You can subscribe to this list here.
| 2001 | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct (35) | Nov (38) | Dec (112) | 
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 | Jan (20) | Feb (24) | Mar (47) | Apr (18) | May (28) | Jun (17) | Jul (15) | Aug (40) | Sep (14) | Oct (5) | Nov (26) | Dec (31) | 
| 2003 | Jan (8) | Feb (14) | Mar (38) | Apr (34) | May (33) | Jun (32) | Jul (24) | Aug (9) | Sep | Oct (20) | Nov (43) | Dec (22) | 
| 2004 | Jan (23) | Feb (25) | Mar (15) | Apr (3) | May (31) | Jun (13) | Jul (3) | Aug (3) | Sep (13) | Oct (15) | Nov (3) | Dec (5) | 
| 2005 | Jan | Feb | Mar (16) | Apr (24) | May | Jun (2) | Jul | Aug (5) | Sep (4) | Oct | Nov (3) | Dec (2) | 
| 2006 | Jan | Feb (1) | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | 
| 
      
      
      From: <Tho...@be...> - 2001-12-07 15:07:06
      
     | 
| Hello.
Is it possible to modify the search criteria from exact to near in
dependency of the security level?
We tried like this :
sub MY_SEARCH_FIELDS           { return qw( field1 field2 ) }
sub MY_SEARCH_FIELDS_EXACT      { 
				 my ( $p ) = @_;
				if ($p->{tmp_security_level} <
SEC_LEVEL_WRITE) {
				return qw( field1 field2 )';
				}else{ 
				return ();
				}
				
}
If the security level is 'write' the user should be able to search without
exact phrases.
How do we manage this ?
Greets.
Thomas
 | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-07 12:58:50
      
     | 
| * Tho...@be... (Tho...@be...) [011207 07:55]: > Is it possible to make the search_form page from the Classified > package secure ? If I edit the security of the object and the > handler nothing happens. > > Everyone can see the search_form :-( Sure. In OpenInteract/Handler/Classified.pm you currently have: %OpenInteract::Handler::Classified::security = ( DEFAULT_SECURITY_KEY() => SEC_LEVEL_NONE, create => SEC_LEVEL_READ, edit => SEC_LEVEL_READ, remove => SEC_LEVEL_WRITE, ); Set it to: %OpenInteract::Handler::Classified::security = ( search_form => SEC_LEVEL_READ, DEFAULT_SECURITY_KEY() => SEC_LEVEL_NONE, create => SEC_LEVEL_READ, edit => SEC_LEVEL_READ, remove => SEC_LEVEL_WRITE, ); OR, if you want to ensure that people have access to execute any operation: %OpenInteract::Handler::Classified::security = ( DEFAULT_SECURITY_KEY() => SEC_LEVEL_READ, remove => SEC_LEVEL_WRITE, ); Re-deploy this and restart the server, and you should be good. Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: <Tho...@be...> - 2001-12-07 12:45:26
      
     | 
| Hey there. Is it possible to make the search_form page from the Classified package secure ? If I edit the security of the object and the handler nothing happens. Everyone can see the search_form :-( Greets. Thomas | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-07 12:38:05
      
     | 
| * Victor Piterbarg (ope...@ha...) [011207 02:32]: > I've been trying to run "oi_manage check_package" on a package that I felt > like testing. However, I was greeted with the unpleasantess of the > Segmentation Fault. I tracked it down to the check() function in > OpenInteract::Package where it checks the files under conf, and it was > choking on spops.perl. It seems that the reson for this, was some creative > syntax on my part: > track => 'create => 1, update => 1, remove => 1}, > > However, why a segfault? After I fixed the prob., I got a few other nice > messages about syntax errors from eval(), but nothing caused a segfault. I > just though it was peculiar. I know -- this happens every once in a while and I haven't looked into it enough to explain it. I just put it in the FAQ. Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-07 12:34:17
      
