Let me know, as this would either need to cause a change in legacy webcit, or
allow further development of ctdlsh to adjust those settings...
> Thu May 18 2017 10:26:10 PM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Package deps
>
>Data conversion code is done, table format is
will do.
> Wed May 17 2017 10:43:57 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Package deps
>
>I haven't pushed anything upstream yet because the build is broken (due
>to being in the middle of the change)
>
>but...
>
>If you just want to look, you can "git clone
Good to hear. I am making progress on getting it into pkgsrc.
Let me know when you have something for me to look at :)
> Tue May 16 2017 10:57:26 PM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Package deps
>
>They're stopped at those versions because I haven't updated
They're stopped at those versions because I haven't updated them. Latest
of each should work fine.
I am working to get citadel ported to pkgsrc for netbsd etc...
> Tue May 16 2017 09:15:02 AM EDT from bennabiy @ Uncensored Subject: Re:
>building on solarish...
>
>
>
>Actually I misunderstood the data structure. /usr is ro, but /usr/local/ is
>mounted rw, so it was not the issue. For
Actually I misunderstood the data structure. /usr is ro, but /usr/local/ is
mounted rw, so it was not the issue. For some reason, if I use the easy
install, libcurl and libcitadel do not compile, but if I manually go into the
directory and run the same configure command, it completes just fine...
Yes, I saw that coming, but hopefully not too much.
> Mon May 15 2017 03:05:42 PM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Eureka moment?
>
>
>>And yes, I can build a test system based on the git. Would you like me to
>
>
>>pull a branch, or master?
>
>It'll be in
>does this still need to be built with --enable-ares=$SUPPORT ? If so, then
it
>still depends on c-ares being built and installed in the system...
I thought I got rid of all those references, but if there's still one around
somewhere, then yes, we have to remove it. There is nothing
>I think I discovered the issue... will find out in a moment. It seems that
>/usr (which is mounted ro on this system) makes it impossible for
>/usr/local/blah to be created. I changed all instances in easyinstall script
>to use /opt/local/ instead, but do I need to also manually edit
>And yes, I can build a test system based on the git. Would you like me to
>pull a branch, or master?
It'll be in master when I commit it. I'm still working on the data structures
and conversion code for old databases ... trust me, it's easier do do the
conversion first than to try
still crashed... thoughts?
> Mon May 15 2017 09:20:01 AM EDT from bennabiy @ Uncensored Subject: Re:
>building on solarish...
>
>
>
>I think I discovered the issue... will find out in a moment. It seems that
>/usr (which is mounted ro on this system) makes it impossible for
I think I discovered the issue... will find out in a moment. It seems that
/usr (which is mounted ro on this system) makes it impossible for
/usr/local/blah to be created. I changed all instances in easyinstall script
to use /opt/local/ instead, but do I need to also manually edit all the
And yes, I can build a test system based on the git. Would you like me to
pull a branch, or master?
> Sat May 13 2017 10:29:45 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Eureka moment?
>
>
>I'm thinking we're just going to go with the static field in the user record
>for
Now, is there a way for the Aide user to edit the actual system email address
(either through webcit, or ctdlsh or... ?)
Will the email address still be populated from LDAP info?
> Sat May 13 2017 10:29:45 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Eureka moment?
>
>
>I'm
I'm thinking we're just going to go with the static field in the user record
for Internet email addresses. After decades of messing around with this data
model I'm inclined to keep it simple this time.
I was able to fit 16 addresses (with pretty long names in them) in 512
characters,
so
I will reply more tomorrow, but this is independent now of the vcard, so a
user can have as many as they like in the vcard, but in the actual system
mail table, they would only have x amount?
I would average on my systems that each user could have 2 addresses, some as
many as 6.
> Wed
Yes, it's called "my development system" :)
And I'm starting to have second thoughts about how I want to do this. It
might make more sense just to add a field to the user record inet_addrs[SIZ]
to stash as many email addresses as we can fit. SIZ is 4096 bytes. Maybe
1024 bytes would make
Agreed on the timing thing. As long as we are keeping that in mind, that is
"Good Enough For Now"...
