Menu
▾
▴
apc-aa-coders — Discussion only useful to active developers.
You can subscribe to this list here.
| 2002 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(21) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2003 |
Jan
(136) |
Feb
(104) |
Mar
(80) |
Apr
(21) |
May
(36) |
Jun
(18) |
Jul
(22) |
Aug
(5) |
Sep
(11) |
Oct
(7) |
Nov
(3) |
Dec
(12) |
| 2004 |
Jan
(5) |
Feb
|
Mar
(3) |
Apr
(1) |
May
|
Jun
(1) |
Jul
(3) |
Aug
(6) |
Sep
(3) |
Oct
(4) |
Nov
(4) |
Dec
|
| 2005 |
Jan
(1) |
Feb
(4) |
Mar
|
Apr
(7) |
May
(4) |
Jun
(7) |
Jul
(4) |
Aug
(4) |
Sep
|
Oct
(8) |
Nov
(4) |
Dec
(1) |
| 2006 |
Jan
|
Feb
(3) |
Mar
(2) |
Apr
(4) |
May
|
Jun
(6) |
Jul
(2) |
Aug
(1) |
Sep
|
Oct
(3) |
Nov
|
Dec
|
| 2007 |
Jan
|
Feb
(4) |
Mar
(7) |
Apr
(4) |
May
(4) |
Jun
(8) |
Jul
(2) |
Aug
(2) |
Sep
|
Oct
|
Nov
(2) |
Dec
|
| 2009 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(2) |
Jun
(1) |
Jul
(3) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2010 |
Jan
|
Feb
(1) |
Mar
|
Apr
(2) |
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
|
From: Mitra <mi...@ea...> - 2003-02-16 01:12:46
|
That looks like something i did ...
I made the changes to bring all the unaliasing into one place, so
that {....} worked the same wherever.
Looking over this code in detail, it seems that the old version and
the broken version do the same thing , i.e. $item->unalias($value)
calls new_unalias_recurent($value,"",0,0,$this)
The only difference is that the aliases are now accepted directly by
new_unalias_recurent rather than creating a dummy item.
What broke?
- Mitra
At 9:06 PM +0100 15/2/03, Jakub Adamek wrote:
>Hi,
>
>someone broke my code in mail.php3, function send_mail_from_table. My
>own code was commented out and the function new_unalias_recurent was
>there instead. But the parameters to this function were all bad and it
>was not working in the cycle as it should and so on.
>
>Why? Who? Not that it was so much damage, I only deleted the changes and
>returned to my code. But it is strange because the changes were really
>so incorrect.
>
>Jakub
>
>This is the broken state:
>
> if (is_array ($aliases)) {
>/*
> // I don't know how to work with unaliasing. Thus I try to
>pretend
> // having an item.
> reset ($aliases);
> while (list ($alias, $translate) = each ($aliases)) {
> // I create the "columns"
> $cols[$alias][0] = array (
> "value" => $translate,
> "flag" => FLAG_HTML);
> // and "aliases"
> $als [$alias] = array ("fce"=>"f_h", "param"=>$alias);
> }
> $item = new Item ("", $cols, $als, "", "" ,"");
>*/
> while (list ($key, $value) = each ($record))
># $record[$key] = $item->unalias ($value);
> $level = 0; $maxlevel = 0;
> $record[$key] = new_unalias_recurent($value, "", $level,
>$maxlevel, null, null, $aliases);
> }
>
>
>
>-------------------------------------------------------
>This sf.net email is sponsored by:ThinkGeek
>Welcome to geek heaven.
>http://thinkgeek.com/sf
>_______________________________________________
>Apc-aa-coders mailing list
>Apc...@li...
>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
--
Mitra Technology Consulting - www.mitra.biz - mi...@mi...
02-6684-8096 or 0414-648-0722
Life is a Mystery to be Lived, not a Problem to be Solved
|
|
From: Jakub A. <jak...@se...> - 2003-02-15 20:07:03
|
Mitra, you had an error on line 185 in zids.php3, qqquote instead of qqqquote. Fixed. Jakub |
|
From: Jakub A. <jak...@se...> - 2003-02-15 20:07:02
|
Hi,
someone broke my code in mail.php3, function send_mail_from_table. My
own code was commented out and the function new_unalias_recurent was
there instead. But the parameters to this function were all bad and it
was not working in the cycle as it should and so on.
Why? Who? Not that it was so much damage, I only deleted the changes and
returned to my code. But it is strange because the changes were really
so incorrect.
Jakub
This is the broken state:
if (is_array ($aliases)) {
/*
// I don't know how to work with unaliasing. Thus I try to
pretend
// having an item.