     | 
| * Victor Piterbarg (ope...@ha...) [011207 00:44]:
> I am creating a fairly simple object, which is a user profile. The
> user profile contains a country. I've created a lookup table for all
> the countries using the lookup package. Now, I am creating a
> template inside which I would like to have a dropdown with all the
> countries listed. I am using the OpenInteract::CommonHandler object
> to manage the user profiles, so I am assuming I need to retrieve the
> list of countries somewhere in _show_customize(). Is there a
> pleasant way to do this, or do I need to just run a query on the
> country lookup table? Thanks.
You could do something like (assuming your object is named 'country',
the key is 'abbreviation' and the name is 'name' ):
sub _show_customize {
    my ( $class, $params ) = @_;
    my $R = OpenInteract::Request->instance;
    $params->{country_list} = $R->country->fetch_group({
                                     order => 'name' });
}
And then in your template:
[% INCLUDE label_form_select_row( 
      label       = 'Country', list = country_list,
      value_field = 'abbreviation',
      label_field = 'name',
      picked      = user.country ) %]
Make sense? Like I mentioned yesterday, I'm going to try to update the
tutorial to have this new stuff in there. That should help.
Chris
-- 
Chris Winters (ch...@cw...)
Building enterprise-capable snack solutions since 1988.
 | 
| 
      
      
      From: Victor P. <ope...@ha...> - 2001-12-07 07:22:41
      
     | 
| Hello,
I've been trying to run "oi_manage check_package" on a package that I felt
like testing. However, I was greeted with the unpleasantess of the
Segmentation Fault. I tracked it down to the check() function in
OpenInteract::Package where it checks the files under conf, and it was
choking on spops.perl. It seems that the reson for this, was some creative
syntax on my part:
      track       => 'create => 1, update => 1, remove => 1},
However, why a segfault? After I fixed the prob., I got a few other nice
messages about syntax errors from eval(), but nothing caused a segfault. I
just though it was peculiar.
-Victor
---
"Tatarsky, of course,  hated most of the manifestations of  Soviet power,
but he still couldn't understand why  it was worth exchanging an evil empire
for an evil banana republic" --Viktor Pelevin, Generation P.
 | 
| 
      
      
      From: Victor P. <ope...@ha...> - 2001-12-07 05:34:43
      
     | 
| Hello, I am creating a fairly simple object, which is a user profile. The user profile contains a country. I've created a lookup table for all the countries using the lookup package. Now, I am creating a template inside which I would like to have a dropdown with all the countries listed. I am using the OpenInteract::CommonHandler object to manage the user profiles, so I am assuming I need to retrieve the list of countries somewhere in _show_customize(). Is there a pleasant way to do this, or do I need to just run a query on the country lookup table? Thanks. -Victor ---- "Tatarsky, of course, hated most of the manifestations of Soviet power, but he still couldn't understand why it was worth exchanging an evil empire for an evil banana republic" --Viktor Pelevin, Generation P. | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-06 21:08:44
      