Can you possibly make a version which does not "upgrade" per se, and just
starts with the correct data structure, and I could import the data into it
for testing through imapcopy or the
Changing the table format is the easy part. The time consuming part is writing the code to convert the old data format, add it to export/import, etc. So it'll take a bit of time. But I think it's worth it to get rid of a piece of the data model I've never liked too much.
I have some thoughts
Yes, I cannot see it working the way we would like without a couple new
tables (unless you want to be looking at it a couple years down the road
wondering why all the mess and hackwork...)
Those tables are not going to be *too* large though, just one more thing to
export. Also, would it be the
Still bellyaching over where we're going to put all this data. I do think
it's going to require some new tables.
ok, so that was strange... crashing might just be LONG lockups, but this last
one, the system was unresponsive (no activity taking place, but not
registering as crashed, IMAP timing out) but then I could go to webcit and
interact, and then it froze momentarily, and then was back up and working
iorarf_list... I love it.
> Thu May 04 2017 11:35:27 AM EDT from bennabiy @ Uncensored Subject: Re:
>email... and the data model.
>
>
>
>Looking over the serv_upgrade.c file now... (and back by the way).
>> Thu May 04 2017 10:20:36 AM EDT from IGnatius T Foobar @ Uncensored
>>Subject:
Looking over the serv_upgrade.c file now... (and back by the way).
> Thu May 04 2017 10:20:36 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: email... and the data model.
>
>
>>Is there somewhere in the database which keeps track of which version
>>things
>
>
>>are at
>Is there somewhere in the database which keeps track of which version things
>are at (which revision of the database?) It would probably be good to mark
>that this upgrade is a non reversible one (like going to 902 from 9.01).
Yes, that is almost exactly how it works. Citadel
It would be good in this to consider LDAP or other means of getting valid
info for users... What I mean by that is that there would be a way to
customize the way that info is imported from LDAP into the vCard or
CDB_DIRECTORY to allow for listings to be the way each system wants them (so
not
I think you and I are thinking similarly. I will reply to this more later,
but I think those are some good steps to take. If you can eliminate the
vCards being authoritative, and being purely informational, and yet, allow
for the aide to edit the authoritative entry somewhere, and allow for hidden
(continued)
Obviously this needs to change. The vCard shouldn't be the authoritative
source of the user's email address(es). And of course we're not going to
share the Global Address Book with other Citadels anymore ... or if we do,
it won't be driving a downstream address database.
I'm still thinking this through. It's going to fix a part of the data model
that I've never really liked. The email addresses in vCards were originally
used in the BBS days to store an email address where a user could be contacted.
Later on we overloaded it to store the user's email address on
Thank you for getting back to me. I was starting to wonder what rabbit hole
it was this time ;)
I think in general, I am with you on that. If the flag already exists in the
user database, then we can pretty easily make that selectable, either through
ctdlsh or a sendcommand, correct? I do
Actually, that might be doable. We have a flag in the user database to make
a user of the system "unlisted." It is a carryover from the BBS days, when
the ability to have such a thing seemed important ... but I'm having a hard
time remembering why.
My rethink of the Global Address Book
I think we should press on towards the Address Book issues (and LDAP in
general) unless I can get something solid to work with on the remaining
errors with this.
So, Can we make the possibility to make "invisible" users which have an entry
in the mailer table, as selectable by the aide
Ok, we might need to revisit the last fix a little as one of my users is
experiencing the message copying, but remaining in the source folder. Not
sure if it is a bug in evolution or with the patch. I tried it in Evolution
and "It worked for me" so...
Are you available today to hash some
I keep forgetting you go offline on Friday evening. Shabbat shalom!
That RFC looks a little tricky -- I'd rather work on the address book thing next.
NM, found it in our IRC conversation...
>
>
> This latest update removes the dependencies on libcares
>and libev
> Thu Apr 20 2017 02:45:33 PM EDTfrom bennabiy @ Uncensored Subject: Build
>requirements
>
>
>
>What dependencies got stripped from the build? libev and?
>
>
I fully agree...
