ldapsh-devel Mailing List for LDAP shell
Status: Beta
Brought to you by:
rcorvalan
Menu
▾
▴
ldapsh-devel — Developpers of the tool
You can subscribe to this list here.
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(7) |
Aug
|
Sep
(1) |
Oct
(1) |
Nov
(1) |
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(10) |
Jun
|
Jul
(3) |
Aug
(6) |
Sep
(5) |
Oct
(1) |
Nov
|
Dec
|
| 2004 |
Jan
|
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
(2) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2005 |
Jan
|
Feb
|
Mar
|
Apr
(5) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2006 |
Jan
|
Feb
(4) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Ian A. T. <yo...@re...> - 2006-02-16 17:46:33
|
On Wed, Feb 15, 2006 at 10:21:04PM +0800, James Devenish wrote: > I'm no sure if I understand all your ideas, but they sound good. Are > you able to do the programming yourself? If so, please feel free to > submit them via the SourceForge tracker > <http://sourceforge.net/tracker/?atid=472733&group_id=54115&func=browse>. > I think we all still have our committers' access, so the patches can > be applied. Rafael Corvalan (project controller) may offer further > comment if he is still reading. I'm going to try and implement most of my suggestions; I've also been taking a look at: http://search.cpan.org/~loosifer/ldapsh-2.00/ I wanted to know what you guys thought about combining ideas and efforts with this other project. There are features and ideas from both that are attractive. For example, the command completion with 0.9.4 is very handy while the 'ls -l' on 2.00 is very informative: default:ou=People,dc=madstop,dc=com> ls -l Creator Created Modifier Modified Sub Name ------------------------------------------------------------------------ cn=Manager 05/06/2004 21:47:58 - 05/06/2004 21:47:58 - uid=luke uid=luke 07/21/2004 11:36:29 - 07/21/2004 11:40:57 - uid=hostmaster uid=luke 07/26/2004 21:57:39 - 07/29/2004 16:03:19 - uid=testing uid=luke 07/26/2004 23:59:14 - 07/26/2004 23:59:14 - uid=admin Thoughts? -- ian |
|
From: James D. <j-d...@us...> - 2006-02-15 14:21:20
|
Hi Ian, I'm no sure if I understand all your ideas, but they sound good. Are you able to do the programming yourself? If so, please feel free to submit them via the SourceForge tracker <http://sourceforge.net/tracker/?atid=3D472733&group_id=3D54115&func=3Dbrow= se>. I think we all still have our committers' access, so the patches can be applied. Rafael Corvalan (project controller) may offer further comment if he is still reading. I also think I still have patches and features to commit, simply haven't done it. |
|
From: <pe...@kl...> - 2006-02-15 13:16:56
|
"Ian A. Tegebo" <yo...@re...> writes: > Hello, > > I started using ldapsh recently and noticed that there are some little > issues I'd like to resolve and features I'd like to add: > > # Use and create tmp dir for ldif files > # Create function for selecting in-progress ldif files > # Allow suspension of ldapsh and editor > # Create function for querying schema > > Since this list has been quiet for awhile, I was wondering if anyone was > still listening. Listen - yes. Work on ldapsh - unfortunately no. I've changed gigs since I got involved with ldapsh, and at my current place of emplyment, we don't use LDAP at all. Mostly, we don't even use Perl. ...Peder... -- It's not called hacking or trial and error! It's called rapid prototyping and extreme programming. - Kristoffer Gleditsch |
|
From: Ian A. T. <yo...@re...> - 2006-02-15 12:46:13
|
Hello, I started using ldapsh recently and noticed that there are some little issues I'd like to resolve and features I'd like to add: # Use and create tmp dir for ldif files # Create function for selecting in-progress ldif files # Allow suspension of ldapsh and editor # Create function for querying schema Since this list has been quiet for awhile, I was wondering if anyone was still listening. -- Ian Tegebo |
|
From: Ski K. <kac...@co...> - 2005-04-07 21:31:15
|
Ok,
Found one problem, but still not getting apply to work. I have to put
'' around the ls uid=ski* like 'ls uid=ski*'. Now I get:
[LDAP @ ] > laptop
Password > ********
[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls 'uid=ski*'
uid=ski
uid=ski1
-----
2 entries found.