     | 
| * John Sequeira (js...@me...) [011206 15:10]: > I'd like to clarify how a package development cycle should go. I > followed through the developer's guide with the fruit example, and > read through some of the docs on packaging and am still a little > confused. These can do with an update, and I'm hoping to do so relatively soon (~within 3-6 weeks) with a not-so-trivial example. > After I've installed a package version in OI, and applied it to a > website, where does the development occur? In other words, which > files should I be messing with? > > 1) in the initial folder you used to create the package? This makes > sense, but then a small edit implies an install/apply cycle vi > oimanager, right? Correct. > 2) In the OI /pkgs folder ? which means you have to apply it before > testing, and restart apache or use StatINC? Correct, but that is definitely not the place to make updates. > 3) In the applied package files /your/web/site/pkg/Fruit2 These are > the files apache is actually executing when you look at a URL, so > this makes sense, but then you have to patch up your original source > tree when creating an official version, right? Right. The best way I've found is to do (1) until you reach a point where you're modifying very samll things (particularly templates), then do (3) until you get it looking right. (With many server restarts along the way.) Then manually port all the changes back to (1). Since I generally only do (3) with templates, manually porting the changes is just copying the templates over to the -dev directory. > Although I can probably figure out a way to do this, it seems that > the best way eludes me. The *best* way would probably be to do (3) until you reach a milestone, and then run a routine to generate the original package code. There's actually a feature request on Sourceforge to do this [1], but it's a relatively low priority. I've been thinking about other ways to go about this, but they all require a mod_perl server per website. This might not be such a bad thing, however, and would fit better with plans to run under different environments. > p.s. Thanks for the Win32 patch - it worked like a charm. Excellent! Chris [1] http://sourceforge.net/tracker/index.php?func=detail&aid=470097&group_id=16810&atid=366810 -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: John S. <js...@me...> - 2001-12-06 20:00:16
      
     | 
| I'd like to clarify how a package development cycle should go. I followed through the developer's guide with the fruit example, and read through some of the docs on packaging and am still a little confused. After I've installed a package version in OI, and applied it to a website, where does the development occur? In other words, which files should I be messing with? 1) in the initial folder you used to create the package? This makes sense, but then a small edit implies an install/apply cycle vi oimanager, right? 2) In the OI /pkgs folder ? which means you have to apply it before testing, and restart apache or use StatINC? 3) In the applied package files /your/web/site/pkg/Fruit2 These are the files apache is actually executing when you look at a URL, so this makes sense, but then you have to patch up your original source tree when creating an official version, right? Although I can probably figure out a way to do this, it seems that the best way eludes me. JS p.s. Thanks for the Win32 patch - it worked like a charm. | 
| 
      
      
      From: <And...@Be...> - 2001-12-06 12:58:12
      
     | 
| .. hard to tell on a test site - it seems, it does. But we shall see = on the production site... BTW: recently I posted our idea of using sys_security with = skip_security as a join to boost performance: this works great and speeds things up = noticebly with our critical requests, that is those, where the actual request = returns a lot of records, of which most are kicked out by the security = mechanism ( with one fetch for each record ! ). later, Andreas -----Urspr=FCngliche Nachricht----- Von: Chris Winters [mailto:ch...@cw...] Gesendet am: Donnerstag, 6. Dezember 2001 14:11 An: Nolte, Andreas, D22K-MSN Cc: ope...@li... Betreff: Re: [Openinteract-help] OI with GEMINI * Andreas Nolte (And...@Be...) [011206 04:01]: > so far, innodb tables seem to work right. Catch: no fulltext index at = the > moment. I did not get around to testing gemini. I`ll keep this list = up to > date. Good news! How has this affected the performance issues? Chris --=20 Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. _______________________________________________ openinteract-help mailing list ope...@li... https://lists.sourceforge.net/lists/listinfo/openinteract-help | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-06 12:50:14
      
     | 
| * Andreas Nolte (And...@Be...) [011206 04:01]: > so far, innodb tables seem to work right. Catch: no fulltext index at the > moment. I did not get around to testing gemini. I`ll keep this list up to > date. Good news! How has this affected the performance issues? Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-06 12:46:22
      
     | 
| * Tho...@be... (Tho...@be...) [011206 05:40]: > I've problems with the uploadfunction in StaticPage 1.51. Is it possible to > upload a html document without any entry in the "Document"-field, so that > the content of this html-document is displayed as new page ? It should be possible -- that's the purpose. Can the user that Apache runs under write files and create directories in the /html directory tree? Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: <Tho...@be...> - 2001-12-06 10:30:32
      
     | 
| Hello. I've problems with the uploadfunction in StaticPage 1.51. Is it possible to upload a html document without any entry in the "Document"-field, so that the content of this html-document is displayed as new page ? Greets Thomas | 
| 
      