I figured it would be an easy fix. I almost put together a patch for it, but
I did not have time.
> Tue Apr 18 2017 09:35:31 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: LDAP and users
>
>I'm going to have to review the RFC on that one. It looks pretty easy
I'm going to have to review the RFC on that one. It looks pretty easy to
fix. Citadel does use 1 as a fixed UIDVALIDITY value, since the message numbers
in our store are both persistent and strictly-ascending, so we can use them
as UIDs without resorting to any ugly hacks.
(Actually, the
That works for me. I am working on one of our delis tonight doing some
repairs.
I had one thing jump to the front of the Queue.
This is another broken implementation of IMAP which needs to be fixed to be
correct according to RFC.
There is no UIDVALIDITY being returned by the server
I was out for a week, and I'm back now. Tonight I have to finish my taxes
:(But this issue is still at the front of the queue and we'll have a look
at it this week.
Where are we at? Been about a week since this, with no input (I know you said
you were going to be out for about a week, so I am just checking where things
are at).
> Fri Apr 07 2017 11:54:59 AM EDT from bennabiy @ Uncensored Subject: LDAP
>and users
>
>
>
>I guess my last post was too
So far I like ctdlsh, but like you said, it is not mature. I tried export,
which took a long time, but for the life of me, I cannot find the resulting
file.
> Fri Apr 07 2017 04:04:51 PM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: WebCit and CitServer upgrade
>
>
>>Does ctdlsh
Good to hear... are the commands documented anywhere other than in the code?
Also, I just fired up webcit-ng and am staring at a 404 error from
uncensored. I noticed at one point you had w3.css somewhat in place, but it
seems that is gone...
> Fri Apr 07 2017 04:04:51 PM EDT from
>Does ctdlsh take standard commands as if sent through sendcommand?
No, ctdlsh takes ctdlsh commands. It communicates to the server through
the admin socket, just like sendcommand does, but where sendcommand is intended
to give you a way to speak to Citadel in its own language, ctdlsh is
Does ctdlsh take standard commands as if sent through sendcommand?
> Tue Feb 14 2017 11:24:16 PM EST from IGnatius T Foobar @ Uncensored
>Subject: Re: WebCit and CitServer upgrade
>
>
>Who was asking about ctdlsh? I made an important change to it tonight.
>
>You can now enter a single
Working for most: Thunderbird, Claws-mail, outlook 2010 and before (with
help) mac mail (with help)
Not working: EssentialPIM, Evolution (although the devs there were nice
enough to patch their client to work with the faulty server...) ,
outlook2013, Chaos Intellect (I am sure there were
Geez. I don't understand how IMAP could have been that broken for that long
without more people noticing it. Can you tell me which clients worked and
which ones didn't (before we fixed it) ?
I'm in US Eastern time too, but I usually hack Citadel late at night. Also
note that next week I
I am sure it works. I have so far tested it on a couple trouble clients (I do
not have outlook on hand to test, but honestly, if everything else works...)
Go ahead and push the change and I will commit to my production server.
Is it intentional to set the rcpt to limit to 172? I tried
Ok, I'm going to commit this change, because I'm pretty confident about it
-- but not 100.0% confident until you give it some more testing with various
clients. Let me know how the testing goes, and if you're satisfied with it
we will move on.
Great. I pulled from latest git master, patched and recompiled. looking good
so far on tested clients.
> Thu Apr 06 2017 01:30:34 PM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Status?
>
>
>>send me a patch and I can recompile :)
>
>Ok then ... I'll see your pastebin and
>send me a patch and I can recompile :)
Ok then ... I'll see your pastebin and raise you another pastebin.
https://pastebin.mozilla.org/9018168
Apply that patch and tell me if it fixes your IMAP client issue. If it fixes
the issue without breaking any other clients (and I
send me a patch and I can recompile :)
> Thu Apr 06 2017 11:48:32 AM EDT from IGnatius T Foobar @ Uncensored
>Subject: Re: Status?
>
>
>Ok, I think I see the problem.