[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply {$_->attributes}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > nbentries
2
[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply{$_->get_value('cn')}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > print
apply{$_->get_value('cn')}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > printf
apply{$_->get_value('cn')}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > printf
apply{$_->get_value('uid')}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply{$_->get_value('uid')}
[LDAP @ localhost] ou=people,dc=nsd,dc=org > $a =
apply{$_->get_value('uid')}
Thanks for your help.
cheers,
ski
Rco Mli wrote:
> Hello Ski,
>
> I suspect a conflict with psh. Have you psh? If so, can you try running ldapsh with psh deactivated (set env variable LDAPSH_PARSER to ldapsh)?
>
> If this works, my suggestion depend on whether you WANT use Psh or not. I don't. By the way, let me know the result of this and then we can go further.
>
> Cheers,
>
> Rafael
>
> -----Original Message-----
>
>>De: Ski Kacoroski <kac...@co...>
>>A: lda...@li...
>>Sujet: [LDAPsh-devel] Cannot get apply to work
>>Date: Mon, 04 Apr 2005 17:01:44 -0700
>>
>>Attached is a listing of what happens.
>>
>>ski@apple:~$ ldapsh
>>[LDAP @ ] > laptop
>>Password > ********
>>[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls uid=ski*
>>uid=ski
>>uid=ski1
>>-----
>>2 entries found.
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > nbentries
>>2
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply {$_->attributes}
>>Can't use string ("{->attributes}") as a subroutine ref while "strict
>>refs" in use at /
>>LDAP @ localhost] ou=people,dc=nsd,dc=org > version
>>LDAP Shell (ldapsh) by Rafael Corvalan, Peder O. Klingenberg
>>CVS File ID: $Id: ldapsh,v 1.40 2003/12/23 10:46:45 rcorvalan
>>Shell release: 0.9.4pre1
>>http://ldapsh.sourceforge.net/
>>
>>
>>
>>Thanks for your help.
>>
>>cheers,
>>
>>ski
>>
>>
>>-------------------------------------------------------
>>SF email is sponsored by - The IT Product Guide
>>Read honest & candid reviews on hundreds of IT Products from real users.
>>Discover which products truly live up to the hype. Start reading now.
>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>_______________________________________________
>>LDAPsh-devel mailing list
>>LDA...@li...
>>https://lists.sourceforge.net/lists/listinfo/ldapsh-devel
>>
>>
>
>
> -------------------------------------------------------------
> NetCourrier, votre bureau virtuel sur Internet : Mail, Agenda, Clubs, Toolbar...
> Web/Wap : www.netcourrier.com
> Téléphone/Fax : 08 92 69 00 21 (0,34 € TTC/min)
> Minitel: 3615 NETCOURRIER (0,16 € TTC/min)
>
>
>
>
--
"When we try to pick out anything by itself, we find it
connected to the entire universe" John Muir
Chris "Ski" Kacoroski, kac...@co..., 425-681-0092
|
|
From: Ski K. <kac...@co...> - 2005-04-07 21:04:08
|
Rafael,
That doesn't help. Now I cannot even do the ls uid=ski* (this used to
work before). BTW, I am running debian sarge on a 2.6 kernel. I did
have psh installed.
ski@apple:~$ ldapsh
[LDAP @ ] > laptop
Password > ********
[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls uid=ski*
uid=ski
uid=ski1
-----
2 entries found.
[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply {$_->attributes}
Can't use string ("{->attributes}") as a subroutine ref while "strict
refs" in use at /usr/local/bin/ldapsh line 2538, <DATA> line 283.
[LDAP @ localhost] ou=people,dc=nsd,dc=org > quit
ski@apple:~$ export LDAPSH_PARSER=ldapsh
ski@apple:~$ ldapsh
[LDAP @ ] > laptop
Password > ********
[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls uid=ski*
syntax error at (eval 25) line 2, at EOF
If I remove psh, I get the same error. Thanks for your time in helping
me to debug this.
cheers,
ski
Rco Mli wrote:
> Hello Ski,
>
> I suspect a conflict with psh. Have you psh? If so, can you try running ldapsh with psh deactivated (set env variable LDAPSH_PARSER to ldapsh)?
>
> If this works, my suggestion depend on whether you WANT use Psh or not. I don't. By the way, let me know the result of this and then we can go further.
>
> Cheers,
>
> Rafael
>
> -----Original Message-----
>
>>De: Ski Kacoroski <kac...@co...>
>>A: lda...@li...
