAll Re: [code] should be a sort of <CDATA >

Emiliano Gabrielli wrote:
Using the [code] elocode should be intended also to preserve the tagged text from beeing parsed as html or elcode itself ..

this is an example:

Quote:
Note that, for security reasons, you should check the MD5 FINGERPRINT of the SSL certificate issued by the server agaist the following one:

MD5 Fingerprint = 23:A7:AD:33:3C:08:BE:2A:62:6E:85:DF:B8:00:23:40


Thank you


As you can see, your entry with the [code] section is now shown without interpretation. So everything between [code] and [/code] is not interpreted as ELCode tags. The modification is committed to CVS.
Goto page Previous  1, 2, 3 ... 192, 193, 194 ... 799, 800, 801   Next  
ELOG V3.1.5-3fb85fa6