>
>modules/imap/imap_fetch.c , line 190
>
>The \r\n sequence probably needs to be sent into the redirect buffer, not
Ok, I think I see the problem.
modules/imap/imap_fetch.c , line 190
The \r\n sequence probably needs to be sent into the redirect buffer, not
straight to the client.
Is there a way I can arrange for you to test your various clients without
outright putting that change into
I forgot to put the closing ) after the header... should go on a line by
itself
> Wed Apr 05 2017 05:12:19 PM EDT from bennabiy @ Uncensored Subject: Re:
>Status?
>
>
>
>And to note, other IMAP4rev1 servers I connect to issue the whole response
>before sending the headers (or message,
And to note, other IMAP4rev1 servers I connect to issue the whole response
before sending the headers (or message, etc)
so the response should be
>
>
>command: A007 UID FETCH 375286, 145418 (RFC822.SIZE RFC822.HEADER
>FLAGS)
>
>response: * 5 FETCH (UID 375286 RFC822.SIZE 3223
Here is a pastebin of a session with highlighting indicating the places where
trouble is caused...
https://pastebin.mozilla.org/9018063
Notice that the return from citserver on lines 41 and 42 should all be on one
line with no \r \n between them. Some clients are erroring because they never
That's a little confusing but I'll look at it later and try to pick it apart.
The ideal situation would be if you could give me the full RFC822 source
of a message that produces the problem (redacted of private information if
necessary) along with the IMAP command you are sending. Then if we
No, and thank you for asking.
Priority is getting IMAP working and fully implemented for all clients
(except outlook...) It would be beneficial to support the IMAPx, for the
server to be able to push email to clients by supporting IDLE, but if that is
too difficult to start with, then
Ok, for starters I have to apologize for appearing to have gone missing for
a little while in terms of development. I had a little thing I wanted to
do with citserver and ended up going down a rabbit hole that was way deeper
than I planned.
So ... is the address book on room access
Great. Looking forward to it in production.
> Mon Mar 27 2017 12:10:51 PM EDT from IGnatius T Foobar @
>Uncensored Subject: Re: A Side note
>
>
>>Have you considered using node.js for webcit backend?
>>
>
>Writing server-side code in JavaScript is all the
>Have you considered using node.js for webcit backend?
Writing server-side code in JavaScript is all the rage these days, and I'm
really happy about that.
WebCit-NG isn't a *complete* rewrite, though. I'm bringing over a lot of
code from legacy WebCit, which is written in C. The web
I'd prefer to discuss here; I don't usually frequent the IRC channel.
Do you have a better suggestion? I am often in the #citadel room in IRC, but
not much happens there...
I have somewhat of a priority list, but I think it would take a little
thought on how to word the first priority. I will wait to hear where you
would like the discussion to take place, and
Still on it, I'm just a little swamped at work right now. Let's figure out
what your biggest priorities are and we'll tackle them one at a time. We
can do it in here if you want to, but again, *one* at a time. Consider yourself
part of the QA team.
Sort of. In server.h there's an enum type, so start at 0 and count up :)
enum {
CDB_MSGMAIN,/* message base */
CDB_USERS, /* user file */
CDB_ROOMS, /* room index*/
That is great! I realized that was my issue when I tried to compile against
the git is that I still was following the webcit-ng branch, so the compile
failed.
It looks like it worked this time. Will let you know if there are issues.
> Sun Mar 12 2017 12:34:44 AM EST from IGnatius T
I remembered your original post about webcit-ng being hard coded to
uncensored, so, good to hear it is able to be switched.
I tried to do an update to my system, but it failed on webcit (I will have to
check to see if I have it still on the webcit-ng branch) I will let you know
tomorrow.
Yes, die in a car fire :)
webcit-ng can point to any Citadel server you want. Just append the host
and, optionally, port number of your Citadel server to the webcit command
line.
Of course, webcit-ng is kind of broken right now because the switch to w3.css
isn't fully absorbed,
So, die in a car fire?
> Wed Mar 08 2017 04:40:24 PM EST from IGnatius T Foobar @ Uncensored
>Subject: Re: git.citadel.org down?
>
>... So ... thank you for that.