reset ($aliases);
while (list ($alias, $translate) = each ($aliases)) {
// I create the "columns"
$cols[$alias][0] = array (
"value" => $translate,
"flag" => FLAG_HTML);
// and "aliases"
$als [$alias] = array ("fce"=>"f_h", "param"=>$alias);
}
$item = new Item ("", $cols, $als, "", "" ,"");
*/
while (list ($key, $value) = each ($record))
# $record[$key] = $item->unalias ($value);
$level = 0; $maxlevel = 0;
$record[$key] = new_unalias_recurent($value, "", $level,
$maxlevel, null, null, $aliases);
}
|
|
From: Mitra <mi...@ea...> - 2003-02-14 22:52:15
|
Ok - so I've tracked it down a bit, but I really need someone else (this is supposed to be open-source! ) to help with some of this. Take a look at: http://earth.path.net/apc-aa+/view.php3?vid=45&cmd[45]=x-45-467&debugcache=1 and http://earth.path.net/apc-aa+/view.php3?vid=45&cmd[45]=x-45-2240&debugcache=1 They are both using the same query string to pagecache->get which has no mention of the item id in it. - Mitra >Something in the latest CVS seems to have completely broken caching >- I'm seeing an index.shtml with multiple views turning up the same >info on each, if I run it with nocache=1 the page works ok. > >Anyone got any ideas? Anyone else seeing this, This is really nasty >- and I would recommend against updating to the CVS unless you are >planning on helping debug it! > >Unfortunately I have to go out so can't look at it myself. > >- Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-14 22:37:38
|
Thanks Marek - I understood this much already, the question is how it was intended to include them - i.e. at what level, in slice.php3 or view.php3 or what? Jakub - this has started breaking since you did all the "require_once" stuff, can you look at it. Marek - I'm not going to get a chance to look at the problem you are describing for several days! I'vegot to fix the cache problem (which may be whats causing this, and is not code I've worked on before) and then I've got to move a site - having wasted several days trying to get import/export working. Anyway - if you want help looking at a problem like this, you have to provide as simple a URL as possible that fails, so I can run it with debug=3D1 and see what is happening. Please look at the FAQ on hints on debugging. - Mitra At 10:33 AM +0100 14/2/03, Marek Tichy wrote: >Hi Mitra, >first - could you please look at this first (not sure wheather it's >related or not): > >> >Another weird problem I have is that slices don't show up - the >> >server returns blank document, as if it had crashed php. It happens >> >in slice.php3 at this spot: >> > >> >$zids=3DQueryZIDs($fields, $slice_id, $conds, $sort, $slice_info[group_= by], >> > "ACTIVE", $slices, $neverAllItems, 0, >> >$defaultCondsOperator, true ); > >OK. Generally locsess is a normal phplib session management >(localised for . While locsessi is patched to get session parameters >also from QUERY_STRING_UNESCAPED. (i at the end stands for >Includes). So as a rule of thumb, locsessi should only be used in the >"public" part of AA - slice, view etc... Anywhere else, where the >page is generated as a whole (admin part) and not being included >vi SSI, using locsess is fine. > >Hope it helps. > >Marek > > > > > > >> -----Original Message----- >> From: apc...@li... >> [mailto:apc...@li...]On Behalf Of Mitra >> Sent: 13. =FAnora 2003 23:01 >> To: Marek Tichy; apc...@so... >> Subject: RE: [Apc-aa-coders] extsess and extsessi >> >> >> Yes - that sounds like the problem, yes - extsess is only included >> by locsess and extsessi by locsessi, but we have some cases where >> locsess is being included AS WELL as locsessi and other cases where >> neither are being included. >> >> I'd fix it, but i can't figure out what, whoever wrote this, >> intended. In particular when to include one and when the other. I >> think we need a single file to include "locsesseither.php3" which >> includes one and only one of the other files, but I don't understand >> when one should be included, and when the other. >> >> - Mitra >> >> At 2:18 PM +0100 13/2/03, Marek Tichy wrote: >> >Hi Mitra, >> >I'm getting an error in latest cvs, which i think is related to this. > > > > > > Call to a member function on a non-object in > > ><b>/raid/www/htdocs/aa.ecn.cz/apc-aa- > > > cvs/include/itemfunc.php3</b> on line <b>1172</b><br /> > > > > > >Which is: > > > relatedwindow =3D open( "'. $sess->url("related_sel.php3") ........= =2E.. >> > >> >As far as I remember, extsess and extsessi are extensions to the >> >phplib session class. >> >I can't see any reason why would those two be included anywhere else >> >than locsess /locsessi, and also in locauth. >> > >> > >> >Don't think I can move this any forward. > > > > > >Best > > > Marek > > > > > > > > >> -----Original Message----- > > >> From: apc...@li... > > >> [mailto:apc...@li...]On Behalf Of Mitr= a > > >> Sent: 13. =FAnora 2003 4:46 >> >> To: apc...@so... >> >> Subject: [Apc-aa-coders] extsess and extsessi >> >> >> >> >> >> The change to require_once seems to have worked fairly well, except >> >> for one problem. >> >> >> >> There are some problems with extsess and extsessi >> >> >> >> If you include one, you can't include the other. >> >> >> >> But .... code has to include one or the other to function, > > >> >> >> What I'm finding is that certain combinations include either none or >> >> both and so break. >> >> >> >> Where SHOULD these be included from ? >> >> >> >> - Mitra >> >> >> >> -- >> >> Mitra Technology Consulting - www.mitra.biz - mi...@mi... >> >> 02-6684-8096 or 0414-648-0722 >> >> >> >> Life is a Mystery to be Lived, not a Problem to be Solved >> >> >> >> >> >> >> >> ------------------------------------------------------- >> >> This sf.net email is sponsored by:ThinkGeek >> >> Welcome to geek heaven. > > >> http://thinkgeek.com/sf >> >> _______________________________________________ >> >> Apc-aa-coders mailing list >> > > Apc...@li... >> > > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders >> > > >> >> >> -- >> Mitra Technology Consulting - www.mitra.biz - mi...@mi... >> 02-6684-8096 or 0414-648-0722 >> >> Life is a Mystery to be Lived, not a Problem to be Solved >> >> >> >> ------------------------------------------------------- >> This SF.NET email is sponsored by: FREE SSL Guide from Thawte >> are you planning your Web Server Security? Click here to get a FREE >> Thawte SSL guide and find the answers to all your SSL security issues. >> http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en >> _______________________________________________ >> Apc-aa-coders mailing list >> Apc...@li... >> https://lists.sourceforge.net/lists/listinfo/apc-aa-coders >> > > > >------------------------------------------------------- >This SF.NET email is sponsored by: FREE SSL Guide from Thawte >are you planning your Web Server Security? Click here to get a FREE >Thawte SSL guide and find the answers to all your SSL security issues. >http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en >_______________________________________________ >Apc-aa-coders mailing list >Apc...@li... >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-14 22:37:33
|