>>Sujet: [LDAPsh-devel] Cannot get apply to work
>>Date: Mon, 04 Apr 2005 17:01:44 -0700
>>
>>Attached is a listing of what happens.
>>
>>ski@apple:~$ ldapsh
>>[LDAP @ ] > laptop
>>Password > ********
>>[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls uid=ski*
>>uid=ski
>>uid=ski1
>>-----
>>2 entries found.
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > nbentries
>>2
>>[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply {$_->attributes}
>>Can't use string ("{->attributes}") as a subroutine ref while "strict
>>refs" in use at /
>>LDAP @ localhost] ou=people,dc=nsd,dc=org > version
>>LDAP Shell (ldapsh) by Rafael Corvalan, Peder O. Klingenberg
>>CVS File ID: $Id: ldapsh,v 1.40 2003/12/23 10:46:45 rcorvalan
>>Shell release: 0.9.4pre1
>>http://ldapsh.sourceforge.net/
>>
>>
>>
>>Thanks for your help.
>>
>>cheers,
>>
>>ski
>>
>>
>>-------------------------------------------------------
>>SF email is sponsored by - The IT Product Guide
>>Read honest & candid reviews on hundreds of IT Products from real users.
>>Discover which products truly live up to the hype. Start reading now.
>>http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
>>_______________________________________________
>>LDAPsh-devel mailing list
>>LDA...@li...
>>https://lists.sourceforge.net/lists/listinfo/ldapsh-devel
>>
>>
>
>
> -------------------------------------------------------------
> NetCourrier, votre bureau virtuel sur Internet : Mail, Agenda, Clubs, Toolbar...
> Web/Wap : www.netcourrier.com
> Téléphone/Fax : 08 92 69 00 21 (0,34 € TTC/min)
> Minitel: 3615 NETCOURRIER (0,16 € TTC/min)
>
>
>
>
--
"When we try to pick out anything by itself, we find it
connected to the entire universe" John Muir
Chris "Ski" Kacoroski, kac...@co..., 425-681-0092
|
|
From: Rco M. <rc...@ma...> - 2005-04-06 01:16:00
|
Hello Ski,
I suspect a conflict with psh. Have you psh? If so, can you try running =
ldapsh with psh deactivated (set env variable LDAPSH_PARSER to ldapsh)?
If this works, my suggestion depend on whether you WANT use Psh or not. =
I don't. By the way, let me know the result of this and then we can go =
further.
Cheers,
Rafael
-----Original Message-----
>De: Ski Kacoroski <kac...@co...>
>A: lda...@li...
>Sujet: [LDAPsh-devel] Cannot get apply to work
>Date: Mon, 04 Apr 2005 17:01:44 -0700
>
>Attached is a listing of what happens.
>
>ski@apple:~$ ldapsh
>[LDAP @ ] > laptop
>Password > ********
>[LDAP @ localhost] dc=3Dnsd,dc=3Dorg > cd ou=3Dpeople
>[LDAP @ localhost] ou=3Dpeople,dc=3Dnsd,dc=3Dorg > ls uid=3Dski*
>uid=3Dski
>uid=3Dski1
>-----
>2 entries found.
>[LDAP @ localhost] ou=3Dpeople,dc=3Dnsd,dc=3Dorg > nbentries
>2
>[LDAP @ localhost] ou=3Dpeople,dc=3Dnsd,dc=3Dorg > apply =
{$_->attributes}
>Can't use string ("{->attributes}") as a subroutine ref while "strict=20
>refs" in use at /
>LDAP @ localhost] ou=3Dpeople,dc=3Dnsd,dc=3Dorg > version
>LDAP Shell (ldapsh) by Rafael Corvalan, Peder O. Klingenberg
>CVS File ID: $Id: ldapsh,v 1.40 2003/12/23 10:46:45 rcorvalan
>Shell release: 0.9.4pre1
>http://ldapsh.sourceforge.net/
>
>
>
>Thanks for your help.
>
>cheers,
>
>ski
>
>
>-------------------------------------------------------
>SF email is sponsored by - The IT Product Guide
>Read honest & candid reviews on hundreds of IT Products from real =
users.
>Discover which products truly live up to the hype. Start reading now.
>http://ads.osdn.com/?ad_id=3D6595&alloc_id=3D14396&op=3Dclick
>_______________________________________________
>LDAPsh-devel mailing list
>LDA...@li...