>
>
>
>
Whoops ... I put that server behind a firewall a couple of weeks ago and didn't
open port 9418. Should be fixed now.
Thanks for tracking the latest changes. For some reason, your feedback has
really been great for the project recently. It must be that you are providing
just the right
I tried it out, and while it looks like certain aspects are working quite
well, there are some issues with rooms loading when logged in vs not logged
in through Firefox and Opera.
I have not looked at the backend implementation yet.
> Thu Feb 23 2017 09:49:05 AM EST from IGnatius T Foobar @
So far it's working pretty well. I used the guide at
http://youmightnotneedjquery.com/
to figure out how to do things without jQuery. Now maybe I've done something
horrible, because I've only tested so far on Chrome, but we'll see.
The inspiration was of course the fact that W3CSS doesn't
That is great. I think these days people are way too dependent on jQuery and
the like when often times it can be done easier.
> Mon Feb 20 2017 09:11:30 PM EST from IGnatius T Foobar @ Uncensored
>Subject: Re: Choosing the right framework
>
>
>Update on that:
>
>I did it. No more jQuery.
Hey the_mgt , check out the current code in the webcit-ng branch ... I have
*roughly* fit in one of the W3CSS templates. Looks promising but it needs
some tweaking (which you can probably do better than I can).
Update on that:
I did it. No more jQuery.
There wasn't a thing in there that couldn't be done a lot easier by simply
making calls to getElementById() and XMLHttpRequest() directly.
I also noticed that the w3css framework is "pure CSS no javascript". Well
not quite, the framework is defined in 100% pure CSS but you still need to
use javascript, for example, to invoke and hide a modal.
But the interesting thing is that, unlike pretty much every other framework
out there,
All right, nothing interesting there. :(
Here is a directory listing of /usr/local/citadel
>
>
>/usr/local/citadel# ls -lah
>
>total 4.7M
>
>drwxr-xr-x 15 citadel citadel 4.0K Feb 15 03:02 .
>
>drwxr-xr-x 13 root root 4.0K Feb 1 17:30 ..
>
>-rwxr-xr-x 1 citadel staff 100K Jan 26 18:00 aidepost
>
>-rwxr-xr-x 1
I was asking about ctdlsh... and that is very encouraging! I am currently
having to use sendcommand in my scripts when I add a new user.
Can you make a section where you document the current abilities of it? Also,
it would be nice to have the server configuration all in one place and
document
>citserver is not crashing, but it loses or forgets its port number, (I noted
>that you had this same thing happen, but I can reproduce it almost every
time
>citserver is restarted) and it does not remember the setting about not
>denying self service registration. I have rooms which are
Who was asking about ctdlsh? I made an important change to it tonight.
You can now enter a single ctdlsh command on the command line, which makes
ctdlsh usable in scripts. (Basically I was taking a cue from virsh, which
I think was my inspiration for ctdlsh in the first place.)
So
>Also, What do you think about multiple auth? Internal (able to be disabled,
>but always on by default) plus optionally host based,and optionally LDAP
(or
>AD, not both). If the user database / mail table was able to parse a list
of
You know what, we actually used to do that.
Bah, I hate when I make a typo... Compiled successfully
> Mon Feb 13 2017 15:08:35 EST from bennabiy @ Uncensored Subject: Re:
>Choosing the right framework
>
>
>
>All fine and wonderful... I guess libraries help ;)
>>
>>
>>/citadel/webcit# make
>>
>>gcc -c -ggdb http.c -o
All fine and wonderful... I guess libraries help ;)
>
>
>/citadel/webcit# make
>
>gcc -c -ggdb http.c -o http.o
>
>In file included from http.c:15:0:
>
>webcit.h:34:24: fatal error: libcitadel.h: No such file or directory
>
> #include
>
> ^
>
>compilation
Great. I will check it out. What is the current state of ctdlsh? I am
considering the administration abstraction from the UI. If the API docs for
citserver and webcit could be reworked to reflect the current state, we might
be able to get a working Admin console for both. I am thinking of webmin
>If I wanted to fire up webcit-ng, is it in the place where I would be able
to
>compile and test it out? If so, and since you removed the automake stuff,
>what is the recommended method for compiling now?