Your proposal would make sense, except that I'd like to be able to re-import the same site later, along with any new views created. Since this (like just about any real case) is a multi-slice situation, its much more complicated than the auto-increment solution will allow for. I spent several days trying to fix and then extend the import code to handle views, but that was before I spotted the problem with short ids. At the moment I'm looking at either 1: Setting up a completely seperate apc-aa database for imports from this machine, and just copying the database 2: Abandoning hope of re-importing and doing a lot of emacs on the stuff when I move it. 3: Spending a bunch more days changing views to support names as well as vid= s. Unfortunately I need to spend some time building the sites, so don't have time to spend more energy trying to fix apc-aa ! - Mitra At 10:24 AM +0100 14/2/03, Marek Tichy wrote: >Hi Mitra, >yes this is a MySQL autoincrement behaviour. For any autoincrement >field, MySql keeps little pointer which contains the highest >value. This is being updated every time you INSERT/UPDATE the >record, regardles you specify the value for the autoincremented field >or not. >I don't think it's MySQL specific - I'd expect any SQL engine should >behave like this. > >I can't think about any easy solution for this. That's the pain of >merging autoincrements between two systems. BTW, maybe i should >mention it here because recently Honza told me that even he wasn't >aware of what was originally behind the decision to use long IDs. >It was: > >1) Every long ID is unique (in practice) >2) Some SQL databases don't have autoincrement (MySQL was one of >them when we started creating AA ). We wanted to keep the code >universal for any database system > >The 2 has already been broken for various other reasons and also >turned out not to be such an important issue (but may come back to >consideration again with the spread of AA). > >So - I also hate long IDs, especially packed ones, but there is >something about it. > >I'm not sure wheather I'm getting it righ but you seem to already >have some proposal for changing view IDs at the moment of >transfering them between two systems. So you must have sorted out >the problem of changing stuff like "cmd[383]" in all aliases >definitions, format strings and shtml pages. If you can do this, >than why don't you just count how many views you will need for the >import and than generate them as auotincrement, while keeping track >on what the new IDs are ? > >Best > Marek > > > > >> -----Original Message----- >> From: Mitra [mailto:mi...@ea...] >> Sent: 13. =FAnora 2003 22:56 >> To: Marek Tichy >> Subject: RE: [Apc-aa-coders] Inserting Views with out-of-sequence ids >> >> >> Sorry, lets see if I've understood you correctly, a typical example > > might be .... > > > > MySQL creates id's 1-10 for new records > > I add records specifically at 1001-1023 > > MySQL then creates the next id as 1024 > > > > Is that what you mean? >> >> That unfortunately is not very good behavior because part of the >> point of putting the added ids at 10001... >> is to get them out of the way of the auto-created ids. So that if a >> second import is done from the same slice they can safely overwrite >> using the same algorithm - e.g. adding 10,000 >> >> Is this APC-AA behavior, or something built into MySQL? >> >> - Mitra >> >> >> At 12:46 PM +0100 13/2/03, Marek Tichy wrote: >> >Hi, >> > >> >> As part of importing / exporting, I've got a problem with shortids. >> >> >> >> Lets take views for example - they have shortids which are an >> >> auto-increment field, >> >> >> >> When I import from one system to another, I want to add an offset to >> >> avoid clashes, for example I might add 1000 so that I can a) do it i= n >> >> emacs and b) understand the result and fix mistakes. >> >> >> >> Will this cause problems with the autoincrement field in any way? Fo= r >> >> example there might be views 1-10 and 1001 -1024, so will the next > > >> auto-incremented view by 11 ? And what happens when it >> > > auto-increments to 1000? >> > >> >It will always start from 1024. >> >> >> -- >> Mitra Technology Consulting - www.mitra.biz - mi...@mi... >> 02-6684-8096 or 0414-648-0722 >> >> Life is a Mystery to be Lived, not a Problem to be Solved >> > > > >------------------------------------------------------- >This SF.NET email is sponsored by: FREE SSL Guide from Thawte >are you planning your Web Server Security? Click here to get a FREE >Thawte SSL guide and find the answers to all your SSL security issues. >http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en >_______________________________________________ >Apc-aa-coders mailing list >Apc...@li... >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-14 22:37:25
|
Something in the latest CVS seems to have completely broken caching - I'm seeing an index.shtml with multiple views turning up the same info on each, if I run it with nocache=1 the page works ok. This is a simple site - no use of the site module or anything that might trigger problems we've seen recently on more complex sites. Anyone got any ideas? Anyone else seeing this, This is really nasty - and I would recommend against updating to the CVS unless you are planning on helping debug it! Unfortunately I have to go out soon to the big peace rally so I might not be able to fix it quickly - I don't think its to do with any code I've changed, so its going to be harder to fix. - Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Jakub A. <jak...@se...> - 2003-02-14 18:44:55
|
I think that you are not right. The idea is to only allow to use the same username / password combination for the web (HTTP auth) and AA (sessions). No other settings are shared between Reader Management slice and AA permitions. Thus the only thing we need is when the user changes his or her password on Reader Personal Details page, so that the new password is propagated into AA. But I believe that it would be easy to do it the other way too if desired. Jakub > -----Original Message----- > From: apc...@li... > [mailto:apc...@li...] On Behalf Of Mitra > Sent: Thursday, February 13, 2003 4:08 AM > To: Jakub Adamek; apc...@so... > Subject: RE: [Apc-aa-coders] Reader Management Slice > > > I think if it is not possible to merge them, then what's needed is > almost the opposite of what you suggest. > > I.e. there are typically more Readers than AA users, so when a user's > password is changed in AA, then they should be automatically added to > the Readers slice so that HTTP authentication can find it? > > Ram and I were doing something different, he had HTTP authentication > working off the normal AA permissions - and I was adding a "READ" > permission, my part never got finished, and Ram's wasn't using Groups > yet, so probably what you have will be better - especially since > non-superadmins can manage it. > > - Mitra > > > At 6:39 PM +0100 11/2/03, Jakub Adamek wrote: > >Norbert, Mitra and all, > > > >we were discussing in Econnect the unhappy situation of two user > >management systems and it does not seem possible to merge > them. The AA > >permission system has many features which would be really > hard to model > >by a slice. On the other hand the slice approach has such features > >which can't be modelled by the AA system. > > > >Our proposal is to add another Reader Management module, which will > >synchronize the AA permissions from the slice. It would be > only one-way > >synchronization, i.e. users are added to AA permissions on > addition to > >the slice and their password and email info changed in AA pemissions > >when changed in the slice. > > > >Please do think about it and write any ideas you have. It was a hard > >way to come to the decision to put the Readers into a slice, > perhaps it > >would be possible to add AA permissions there also. But at > the moment I > >do not much believe in it. > > > >Jakub > > > >> -----Original Message----- > >> From: apc...