>https://lists.sourceforge.net/lists/listinfo/ldapsh-devel
>
>
-------------------------------------------------------------
NetCourrier, votre bureau virtuel sur Internet : Mail, Agenda, Clubs, =
Toolbar...
Web/Wap : www.netcourrier.com
T=C3=A9l=C3=A9phone/Fax : 08 92 69 00 21 (0,34 =E2=82=AC TTC/min)
Minitel: 3615 NETCOURRIER (0,16 =E2=82=AC TTC/min)
|
|
From: Ski K. <kac...@co...> - 2005-04-05 00:01:53
|
Attached is a listing of what happens.
ski@apple:~$ ldapsh
[LDAP @ ] > laptop
Password > ********
[LDAP @ localhost] dc=nsd,dc=org > cd ou=people
[LDAP @ localhost] ou=people,dc=nsd,dc=org > ls uid=ski*
uid=ski
uid=ski1
-----
2 entries found.
[LDAP @ localhost] ou=people,dc=nsd,dc=org > nbentries
2
[LDAP @ localhost] ou=people,dc=nsd,dc=org > apply {$_->attributes}
Can't use string ("{->attributes}") as a subroutine ref while "strict
refs" in use at /
LDAP @ localhost] ou=people,dc=nsd,dc=org > version
LDAP Shell (ldapsh) by Rafael Corvalan, Peder O. Klingenberg
CVS File ID: $Id: ldapsh,v 1.40 2003/12/23 10:46:45 rcorvalan
Shell release: 0.9.4pre1
http://ldapsh.sourceforge.net/
Thanks for your help.
cheers,
ski
|
|
From: Ski K. <kac...@co...> - 2005-04-04 22:51:56
|
Hi, Really like the program, but needed to have an interactive option on deleting objects. The attached patch does it for me and may be useful for other folks. Not sure if I created the patch correctly, but to apply it I: patch -p1 ldapsh ldapsh_rmi-patch.diff cheers, ski |
|
From: James D. <j-d...@us...> - 2004-06-05 08:34:45
|
In message <uju...@nf...> on Thu, Jun 03, 2004 at 05:47:14PM +0200, Peder O. Klingenberg wrote: > The functionality is of course still available through the aliases mv > and create. I personally prefer to use mkdir interactively, because it "feels more like a shell". However, I don't object to your changes. |
|
From: <po...@li...> - 2004-06-03 15:47:21
|
I've just posted (not committed!) a patch to SF removing these two function names from ldapsh. The functionality is of course still available through the aliases mv and create. <http://sourceforge.net/tracker/index.php?func=detail&aid=965818&group_id=54115&atid=472733> The problem is that these two function names collide with perl builtin functions, IMO gratuitously so. As explained in the patch tracker, a client application built on top of ldapsh makes use of these functions assuming they will do what they usually do in perl. Again in my opinion, we should not break such assumptions without good reason. As far as I can tell, with this patch applied, the only builtin function we shadow is exit(), and I think that one is justifiable. Comments? Should I apply? -- Peder O. Klingenberg | Seniorkonsulent/Senior Consultant Linpro AS | URL: http://www.linpro.no/ Wdm. Thranes Gt 98b | Tlf: +47 22 87 11 80 NO-0175 OSLO | Fax: +47 22 87 11 81 |
|
From: James D. <j-d...@us...> - 2004-02-03 01:18:54
|
Hi,
In message <200...@ka...>
on Mon, Feb 02, 2004 at 05:29:02PM +0100, Sebastien Gross wrote:
> - when a C-t is done, the command line should be cleared the
> begin of the command redisplayed
Thanks; fixed in CVS.
> I was wondering if some feature/enhencement could be included:
> - vi with option to fetch and add MUST/MAY fields in the edited
> file
I added MUST/MAY for /new/ entries, though it is difficult to add these
when editing /existing/ entries. The reason is that the output file is
created by a function that is part of the Net::LDAP::LDIF Perl module.
We would have to filter the output in order to intersperse MUST and MAY
annotations. On the up side, this might also give us the opportunity to
sort attribute names alphabetically. Do you think it would be better to
put unused MAY attributes at the bottom of the LDIF file or better to
put them in their alphabetical location?
> - when I create a new user I got the following message:
>
> new cn=zero top person inetOrgPerson organizationalPerson posixAccount
> LDAP Error adding entry 'cn=zero, ou=users,dc=chezwam,dc=org'. Code:21.