      
      From: <And...@Be...> - 2001-12-06 09:22:55
      
     | 
| Yes, we did. It is called NATS (Network activity tracking system )but = right now is tweaked for LDAP usage. It should work with standard OI, but we = did not test it yet and there ist very little docu. The system is multi client with 3 roles: Users, which can pu in orders = ( CREATOR role ), the working folks ( PROCESSOR role ) and admins ( ADMIN = role ). A combination of these 3 roles is called a client. You can have = multiple groups playing different roles in different client setups, e.g. one IT support processor group with multiple creator groups as multiple = clients. You need to define order numbers for each client and define what we = call "actions" for each order number, so you can determine, what may be done = for which order number. The processors then can book their working hours to = some order number. The general NATS approach is implemented to for a NATS ticket system. Is this interesting for y`all? You would probably need to do some = tweaking to make it usable in a general way ( that`s why we did not post it = here, so far), but the basis works fine.=20 later, Andreas -----Urspr=FCngliche Nachricht----- Von: Bryce Harrington [mailto:br...@ne...] Gesendet am: Dienstag, 4. Dezember 2001 00:45 An: ope...@li... Betreff: [Openinteract-help] OI Task / Request manager? Has anyone created a package for doing task management? One thought I've been dinging around with is integrating task / request / bug tracking in a sort of integrated manner with particular packages. So for example, if you had a setup where you had 'projects', 'people', = and 'hardware', you could associate tasks that involve projects (putting = out product releases, doing testing runs, etc.), or people (setting up a person's account, assigning him resources, etc.), or hardware (installing a new memory card, figuring out why it crashes on full moons). Anyway, has anyone already done something like this (or even vaguely close)? Background: We've got a collection of technical and non-technical users, with pretty disparate needs and desires. We've tried Bugzilla and Request Tracker, and have considered a few others, but they were = all either too complicated to use (for the non-technical) or too hard to customize and extend (for the technical), and I'd like to give a shot = at seeing if something could be done with OpenInteract that might provide the optimum solution. Bryce _______________________________________________ openinteract-help mailing list ope...@li... https://lists.sourceforge.net/lists/listinfo/openinteract-help | 
| 
      
      
      From: <And...@Be...> - 2001-12-06 08:52:21
      
     | 
| Hey, y`all, so far, innodb tables seem to work right. Catch: no fulltext index at = the moment. I did not get around to testing gemini. I`ll keep this list up = to date. later, Andreas -----Urspr=FCngliche Nachricht----- Von: Chris Winters [mailto:ch...@cw...] Gesendet am: Montag, 3. Dezember 2001 04:24 An: Nolte, Andreas, D22K-MSN Cc: ope...@li... Betreff: Re: [Openinteract-help] OI with GEMINI * Andreas Nolte (And...@Be...) [011201 16:34]: > has anyone of you used OI with mysql and the GEMINI table type? We > have some concurrency problems, since we have a lot of updates and > inserts with auto_increment columns. The docs of mysql say, that > MyISAM talbes ( the default ) lock the whole table, before an > auto_increment insert can be made - this hurts performance badly in > our environment, where multiple users do updates / inserts at the > same time. I'd never heard this before, that it locks the whole table. Sheesh. Another alternative is dumping auto-incrementing fields entirely and using something else -- it probably wouldn't be too difficult to create a socket server (in Perl even!) to generate unique IDs. But this might have bigger problems that using the Gemini/InnoDB tables -- another point of failure, etc. Good luck Chris --=20 Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: Crisha W. <re...@tr...> - 2001-12-05 04:04:16
      