@li... > >> [mailto:apc...@li...] On Behalf Of > >> Norbert Brazda > >> Sent: Sunday, February 09, 2003 1:38 PM > >> To: apc...@so... > >> Subject: Re: [Apc-aa-coders] Reader Management Slice > >> > >> > >> From: "Mitra" <mi...@ea...> > >> > One concern .... a common situation I have is where the user's > >> have > ability to edit some parts of a site, and > controlled reader > >> access to > those and other parts of a site. For example > this is the > >> case where > you have a the board of an organization wanting a > >> shared space to > post news articles. I'd like NOT to > have to give > >> these people two > different sets of login ids, and > manage them in > >> two different places. > >> > Can we integrate the existing user management process for > >> > apc-aa/admin with this? So that one interface can be > used to add > >> > users to groups for managing parts of other slices.. > > > > > > So why not to put all user (not only reader) related > > > information into the slice? > > > > > > And I would like to suggest one more change if you are > going > to > > do it: User have no longer access to the anonymous editing > >> when is the item changed by administrator > >> (ITEM_FLAG_ANONYMOUS_EDITABLE is reset through admin > interface). It > >> would be good to have this as a option. In come cases we > want users > >> to change thier data even after they were changed by the > >> adminstrator (e.g. directory of NGOs - sometime NGOs > update data by > >> themseleves, sometimes it is done by our editors). > >> > >> > >> > >> > >> > >> Norbert Brazda > >> br...@ch... > >> tel: 0905-729359 > >> > >> ChangeNet - informacny servis o obcianskej spolocnosti > Mlynske nivy > >> 41, 821 09 Bratislava, Slovakia > >> tel/fax: 02-55560026, eMail: in...@ch... > >> http://www.changenet.sk > >> > >> SPAJAME LUDI, KTORI MENIA SVET > >> > >> > >> > >> > >> ------------------------------------------------------- > >> This SF.NET email is sponsored by: > >> SourceForge Enterprise Edition + IBM + LinuxWorld > >> http://www.vasoftware.com > >> _______________________________________________ > >> Apc-aa-coders mailing list Apc...@li... > >> https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > >> > > > > > > > >------------------------------------------------------- > >This SF.NET email is sponsored by: > >SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! > >http://www.vasoftware.com > >_______________________________________________ > >Apc-aa-coders mailing list > >Apc...@li... > >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > > > -- > Mitra Technology Consulting - www.mitra.biz - > mi...@mi... 02-6684-8096 or 0414-648-0722 > > Life is a Mystery to be Lived, not a Problem to be Solved > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Welcome to geek heaven. > http://thinkgeek.com/sf > _______________________________________________ > Apc-aa-coders mailing list > Apc...@li... > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > |
|
From: Jakub A. <jak...@se...> - 2003-02-14 18:38:07
|
Mitra, IMHO I believe export / import was never so much working that anybody would have any data of value exported. I believe you can do without backward compatibility. But it's only my humble opinion. And as I write this, I will add that: It's really great for AA that you are working on this. If you care about history, the slice export / import in a very simple version was created by me in my very beginnings with AA and PHP (one and a half year ago). It was re-written by some extern staff in Econnect who didn't care enough and who was confused with my beginners code. It was never working for me when I needed it. I am not sure if others have any experiences with export / import? Jakub > -----Original Message----- > From: apc...@li... > [mailto:apc...@li...] On Behalf Of Mitra > Sent: Thursday, February 13, 2003 12:32 PM > To: apc...@so... > Subject: [Apc-aa-coders] Import / Export backward compatability > > > Does anyone think that sliceimport and export are being used for > storing data over a period. > > I'm wondering, because it would make life easier if I could break > backward compatability - the code is in such poor condition, that > being able to change the output format and corresponding input > processing would be much simpler if I don't have to handle older > versions. > > My assumption is that people are > a) Moving data from site to site pretty much instantaneously, i.e. > they export, process then import. > b) that systems at both ends of the link can typically be assumed to > run current versions of apc-aa. > > I'm only making these assumptions because the code appears to be > mostly non-functional, so I can't imaging someone is using it for > archiving / backup purposes. > > Any comments .... > > - Mitra > > -- > Mitra Technology Consulting - www.mitra.biz - > mi...@mi... 02-6684-8096 or 0414-648-0722 > > Life is a Mystery to be Lived, not a Problem to be Solved > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: FREE SSL Guide from > Thawte are you planning your Web Server Security? Click here > to get a FREE Thawte SSL guide and find the answers to all > your SSL security issues. > http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en > > _______________________________________________ > Apc-aa-coders mailing list > Apc...@li... > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > |
|
From: Marek T. <ma...@gn...> - 2003-02-14 09:36:00
|
Hi Mitra, first - could you please look at this first (not sure wheather it's related or not): > >Another weird problem I have is that slices don't show up - the > >server returns blank document, as if it had crashed php. It happens > >in slice.php3 at this spot: > > > >$zids=QueryZIDs($fields, $slice_id, $conds, $sort, $slice_info[group_by], > > "ACTIVE", $slices, $neverAllItems, 0, > >$defaultCondsOperator, true ); OK. Generally locsess is a normal phplib session management (localised for . While locsessi is patched to get session parameters also from QUERY_STRING_UNESCAPED. (i at the end stands for Includes). So as a rule of thumb, locsessi should only be used in the "public" part of AA - slice, view etc... Anywhere else, where the page is generated as a whole (admin part) and not being included vi SSI, using locsess is fine. Hope it helps. Marek > -----Original Message----- > From: apc...@li... > [mailto:apc...@li...]On Behalf Of Mitra > Sent: 13. února 2003 23:01 > To: Marek Tichy; apc...@so... > Subject: RE: [Apc-aa-coders] extsess and extsessi > > > Yes - that sounds like the problem, yes - extsess is only included > by locsess and extsessi by locsessi, but we have some cases where > locsess is being included AS WELL as locsessi and other cases where > neither are being included. > > I'd fix it, but i can't figure out what, whoever wrote this, > intended. In particular when to include one and when the other. I > think we need a single file to include "locsesseither.php3" which > includes one and only one of the other files, but I don't understand > when one should be included, and when the other. > > - Mitra > > At 2:18 PM +0100 13/2/03, Marek Tichy wrote: > >Hi Mitra, > >I'm getting an error in latest cvs, which i think is related to this. > > > > Call to a member function on a non-object in > ><b>/raid/www/htdocs/aa.ecn.cz/apc-aa- > > cvs/include/itemfunc.php3</b> on line <b>1172</b><br /> > > > >Which is: > > relatedwindow = open( "'. $sess->url("related_sel.php3") ........... > > > >As far as I remember, extsess and extsessi are extensions to the > >phplib session class. > >I can't see any reason why would those two be included anywhere else > >than locsess /locsessi, and also in locauth. > > > > > >Don't think I can move this any forward. > > > >Best > > Marek > > > > > >> -----Original Message----- > >> From: apc...