> Message:description: value #0 invalid per syntax
> LDIF stored in LDAPShell_vi_PnfAMwE9zZ.ldif
>
> this should be because some fieds are left empty (or with only
> a space).
I'm not sure if Net::LDAP gives us the facility to check each attribute
value for its correct syntax. And if a blank line is /valid/ syntax, how
do we know whether we should remove that line? I don't know enough about
LDIF/LDAP to decide on the correct behaviour. Perhaps we should add a
warning comment at the top of each new entry? As a consolation prize,
perhaps we could add a -p ("prune") option to the create/new commands so
that users can opt to have empty lines removed.
|
|
From: Sebastien G. <wa...@ch...> - 2004-02-02 15:30:28
|
Hello Rafael,
I recently found your tool LDAPsh. I was searching for a ldap browser in
text mode.
I was wondering if some feature/enhencement could be included:
- vi with option to fetch and add MUST/MAY fields in the edited
file
- when a C-t is done, the command line should be cleared the
begin of the command redisplayed
- when I create a new user I got the following message:
new cn=zero top person inetOrgPerson organizationalPerson posixAccount
LDAP Error adding entry 'cn=zero, ou=users,dc=chezwam,dc=org'. Code:21.
Message:description: value #0 invalid per syntax
LDIF stored in LDAPShell_vi_PnfAMwE9zZ.ldif
this should be because some fieds are left empty (or with only
a space).
Please can you check this?
Thanks in advance.
--
Sebastien Gross
|
|
From: James D. <j-d...@us...> - 2003-10-20 12:40:16
|
Hi, I ran into a problem using `perl -w` for ldapsh in combination with Net::LDAP v0.29. It seems that Net::LDAP::Constant has changed and so something like this... $code != Net::LDAP::Constant::LDAP_PARTIAL_RESULTS ...gives a bareword error. What is the proper way to use Net::LDAP::Constant 0.29? One thing that I tried was this: $code != Net::LDAP::Constant::LDAP_PARTIAL_RESULTS() (Since they're supposed to be subroutines.) What would you do? |
|
From: Rafael C. <Raf...@li...> - 2003-09-23 15:10:16
|
Hello,
You're right. The line:
if (s/;\s*([|>][\s\w\d_\/.-]+)//) {
is historical. Prior to have the ";|" syntax, we got some problems with
someting like:
if ( $a | $b ) {.....
to avoid that, the regexp was very restrictive (you can only have
[\s\w\d_\/.-]+ after the "greater that" or the "pipe"), so the parsing
would be better. But this regexp is broken too.
Since the problem has been partially solved by the "|;" and ";>" syntax,
we can use :
if (s/;\s*([|>].+)//) {
I will replace this regexp and make a new release and publish it.
Bye
Rafael=20
-----Original Message-----
From: James Devenish [mailto:j-d...@us...]=20
Sent: dimanche, 21. septembre 2003 04:11
To: lda...@li...
Subject: Re: [LDAPsh-devel] Enhanced redirection
<...>
> cat 'cn=3Dtest' | grep -i last
>=20
> will give an error.
If you use ;| instead of | then it will work. I.e.:
cat 'cn=3Dtest' ;| grep -i last
This is not ideal (as you have said) but it is what we have included in
our documented example.
PS. Actually, it will not work: grep will be invoked without the "-i
last" options. This is because of the following line:
if (s/;\s*([|>][\s\w\d_\/.-]+)//) {
I am not sure of the purpose of that line, so you might like to make a
decision about whether it should look more like this:
if (s/;\s*([|>].+)//) {
(The above line is what I, personally, use).
|
|
From: James D. <j-d...@us...> - 2003-09-21 02:10:52
|
Hi,
In message <61F...@he...>
on Fri, Sep 19, 2003 at 03:41:07PM +0200, Rafael Corvalan wrote:
> Redirection is still difficult to understand, I think, for someone that
> does not know internals... Mainly due to filehandles lifetime.
I believe that if the user uses the "temporary redirect" syntax...
ldapsh> ls ;| wc -c
----
19 entries found.
319
ldapsh> ls ;| wc -c
----
19 entries found.
319
...then they will get what they want (i.e. wc will print its answer each
time, since the handles are only open for the duration of ldapsh's ls).
If the user uses the "permanent redirect" syntax:
ldapsh> redir "| sort -u"
ldapsh> ls
----
19 entries found.
ldapsh> ls
----
19 entries found.
ldapsh> noredir
...then it is a *feature* that unique, sorted results would be printed
only after noredir.