     | 
| <HTML><HEAD><TITLE></TITLE>
<STYLE type=text/css>TD {
	FONT-SIZE: 12px
}
P {
	FONT-SIZE: 12px
}
LI {
	FONT-SIZE: 12px
}
INPUT {
	FONT-SIZE: 12px
}
SELECT {
	FONT-SIZE: 12px
}
</STYLE>
<META http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<META content="MSHTML 5.00.2920.0" name=GENERATOR></HEAD>
<BODY bgColor=#cccccc leftMargin=0 text=#000000 topMargin=0 marginheight="0" 
marginwidth="0">
<TABLE bgColor=#ffffff border=0 cellPadding=0 cellSpacing=0 width=600>
  <TBODY>
  <TR>
    <TD height=55 width=113><IMG height=140 
      src="http://image.trafficbbs.com/meiyuan.gif" width=100></TD>
    <TD height=55>
      <TABLE border=0 cellPadding=0 cellSpacing=0 width="100%">
        <TBODY>
        <TR>
          <TD><B><FONT face="Arial, Helvetica, sans-serif" size=2><IMG 
            height=18 src="http://image.trafficbbs.com/type_d.gif" width=23>o you 
            want to get maximum exposure for your website? </FONT></B></TD></TR>
        <TR>
          <TD><B><FONT face="Arial, Helvetica, sans-serif" size=2><IMG 
            height=18 src="http://image.trafficbbs.com/type_a.gif" width=26>re you 
            trying to introduce or sell your new product? </FONT></B></TD></TR>
        <TR>
          <TD><B><FONT face="Arial, Helvetica, sans-serif" size=2><IMG 
            height=18 src="http://image.trafficbbs.com/type_a.gif" width=26>re you 
            planning to present your new service or technology?</FONT></B></TD></TR>
        <TR>
          <TD><B><FONT face="Arial, Helvetica, sans-serif" size=2><IMG 
            height=18 src="http://image.trafficbbs.com/type_d.gif" width=23>o you 
            want to learn instant info about new service?</FONT></B></TD></TR>
        <TR>
          <TD><B><FONT face="Arial, Helvetica, sans-serif" size=2><IMG 
            height=18 src="http://image.trafficbbs.com/type_h.gif" width=32>ave you 
            got enough time, energy and cost to spread your idea? 
        </FONT></B></TD></TR></TBODY></TABLE></TD></TR></TBODY></TABLE>
<TABLE bgColor=#514fa3 border=0 cellPadding=0 cellSpacing=0 width=600>
  <TBODY>
  <TR>
    <TD vAlign=top width=21> </TD>
    <TD height=227 width=303>
      <TABLE border=0 cellPadding=0 cellSpacing=0 width="100%">
        <TBODY>
        <TR>
          <TD><FONT color=#ffffff face="Verdana, Arial, Helvetica, sans-serif" 
            size=2><BR>Hello,<BR><BR>You may have spent much on lots of ways to 
            achieve these - search engine registrations, website promotions, 
            press release, email sending¡ Here Traffic BBS presents you a unique 
            method economically and professionally converting a PC into personal 
            message distribution center! Traffic BBS assists you to post your 
            message or ad to over <B>12,000+</B> message boards on the web 
            worldwide. Along with a hyperlink to your website or email address, 
            a message of your business, product, service or offer will be 
            promptly submitted to targeted bulletin boards. You can expect 
            instant response! <!--      For more details please go to <a href=http://www.trafficbbs.com target=_blank><font color=#ffffff>Trafficbbs.com</a>.--></FONT>
            <P><FONT color=#ffffff face="Verdana, Arial, Helvetica, sans-serif" 
            size=2>Get your business, service, product or offer seen! 
</FONT></P>
            <P><FONT color=#ffffff face="Verdana, Arial, Helvetica, sans-serif" 
            size=2>Best Regards,<BR>Crisha Wenston<BR>Sales & Marketing 
            <BR><A href="http://www.trafficbbs.com/" target=_blank><FONT 
            color=#ffffff>www.trafficbbs.com</A> 
      </FONT></P></FONT></TD></TR></TBODY></TABLE></TD>
    <TD vAlign=top>
      <TABLE border=0 cellPadding=0 cellSpacing=0>
        <TBODY>
        <TR>
          <TD align=right background="http://image.trafficbbs.com/image932/7/80/img4.jpg" height=220 vAlign=top width=275>
          <A href='http://www.openinteract.org' target=_blank><IMG height=227 border=0 src="http://image.trafficbbs.com/image_kuang.gif" width=275></A></TD></TR>
        <TR height="100%" width="100%">
          <TD align=middle vAlign=center><BR><BR><BR><BR>
            <FORM action=http://www.trafficbbs.com method=post target=_blank><INPUT type=submit value="More Infomation ..."></FORM></TD></TR></TBODY></TABLE></TD></TR></TBODY></TABLE>
<TABLE bgColor=#514fa3 border=0 cellPadding=0 cellSpacing=0 width=600>
  <TBODY>
  <TR>
    <TD vAlign=top width=21> </TD>
    <TD vAlign=top><FONT color=#ffffff 
      face="Verdana, Arial, Helvetica, sans-serif" size=2>
      <P><BR></P></FONT></TD></TR></TBODY></TABLE>
<TABLE bgColor=#ffffff border=0 cellPadding=0 cellSpacing=0 width=600>
  <TBODY>
  <TR>
    <TD height=54 vAlign=top width=21> </TD>
    <TD height=54 vAlign=bottom width=222><FONT color=#ffffff size=2><IMG 
      height=41 src="http://image.trafficbbs.com/logo.gif" width=225> </FONT></TD>
    <TD height=54 vAlign=bottom width=357> </TD></TR>
  <TR>
    <TD height=40 vAlign=top width=21> </TD>
    <TD align=right height=40> </TD>
    <TD height=40><FONT color=#000000 
      face="Verdana, Arial, Helvetica, sans-serif" size=1>Copy right©2001 , 
      TrafficBBS™All Rights Reserved.<BR>TrafficBBS™is a trademark of 
      TrafficBBS.COM inc.</FONT></TD></TR></TBODY></TABLE></BODY></HTML>
 | 
| 
      