Very easy. :)
1. git checkout webcit-ng
2. cd webcit
3. make
4.
If I wanted to fire up webcit-ng, is it in the place where I would be able to
compile and test it out? If so, and since you removed the automake stuff,
what is the recommended method for compiling now?
> Mon Feb 13 2017 10:16:43 EST from IGnatius T Foobar @ Uncensored Subject:
>Re: Choosing
http://www.citadel.org/doku.php/documentation:appproto:system_config lists
>
>
>
>CONF (get or set global CONFiguration options)
>
>
>
>
>Retrieves or sets various system-wide configuration and policy options. This
>command is only available to Admins.
>
>The six forms of this command
Sun Feb 12 2017 18:07:32 ESTfrom IGnatius T Foobar @ Uncensored Subject: Re: Choosing the right framework
I understand IG's wish to choose something wide spread and well supported, so that it stays with us and our slow web-technique adaption cycle. But we also need something that both
Wed Feb 08 2017 12:21:22 PM EST from bennabiy @ Uncensored Subject: Re: WebCit and CitServer upgrade
As to the clobbering, I think you are right. The login hook seems to be the culprit. I would be fine with it if I could map LDAP attributes to fields on the server without having to
Looking in the source, I see you use PUTVAL rather than the documented SETVAL. What is the difference?
According to the code, PUTVAL is correct; SETVAL is not. Did I write it as SETVAL somewhere? If so please point out where and I'll fix it!
I understand IG's wish to choose something wide spread and well supported, so that it stays with us and our slow web-technique adaption cycle. But we also need something that both programmers and web designers (or rather the guy that patches the html together) can understand fast. That is why
Looking in the source, I see you use PUTVAL rather than the documented
SETVAL. What is the difference?
> Sun Feb 12 2017 12:21:30 EST from bennabiy @ Uncensored Subject: DB
>frontend for configuration troubleshooting
>
>
>
>I tried to set the registration setting:
>
>CONF
That looks nice and simple. I agree, XHTML is not fun to work with.
> Sun Feb 12 2017 11:16:44 EST from the_mgt @ Uncensored Subject: Choosing
>the right framework
>
>
>
>So I was in the mood for some webdesign this weekend and searched again for
>the proper tool to use.
>
>I already
I use OpenLDAP on Debian and Ubuntu primarily, so standard rfc2307.
I do not have the citadel.schema or rfc2739.schema in place.
As to the clobbering, I think you are right. The login hook seems to be the
culprit. I would be fine with it if I could map LDAP attributes to fields on
the server
Regarding the vCard update issue:
Ctdl_LDAP_to_vCard() appears to be the function that painstakingly copies
LDAP attributes into a user's vCard. "git blame" has my name and a 2014
timestamp
on it, but I don't remember writing it ... maybe it was submitted as a patch
and I put it in.
That was my thought as well. It would be good to make clear API documentation for it.
What is the current state of ctdlsh?
For the webmin module, check out what we're doing with ctdlsh. The eventual
goal is for that to be a one-stop-shop for doing server configuration without
having to actually log in to Citadel. A webmin module (or any other type
of tool) could theoretically make calls to ctdsh to do its work.
>From IGnatius T Foobar in email with my answers and comments interspersed
>
>
>*/Let's have this conversation in the Citadel Development room. I like
>what you're saying and we should work on it./*
>
>*/If I understand you correctly, you're talking about the way Citadel
>populates a user's
Well, there's more than just the database. So far we've discussed the current effort to get 100% of Citadel's persistent storage into the database. We've already moved citadel.config and citadel.control, along with user profiles (bios), user photos (userpic), netconfigs, etc. We should have
BerkeleyDB docs suggest the way to go is "a thin client side replication layer", but I don't think that's the best route because it's not 'master-master'.
Dovecot has a 'dsync' module, it's a 'master-master' answer. If citadel could add that, then this equation would be true for groups 500 and
501 - 600 of 1530 matches
Mail list logo