However, you are quite correct that this can be confusing for new users
(esp. if they then want to invoke help). Also, it is not very convenient
if I want `sort` filehandles to be opened/closed with every ls. Perhaps
we could introduce an abbreviated redirection syntax ;| (with no
trailing characters) to say "repeat the last redirection". Thus:
ldapsh> ls ;| wc -c
----
19 entries found.
319
ldapsh> ls ;|
----
19 entries found.
319
> By the way, redir() is quite broken,
I notice that "ls;|echo" causes an exit code 141 on one of my hosts. I
don't know how to debug perl.
> because command line parsing is
> very simple.
It would be interesting to have a "proper" command-line parser. I
realise the convenience and magic of using the Perl interpreter, but I
think it is a bit limiting for us, and can be a barrier for new users
(well, it was for me). I am not very good at developing parsers, though.
:(
> cat 'cn=test' | grep -i last
>
> will give an error.
If you use ;| instead of | then it will work. I.e.:
cat 'cn=test' ;| grep -i last
This is not ideal (as you have said) but it is what we have included in
our documented example.
PS. Actually, it will not work: grep will be invoked without the "-i
last" options. This is because of the following line:
if (s/;\s*([|>][\s\w\d_\/.-]+)//) {
I am not sure of the purpose of that line, so you might like to make a
decision about whether it should look more like this:
if (s/;\s*([|>].+)//) {
(The above line is what I, personally, use).
|
|
From: Rafael C. <Raf...@li...> - 2003-09-19 19:45:43
|
Hi,
After some of code reviewing, testing and resolving conflicts ;-), I
applied the patch proposed by James for enhanced redirection.
Redirection is still difficult to understand, I think, for someone that
does not know internals... Mainly due to filehandles lifetime.
When you use "redir('| something')", you have to understand that if
"something" is a command "flow based", it works as everyone expects. If
"something" must wait that the input filehandle is closed before
printing anything, some people may not understand. Example:
redir '| grep last'
will work as everyone expects
redir '| wc'
will not output anything until a noredir() is typed, because "wc" is a
command that needs to know the whole input to print anything. And since
the filehandle is still open, the "wc" command is still "alive".
We have 2 options:
1) That's what we expect
2) We expected an output at each command, in which case we must open the
filhandle '| wc' at each command.
I don't have the answer... For me both are good.
By the way, redir() is quite broken, because command line parsing is
very simple. For example:
cat 'cn=3Dtest' | grep -i last
will give an error. This is not 100% a problem of redir. It's also a
command line parsing problem. If someone knows a package that handle
this issue better than we do, please let me know...
Cheers,
Rafael
|
|
From: Rafael C. <Raf...@li...> - 2003-09-17 15:20:45
|
Hi James,=20 Regarding my comment on the code: <R> BUG (TODO): If you add an attribute to an entry (so, <R> if you add a value to an attribute that hadn't <R> previously a value), it will not be detected. <R> We should do a two-way check. You replied: <J> I don't quite understand this comment. Do you have <J> any test cases that exhibit incorrect behaviour? <J> The command was designed to detect new attributes, <J> and new values for existing attributes, without <J> using a two-way check. The task that is not possible <J> is the addition of new entries (i.e. new DNs) or <J> removal of existing entries. You're absolutely right. My test case was wrong. It works great, on Linux and Windows ! I will remove the comment. Rafael |
|
From: James D. <j-d...@us...> - 2003-09-04 01:02:32
|
In message <61F...@he...> on Sun, Aug 31, 2003 at 02:04:15AM +0200, Rafael Corvalan wrote: > 1) Before commiting the patch, I did 2 tests : a) adding one new > attribute b) adding one new value to an existing attribute. Maybe my code is not portable? (i.e. it works fine for me even if it doesn't work for you.) > 3) You're patch to enhanced redirection seems quite good. I was > reviewing it on friday to check that there was no side effects. I will > probably commit these changes next week also. > This is how it works until now: > > a. "ls" => output to STDOUT > b. "ls | less" => redirects "ls" result into "less" Oh, "| less" this has never worked for me (I have tried with Solaris, OpenBSD and Linux using Perl 5.8.0, 5.6.1 and 5.6.1, respectively). > d. "redir" => any previous redirection is cleared This has never worked for me. I really like the behaviour you described, though. |
|
From: Rafael C. <Raf...@li...> - 2003-08-31 00:04:42
|
Hi James,
1) Before commiting the patch, I did 2 tests : a) adding one new =
attribute b) adding one new value to an existing attribute.