@li... > >> [mailto:apc...@li...]On Behalf Of Mitra > >> Sent: 13. února 2003 4:46 > >> To: apc...@so... > >> Subject: [Apc-aa-coders] extsess and extsessi > >> > >> > >> The change to require_once seems to have worked fairly well, except > >> for one problem. > >> > >> There are some problems with extsess and extsessi > >> > >> If you include one, you can't include the other. > >> > >> But .... code has to include one or the other to function, > >> > >> What I'm finding is that certain combinations include either none or > >> both and so break. > >> > >> Where SHOULD these be included from ? > >> > >> - Mitra > >> > >> -- > >> Mitra Technology Consulting - www.mitra.biz - mi...@mi... > >> 02-6684-8096 or 0414-648-0722 > >> > >> Life is a Mystery to be Lived, not a Problem to be Solved > >> > >> > >> > >> ------------------------------------------------------- > >> This sf.net email is sponsored by:ThinkGeek > >> Welcome to geek heaven. > >> http://thinkgeek.com/sf > >> _______________________________________________ > >> Apc-aa-coders mailing list > > > Apc...@li... > > > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > > > > > > -- > Mitra Technology Consulting - www.mitra.biz - mi...@mi... > 02-6684-8096 or 0414-648-0722 > > Life is a Mystery to be Lived, not a Problem to be Solved > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: FREE SSL Guide from Thawte > are you planning your Web Server Security? Click here to get a FREE > Thawte SSL guide and find the answers to all your SSL security issues. > http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en > _______________________________________________ > Apc-aa-coders mailing list > Apc...@li... > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > |
|
From: Marek T. <ma...@gn...> - 2003-02-14 09:32:47
|
Hi Mitra, yes this is a MySQL autoincrement behaviour. For any autoincrement field, MySql keeps little pointer which contains the highest value. This is being updated every time you INSERT/UPDATE the record, regardles you specify the value for the autoincremented field or not. I don't think it's MySQL specific - I'd expect any SQL engine should behave like this. I can't think about any easy solution for this. That's the pain of merging autoincrements between two systems. BTW, maybe i should mention it here because recently Honza told me that even he wasn't aware of what was originally behind the decision to use long IDs. It was: 1) Every long ID is unique (in practice) 2) Some SQL databases don't have autoincrement (MySQL was one of them when we started creating AA ). We wanted to keep the code universal for any database system The 2 has already been broken for various other reasons and also turned out not to be such an important issue (but may come back to consideration again with the spread of AA). So - I also hate long IDs, especially packed ones, but there is something about it. I'm not sure wheather I'm getting it righ but you seem to already have some proposal for changing view IDs at the moment of transfering them between two systems. So you must have sorted out the problem of changing stuff like "cmd[383]" in all aliases definitions, format strings and shtml pages. If you can do this, than why don't you just count how many views you will need for the import and than generate them as auotincrement, while keeping track on what the new IDs are ? Best Marek > -----Original Message----- > From: Mitra [mailto:mi...@ea...] > Sent: 13. února 2003 22:56 > To: Marek Tichy > Subject: RE: [Apc-aa-coders] Inserting Views with out-of-sequence ids > > > Sorry, lets see if I've understood you correctly, a typical example > might be .... > > MySQL creates id's 1-10 for new records > I add records specifically at 1001-1023 > MySQL then creates the next id as 1024 > > Is that what you mean? > > That unfortunately is not very good behavior because part of the > point of putting the added ids at 10001... > is to get them out of the way of the auto-created ids. So that if a > second import is done from the same slice they can safely overwrite > using the same algorithm - e.g. adding 10,000 > > Is this APC-AA behavior, or something built into MySQL? > > - Mitra > > > At 12:46 PM +0100 13/2/03, Marek Tichy wrote: > >Hi, > > > >> As part of importing / exporting, I've got a problem with shortids. > >> > >> Lets take views for example - they have shortids which are an > >> auto-increment field, > >> > >> When I import from one system to another, I want to add an offset to > >> avoid clashes, for example I might add 1000 so that I can a) do it in > >> emacs and b) understand the result and fix mistakes. > >> > >> Will this cause problems with the autoincrement field in any way? For > >> example there might be views 1-10 and 1001 -1024, so will the next > >> auto-incremented view by 11 ? And what happens when it > > > auto-increments to 1000? > > > >It will always start from 1024. > > > -- > Mitra Technology Consulting - www.mitra.biz - mi...@mi... > 02-6684-8096 or 0414-648-0722 > > Life is a Mystery to be Lived, not a Problem to be Solved > |
|
From: Mitra <mi...@ea...> - 2003-02-13 22:05:37
|
Yes - that sounds like the problem, yes - extsess is only included by locsess and extsessi by locsessi, but we have some cases where locsess is being included AS WELL as locsessi and other cases where neither are being included. I'd fix it, but i can't figure out what, whoever wrote this, intended. In particular when to include one and when the other. I think we need a single file to include "locsesseither.php3" which includes one and only one of the other files, but I don't understand when one should be included, and when the other. - Mitra At 2:18 PM +0100 13/2/03, Marek Tichy wrote: >Hi Mitra, >I'm getting an error in latest cvs, which i think is related to this. > > Call to a member function on a non-object in ><b>/raid/www/htdocs/aa.ecn.cz/apc-aa- > cvs/include/itemfunc.php3</b> on line <b>1172</b><br /> > >Which is: > relatedwindow =3D open( "'. $sess->url("related_sel.php3") ........... > >As far as I remember, extsess and extsessi are extensions to the >phplib session class. >I can't see any reason why would those two be included anywhere else >than locsess /locsessi, and also in locauth. > >Another weird problem I have is that slices don't show up - the >server returns blank document, as if it had crashed php. It happens >in slice.php3 at this spot: > >$zids=3DQueryZIDs($fields, $slice_id, $conds, $sort, $slice_info[group_by], > "ACTIVE", $slices, $neverAllItems, 0, >$defaultCondsOperator, true ); > >Don't think I can move this any forward. > >Best > Marek > > >> -----Original Message----- >> From: apc...@li... >> [mailto:apc...@li...]On Behalf Of Mitra >> Sent: 13. =FAnora 2003 4:46 >> To: apc...@so... >> Subject: [Apc-aa-coders] extsess and extsessi >> >> >> The change to require_once seems to have worked fairly well, except >> for one problem. >> >> There are some problems with extsess and extsessi >> >> If you include one, you can't include the other. >> >> But .... code has to include one or the other to function, >> >> What I'm finding is that certain combinations include either none or >> both and so break. >> >> Where SHOULD these be included from ? >> >> - Mitra >> >> -- >> Mitra Technology Consulting - www.mitra.biz - mi...@mi... >> 02-6684-8096 or 0414-648-0722 >> >> Life is a Mystery to be Lived, not a Problem to be Solved >> >> >> >> ------------------------------------------------------- >> This sf.net email is sponsored by:ThinkGeek >> Welcome to geek heaven. >> http://thinkgeek.com/sf >> _______________________________________________ >> Apc-aa-coders mailing list > > Apc...@li... > > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > > -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-13 21:45:35