Maybe I did something wrong, but I noticed that the 1st test said =
that the entry didn't changed.
If you implemented the behaviour, this is probably a worng test I =
made and I will remove the comment.
I will do the test on my box at office next week!
2) You're right, the comment should be changed to reflect the new logic
3) You're patch to enhanced redirection seems quite good. I was =
reviewing it on friday to check that there was no side effects. I will =
probably commit these changes next week also.
This is how it works until now:
a. "ls" =3D> output to STDOUT
b. "ls | less" =3D> redirects "ls" result into "less"
c. "redir '>/tmp/file.out'; ls; csv $ENTRIES" =3D> "ls" and "csv" =
output is redirected to "/tmp/file.out" (also any other output after =
those commands until a "redir" without arguments is executed.
d. "redir" =3D> any previous redirection is cleared
e. "redir '| sed -n "/John/p"' =3D> Commands output will be redirectet =
to the "sed" command until a "redir" without arguments is executed.
Promised I take a look at it next week.
Rafael
-----Original Message-----
From: James Devenish [mailto:j-d...@us...]=20
Sent: samedi, 30 ao=FBt 2003 06:24
To: lda...@li...
Subject: [LDAPsh-devel] Re: [LDAPsh-cvs] ldapsh ldapsh,1.20.8.1,1.20.8.2
Hi,
In message <E19...@sc...>
on Thu, Aug 28, 2003 at 03:35:27PM -0700, =
rco...@us... wrote:
> *** ldapsh 27 Jul 2003 07:47:25 -0000 1.20.8.1
> --- ldapsh 28 Aug 2003 22:35:25 -0000 1.20.8.2
> ***************
> *** 1591,1594 ****
> --- 1591,1596 ----
> the search results. Any changes can be viewed using =
L<changes|changes> and will need to
> be committed using L<commit|commit>.
> + BUG (TODO): If you add an attribute to an entry (so, if you add a=20
> + value to an attribute that hadn't previously a value), it will not=20
> + be detected. We should do a two-way check.
I don't quite understand this comment. Do you have any test cases that =
exhibit incorrect behaviour? The command was designed to detect new =
attributes, and new values for existing attributes, without using a =
two-way check. The task that is not possible is the addition of new =
entries (i.e. new DNs) or removal of existing entries.
In fact...does ldapsh provide *any* commands to add or delete entries =
(other than "cp")? Currently, the "add" command is for attributes (which =
is needed more often than the ability to add entries, I think). Perhaps =
"new" (synonym: "create") and "remove" (synonym: "rm") could be the =
command names for adding and deleting entries? Plus "rename" (synonyms: =
"mv" and "move") for modifying DNs?
> ! unless ($ldif->error() or $examined >=3D scalar(@$entries)) {
> print STDERR "Warning: some entries were absent from the edited =
file.\n\n";
> }
I think the message needs to be changed to reflect the new logic. =
Perhaps "Warning: some edited entries were added, removed, or =
unreadable.\n\n"?
PS. Also, I would be glad if someone could comment on my "enhanced =
redirection" patch -- have I improved redirection or have I =
misunderstood the pre-existing redirection features? I would like to be =
able to do
help all;|less
within the shell.
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf _______________________________________________
LDAPsh-devel mailing list
LDA...@li...
https://lists.sourceforge.net/lists/listinfo/ldapsh-devel
|
|
From: James D. <j-d...@us...> - 2003-08-30 04:24:26
|
Hi,
In message <E19...@sc...>
on Thu, Aug 28, 2003 at 03:35:27PM -0700, rco...@us... wrote:
> *** ldapsh 27 Jul 2003 07:47:25 -0000 1.20.8.1
> --- ldapsh 28 Aug 2003 22:35:25 -0000 1.20.8.2
> ***************
> *** 1591,1594 ****
> --- 1591,1596 ----
> the search results. Any changes can be viewed using L<changes|changes> and will need to
> be committed using L<commit|commit>.
> + BUG (TODO): If you add an attribute to an entry (so, if you add a value to an attribute
> + that hadn't previously a value), it will not be detected. We should do a two-way check.
I don't quite understand this comment. Do you have any test cases that
exhibit incorrect behaviour? The command was designed to detect new
attributes, and new values for existing attributes, without using a
two-way check. The task that is not possible is the addition of new
entries (i.e. new DNs) or removal of existing entries.