|
I think Jakub forgot to mention (I no longer have the mail) that some of the changes he's made mean that a sql_update is required - some things break otherwise. - Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-13 21:27:43
|
Does anyone think that sliceimport and export are being used for storing data over a period. I'm wondering, because it would make life easier if I could break backward compatability - the code is in such poor condition, that being able to change the output format and corresponding input processing would be much simpler if I don't have to handle older versions. My assumption is that people are a) Moving data from site to site pretty much instantaneously, i.e. they export, process then import. b) that systems at both ends of the link can typically be assumed to run current versions of apc-aa. I'm only making these assumptions because the code appears to be mostly non-functional, so I can't imaging someone is using it for archiving / backup purposes. Any comments .... - Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Marek T. <ma...@gn...> - 2003-02-13 19:41:34
|
Hi Mitra, I'm getting an error in latest cvs, which i think is related to this. Call to a member function on a non-object in <b>/raid/www/htdocs/aa.ecn.cz/apc-aa- cvs/include/itemfunc.php3</b> on line <b>1172</b><br /> Which is: relatedwindow = open( "'. $sess->url("related_sel.php3") ........... As far as I remember, extsess and extsessi are extensions to the phplib session class. I can't see any reason why would those two be included anywhere else than locsess /locsessi, and also in locauth. Another weird problem I have is that slices don't show up - the server returns blank document, as if it had crashed php. It happens in slice.php3 at this spot: $zids=QueryZIDs($fields, $slice_id, $conds, $sort, $slice_info[group_by], "ACTIVE", $slices, $neverAllItems, 0, $defaultCondsOperator, true ); Don't think I can move this any forward. Best Marek > > -----Original Message----- > > From: apc...@li... > > [mailto:apc...@li...]On Behalf Of Mitra > > Sent: 13. února 2003 4:46 > > To: apc...@so... > > Subject: [Apc-aa-coders] extsess and extsessi > > > > > > The change to require_once seems to have worked fairly well, except > > for one problem. > > > > There are some problems with extsess and extsessi > > > > If you include one, you can't include the other. > > > > But .... code has to include one or the other to function, > > > > What I'm finding is that certain combinations include either none or > > both and so break. > > > > Where SHOULD these be included from ? > > > > - Mitra > > > > -- > > Mitra Technology Consulting - www.mitra.biz - mi...@mi... > > 02-6684-8096 or 0414-648-0722 > > > > Life is a Mystery to be Lived, not a Problem to be Solved > > > > > > > > ------------------------------------------------------- > > This sf.net email is sponsored by:ThinkGeek > > Welcome to geek heaven. > > http://thinkgeek.com/sf > > _______________________________________________ > > Apc-aa-coders mailing list > > Apc...@li... > > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > > |
|
From: <jak...@se...> - 2003-02-13 12:16:00
|
Mitra, Honza is the only author of the Site module. He is off this week. Jakub > -----Original Message----- > From: Mitra [mailto:mi...@ea...] > Sent: Thursday, February 13, 2003 12:13 PM > To: Jakub Adamek; apc...@so... > Subject: Re: [Apc-aa-coders] Auth feature and Docbook and pre-farewell > > > I'll be sorry to see you go, you have contributed much to APC-AA. > > One thing ... in case there is time before you go. I think you did > the Sites module ? (or was that Honza?) if so could you extend the > sliceexp and sliceimp to handle the Site modules. > > I've extended it to handle Views, and also to output XML as an > alternative to Hex/Zipped. I've found this essential since there are > many things that need changing by hand, and the right place to do > this is in the file before importing it. (Of course, it would be > ideal if the PHP did everything wanted, but that is never going to > work - there are for example too many places where a View id appears > inside another view, or in an alias. > > . I will check them into CVS when they are better tested - unless > anyone else wants to work on parts of it. > > - Mitra > > > At 6:39 PM +0100 11/2/03, Jakub Adamek wrote: > >Perhaps you should know that I will finish my job in Econnect in the > >middle of March and perhaps will not any more read the conference nor = do > >any work for AA. Use this knowledge to test all the features I am > >working on so that I can do the fixes needed before I finish. > > -- > Mitra Technology Consulting - www.mitra.biz - mi...@mi... > 02-6684-8096 or 0414-648-0722 > > Life is a Mystery to be Lived, not a Problem to be Solved > --- Odchoz=ED zpr=E1va neobsahuje viry. Zkontrolov=E1no antivirov=FDm syst=E9mem AVG (http://www.grisoft.cz). Verze: 6.0.449 / Virov=E1 b=E1ze: 251 - datum vyd=E1n=ED: 27.1.2003 |
|
From: Mitra <mi...@ea...> - 2003-02-13 11:22:42
|
I'll be sorry to see you go, you have contributed much to APC-AA. One thing ... in case there is time before you go. I think you did the Sites module ? (or was that Honza?) if so could you extend the sliceexp and sliceimp to handle the Site modules. I've extended it to handle Views, and also to output XML as an alternative to Hex/Zipped. I've found this essential since there are many things that need changing by hand, and the right place to do this is in the file before importing it. (Of course, it would be ideal if the PHP did everything wanted, but that is never going to work - there are for example too many places where a View id appears inside another view, or in an alias. . I will check them into CVS when they are better tested - unless anyone else wants to work on parts of it. - Mitra At 6:39 PM +0100 11/2/03, Jakub Adamek wrote: >Perhaps you should know that I will finish my job in Econnect in the >middle of March and perhaps will not any more read the conference nor do >any work for AA. Use this knowledge to test all the features I am >working on so that I can do the fixes needed before I finish. -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-13 11:22:20
|
As part of importing / exporting, I've got a problem with shortids. Lets take views for example - they have shortids which are an auto-increment field, When I import from one system to another, I want to add an offset to avoid clashes, for example I might add 1000 so that I can a) do it in emacs and b) understand the result and fix mistakes. Will this cause problems with the autoincrement field in any way? For example there might be views 1-10 and 1001 -1024, so will the next auto-incremented view by 11 ? And what happens when it auto-increments to 1000? I could probably write a bunch of tests for this, but wondered if anyone already had done so? - Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-13 03:51:03
|
The change to require_once seems to have worked fairly well, except for one problem. There are some problems with extsess and extsessi If you include one, you can't include the other. But .... code has to include one or the other to function, What I'm finding is that certain combinations include either none or both and so break. Where SHOULD these be included from ? - Mitra -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-13 03:14:14
|
I think if it is not possible to merge them, then what's needed is almost the opposite of what you suggest. I.e. there are typically more Readers than AA users, so when a user's password is changed in AA, then they should be automatically added to the Readers slice so that HTTP authentication can find it? Ram and I were doing something different, he had HTTP authentication working off the normal AA permissions - and I was adding a "READ" permission, my part never got finished, and Ram's wasn't using Groups yet, so probably what you have will be better - especially since non-superadmins can manage it. - Mitra At 6:39 PM +0100 11/2/03, Jakub Adamek wrote: >Norbert, Mitra and all, > >we were discussing in Econnect the unhappy situation of two user >management systems and it does not seem possible to merge them. The AA >permission system has many features which would be really hard to model >by a slice. On the other hand the slice approach has such features which >can't be modelled by the AA system. > >Our proposal is to add another Reader Management module, which will >synchronize the AA permissions from the slice. It would be only one-way >synchronization, i.e. users are added to AA permissions on addition to >the slice and their password and email info changed in AA pemissions >when changed in the slice. > >Please do think about it and write any ideas you have. It was a hard way >to come to the decision to put the Readers into a slice, perhaps it >would be possible to add AA permissions there also. But at the moment I >do not much believe in it. > >Jakub > >> -----Original Message----- >> From: apc...@li... >> [mailto:apc...@li...] On Behalf >> Of Norbert Brazda >> Sent: Sunday, February 09, 2003 1:38 PM >> To: apc...@so... >> Subject: Re: [Apc-aa-coders] Reader Management Slice >> >> >> From: "Mitra" <mi...@ea...> >> > One concern .... a common situation I have is where the user's have >> > ability to edit some parts of a site, and controlled reader >> access to >> > those and other parts of a site. For example this is the case where >> > you have a the board of an organization wanting a shared space to >> > post news articles. I'd like NOT to have to give these people two >> > different sets of login ids, and manage them in two >> different places. >> > Can we integrate the existing user management process for >> > apc-aa/admin with this? So that one interface can be used to add >> > users to groups for managing parts of other slices.. > > > > So why not to put all user (not only reader) related > > information into the slice? > > > > And I would like to suggest one more change if you are going > > to do it: User have no longer access to the anonymous editing >> when is the item changed by administrator >> (ITEM_FLAG_ANONYMOUS_EDITABLE is reset through admin >> interface). It would be good to have this as a option. In >> come cases we want users to change thier data even after they >> were changed by the adminstrator (e.g. directory of NGOs - >> sometime NGOs update data by themseleves, sometimes it is >> done by our editors). >> >> >> >> >> >> Norbert Brazda >> br...@ch... >> tel: 0905-729359 >> >> ChangeNet - informacny servis o obcianskej spolocnosti Mlynske nivy >> 41, 821 09 Bratislava, Slovakia >> tel/fax: 02-55560026, eMail: in...@ch... http://www.changenet.sk >> >> SPAJAME LUDI, KTORI MENIA SVET >> >> >> >> >> ------------------------------------------------------- >> This SF.NET email is sponsored by: >> SourceForge Enterprise Edition + IBM + LinuxWorld >> http://www.vasoftware.com >> _______________________________________________ >> Apc-aa-coders mailing list >> Apc...@li... >> https://lists.sourceforge.net/lists/listinfo/apc-aa-coders >> > > > >------------------------------------------------------- >This SF.NET email is sponsored by: >SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! >http://www.vasoftware.com >_______________________________________________ >Apc-aa-coders mailing list >Apc...@li... >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Jakub A. <jak...@se...> - 2003-02-11 17:40:09
|
Norbert, Mitra and all, we were discussing in Econnect the unhappy situation of two user management systems and it does not seem possible to merge them. The AA permission system has many features which would be really hard to model by a slice. On the other hand the slice approach has such features which can't be modelled by the AA system. Our proposal is to add another Reader Management module, which will synchronize the AA permissions from the slice. It would be only one-way synchronization, i.e. users are added to AA permissions on addition to the slice and their password and email info changed in AA pemissions when changed in the slice. Please do think about it and write any ideas you have. It was a hard way to come to the decision to put the Readers into a slice, perhaps it would be possible to add AA permissions there also. But at the moment I do not much believe in it. Jakub > -----Original Message----- > From: apc...@li... > [mailto:apc...@li...] On Behalf > Of Norbert Brazda > Sent: Sunday, February 09, 2003 1:38 PM > To: apc...@so... > Subject: Re: [Apc-aa-coders] Reader Management Slice > > > From: "Mitra" <mi...@ea...> > > One concern .... a common situation I have is where the user's have > > ability to edit some parts of a site, and controlled reader > access to > > those and other parts of a site. For example this is the case where > > you have a the board of an organization wanting a shared space to > > post news articles. I'd like NOT to have to give these people two > > different sets of login ids, and manage them in two > different places. > > Can we integrate the existing user management process for > > apc-aa/admin with this? So that one interface can be used to add > > users to groups for managing parts of other slices.. > > So why not to put all user (not only reader) related > information into the slice? > > And I would like to suggest one more change if you are going > to do it: User have no longer access to the anonymous editing > when is the item changed by administrator > (ITEM_FLAG_ANONYMOUS_EDITABLE is reset through admin > interface). It would be good to have this as a option. In > come cases we want users to change thier data even after they > were changed by the adminstrator (e.g. directory of NGOs - > sometime NGOs update data by themseleves, sometimes it is > done by our editors). > > > > > > Norbert Brazda > br...@ch... > tel: 0905-729359 > > ChangeNet - informacny servis o obcianskej spolocnosti Mlynske nivy > 41, 821 09 Bratislava, Slovakia > tel/fax: 02-55560026, eMail: in...@ch... http://www.changenet.sk > > SPAJAME LUDI, KTORI MENIA SVET > > > > > ------------------------------------------------------- > This SF.NET email is sponsored by: > SourceForge Enterprise Edition + IBM + LinuxWorld > http://www.vasoftware.com > _______________________________________________ > Apc-aa-coders mailing list > Apc...@li... > https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > |
|
From: Jakub A. <jak...@se...> - 2003-02-11 17:40:06
|