In fact...does ldapsh provide *any* commands to add or delete entries
(other than "cp")? Currently, the "add" command is for attributes (which
is needed more often than the ability to add entries, I think). Perhaps
"new" (synonym: "create") and "remove" (synonym: "rm") could be the
command names for adding and deleting entries? Plus "rename" (synonyms:
"mv" and "move") for modifying DNs?
> ! unless ($ldif->error() or $examined >= scalar(@$entries)) {
> print STDERR "Warning: some entries were absent from the edited file.\n\n";
> }
I think the message needs to be changed to reflect the new logic.
Perhaps "Warning: some edited entries were added, removed, or
unreadable.\n\n"?
PS. Also, I would be glad if someone could comment on my "enhanced
redirection" patch -- have I improved redirection or have I
misunderstood the pre-existing redirection features? I would like to be
able to do
help all;|less
within the shell.
|
|
From: James D. <j-d...@us...> - 2003-08-29 00:22:58
|
In message <61F...@he...> on Fri, Aug 29, 2003 at 01:26:23AM +0200, Rafael Corvalan wrote: > So I've done a cvs up -A ldapsh. This generated a lot of conflicts. I > copied the ldapsh I had just prior to run the "cvs up -A" into ldapsh > and commited. This created 1.21. You are probably right! |
|
From: James D. <j-d...@us...> - 2003-08-29 00:21:23
|
In message <61F...@he...> on Fri, Aug 29, 2003 at 12:38:40AM +0200, Rafael Corvalan wrote: > The module Unicode::MapUTF8 seems to be problematic: [...] > 2) I upgraded my Linux box and have now Perl 5.8. I get a lot of errors > installing the module. Interesting...I am using Perl 5.8 with Solaris and Unicode::MapUTF8 is successfully installed within site_perl. |
|
From: Rafael C. <Raf...@li...> - 2003-08-28 23:26:33
|
Hi,
I merged James patches into HEAD.
I'm not a CVS specialist (I'm not a specialist of anything at all ;-))
I hope it's all OK.
I followed the instructions of James:
cvs up -kk -r 1.17 ldapsh && \
cvs up -kk -j 1.17 -j 1.17.2.1 ldapsh && \
cvs up -kk -j 1.17 -j 1.17.4.1 ldapsh && \
cvs up -kk -j 1.17 -j 1.17.6.1 ldapsh && \
cvs up -kkv -j 1.17 -j 1.20 ldapsh && \
cvs up -kk -j 1.20 -j 1.20.2.1 ldapsh && \
cvs up -kk -j 1.20 -j 1.20.4.1 ldapsh && \
cvs up -kk -j 1.20 -j 1.20.6.1 ldapsh
# now, edit the conflict by merely removing the conflict
# markers (and leaving all the new code in there!)
cvs up -kk -j 1.20 -j 1.20.8.1 ldapsh
At the end, I was not able to commit. I got:
cvs server: sticky tag `1.17' for file `ldapsh' is not a branch
So I've done a cvs up -A ldapsh. This generated a lot of conflicts. I
copied the ldapsh I had just prior to run the "cvs up -A" into ldapsh
and commited.
This created 1.21.
If this was not the right way please let me know and I'll try to do the
right way!!
If this it's Ok for you and it runs ok for a few weeks, I will build a
complete release (tarball with the shell, the documentation, sample
.ldapshrc etc...).
Enjoy,
Rafael
|
|
From: Rafael C. <Raf...@li...> - 2003-08-28 22:38:46
|
Hi all, I checked the branches created by James. Excellent work. I really LOVE the new "vi" feature that allows changing data !!! I just fixed a very little bug (release 1.20.8.2). I have a question. The module Unicode::MapUTF8 seems to be problematic: 1) As far as I know, it is not available on Windows 2) I upgraded my Linux box and have now Perl 5.8. I get a lot of errors installing the module. By the other hand, I know that Perl now supports Unicode natively. What this really means? Why I initially loaded the module Unicode::MapUTF8 on the shell? Because the user can then use the function from_utf8() on the shell. And because I planned to auto-detect UTF8 encoded attribute values so I can convert values to a user-defined character set (ISO 8859-1 by default). I'm not an expert of UTF8. Do you know how we could improve UTF8 handling without using Unicode::MapUTF8 ? Thanks a lot! Rafael |