Hi all, I have finished the first "module" (it is not a real module) for the Reader Management Slice: the Auth feature. See the description in the attached file. This file is again the HTML doc/reader.html got from the Docbook file doc/reader.xml. This should be also an advertisement of the Docbook features. See the nice revision info and content at the beginning of the file. Both are created automagically. I would like to recommend Docbook to you for next version of all AA docs. There is also a lovely and powerful free Docbook (but not only Docbook) editor XmlMind, see www.xmlmind.org (if I am not mistaken) which makes creating Docbook documents a pleasure. Perhaps you should know that I will finish my job in Econnect in the middle of March and perhaps will not any more read the conference nor do any work for AA. Use this knowledge to test all the features I am working on so that I can do the fixes needed before I finish. Jakub |
|
From: Mitra <mi...@ea...> - 2003-02-10 08:30:08
|
I lost track of whose edits ended up in CVS, but I still see problems in sites with actions being duplicated. - Mitra At 1:44 PM +1100 11/1/03, Mitra wrote: >The code sending spot_id seemed to work,it wasn't being reset from >the session variable, because I no longer registered r_spot_id with >the session, but your way also works. > >- Mitra > > >At 12:45 AM +0100 11/1/03, Honza Malik wrote: >>Hi Mitra, >> >> I'm the coder of site module so all the the laments let go on my >>head. I'm agree it was not happy decision to have spot_id in session, >>but it is quite easy to fix. I will do it, when I will find a time. >> >> As I see you tried to fix the problem by sending 'spot id' whenever >>you send data to the server, which is good solution, but probably not >>working if you are sending the r_spot_id variable. r_spot_id variable is >>(like any other r_* variables) session variable, so if you send it by >>GET or POST, it is reset in init_page.php3 from session. >> >> I fixed the code (to send site_id instead). It is in CVS. I test it >>just for a while, but it looks it works. Please test it, if you can. >> >> Regards, >> Honza >> >> >>On =DAt, 2003-01-07 at 01:11, Mitra wrote: >>> There is definately a nasty bug in the site editor, its doing things >>> like overwriting the wrong Spot during an edit etc. >>> >>> I'm suspecting its got something to do with the session variable which >>> stores the spot id, although I'm not doing any of the obviously bad >>> things to do when someone is using session variables (like using Back >>> or Forward). >>> >>> <FLAME> >>> What I fail to understand is why whoever coded this (it isn't signed) >>> used a site variable in the first place, this information could EASILY >>> have been incorporated into the URL without the inherent bugs of >>> session variables. >>> >>> Just to be clear - in case you don't know why session variables are a >>> BAD idea, >>> If I do something to change the spot I'm looking at, and then hit >>> Back, then my browser thinks I'm on one spot, while the server thinks >>> I'm on another one. >>> >>> Its fine to use Session Variables to cache stuff like authentication >>> but NEVER use them for maintaining state that changes between screens, >>> you'll get caught time and time again because the web was never >>> designed to work with a one-way flow. (Believe me, I was at the >> > meetings where this was discussed, and a primary design goal of the >> > web - unlike the existing database systems around at the time - was to >> > make it stateless, and therefore able to branch, go back and forth >>> without talking to the server, and be bookmarkable) >>> >>> I'd love whoever wrote this code to say something ..... before I go >>> and poke around in the code, at this point my temptation is to just >>> rip out the session variable code from this module ! >>> >>> </FLAME> >>> - Mitra >>> >>> >>> -- >>> Mitra Technology Consulting - www.mitra.biz - mi...@mi... >>> 02-6684-8096 or 0414-648-0722 >>> >>> Life is a Mystery to be Lived, not a Problem to be Solved >>> >> >> >> >>------------------------------------------------------- >>This SF.NET email is sponsored by: >>SourceForge Enterprise Edition + IBM + LinuxWorld http://www.vasoftware.co= m >>_______________________________________________ >>Apc-aa-coders mailing list >>Apc...@li... >>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders > > >-- >Mitra Technology Consulting - www.mitra.biz - mi...@mi... >02-6684-8096 or 0414-648-0722 >Life is a Mystery to be Lived, not a Problem to be Solved > > > >------------------------------------------------------- >This SF.NET email is sponsored by: >SourceForge Enterprise Edition + IBM + LinuxWorld http://www.vasoftware.com >_______________________________________________ >Apc-aa-coders mailing list >Apc...@li... >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-10 08:18:43
|
I didn't hear anything back on this one, so noone has been working on that code recently? Must be code-rot ! - Mitra At 5:58 PM +1100 3/2/03, Mitra wrote: >Group Permissions seem to be broken, > >Assign New Permissions has a list of slices, but when you select one >and hit Update the page redraws with no changes. > >Anyone been working on this recently, want to take a look at what is wrong? > >- Mitra > >-- >Mitra Technology Consulting - www.mitra.biz - mi...@mi... >02-6684-8096 or 0414-648-0722 > >Life is a Mystery to be Lived, not a Problem to be Solved -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
|
From: Mitra <mi...@ea...> - 2003-02-10 02:08:46
|
Yes - that would be my prefered solution - to put all user information into the slice. Jakub - can you do this? - Mitra At 1:38 PM +0100 9/2/03, Norbert Brazda wrote: >From: "Mitra" <mi...@ea...> >> One concern .... a common situation I have is where the user's have >> ability to edit some parts of a site, and controlled reader access to >> those and other parts of a site. For example this is the case where >> you have a the board of an organization wanting a shared space to >> post news articles. I'd like NOT to have to give these people two >> different sets of login ids, and manage them in two different places. >> Can we integrate the existing user management process for >> apc-aa/admin with this? So that one interface can be used to add >> users to groups for managing parts of other slices.. > >So why not to put all user (not only reader) related information >into the slice? > >And I would like to suggest one more change if you are going to do >it: User have no longer access to the anonymous editing when is the >item changed by administrator (ITEM_FLAG_ANONYMOUS_EDITABLE is reset >through admin interface). It would be good to have this as a option. >In come cases we want users to change thier data even after they >were changed by the adminstrator (e.g. directory of NGOs - sometime >NGOs update data by themseleves, sometimes it is done by our >editors). > > > > > >Norbert Brazda >br...@ch... >tel: 0905-729359 > >ChangeNet - informacny servis o obcianskej spolocnosti >Mlynske nivy 41, 821 09 Bratislava, Slovakia >tel/fax: 02-55560026, eMail: in...@ch... >http://www.changenet.sk > >SPAJAME LUDI, KTORI MENIA SVET > > > > >------------------------------------------------------- >This SF.NET email is sponsored by: >SourceForge Enterprise Edition + IBM + LinuxWorld http://www.vasoftware.com >_______________________________________________ >Apc-aa-coders mailing list >Apc...@li... >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders -- Mitra Technology Consulting - www.mitra.biz - mi...@mi... 02-6684-8096 or 0414-648-0722 Life is a Mystery to be Lived, not a Problem to be Solved |
3 messages has been excluded from this view by a project administrator.