      
      From: Bryce H. <br...@ne...> - 2001-12-05 00:15:28
      
     | 
| On Tue, 4 Dec 2001, Chris Winters wrote: > * Bryce Harrington (br...@ne...) [011203 18:55]: > > Has anyone created a package for doing task management? > > I've created something for personal use, more of a time tracking > thing. But it was cut short of finished and I haven't picked it back > up again. > > I know a fairly substantial user has implemented a trouble ticket sort > of system, but that's for internal use. > > One thing to look at is perhaps wrapping around the RT package. RT is > being used for bug tracking on perl modules (see http://rt.cpan.org/ > for that, and http://www.fsck.com/projects/rt/index.html on RT in > general). I haven't used it, but it seems pretty robust -- hey, it > uses Template Toolkit! I'm not sure a wrapper would even be > appropriate -- maybe for user information/authentication integration, > plus fitting its info into your overall site design. Yeah, that's exactly the direction I've been going, and that's what we've been using so far. I've dug into the RT database tables and explored how the system works. One thing that it's missing is the notion of hierarchical project grouping. I.e., if we have a product, with several teams working on subprojects within it, and those teams want to divide up their work into sub-subprojects, we'd like to be able to represent that. With RT, though, AFAIK it has to be done via keywords, which works but is not quite what folks were after. Also, the notion of being able to "tie" tasks to particular other parts of the database (so you can get a listing of tasks associated with a particular piece of hardware, alongside other information about the hardware) does not seem evident to do... Anyway, in digging into it, it really looks like the design is just a single ticket table with what we need plus a whole bunch of admin tables with stuff that isn't quite what we want, so I've been kicking around ideas for alternative ways of doing this. So who knows, maybe I'll prototype something out using OpenInteract some day, and see how it goes. :-) Bryce | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-04 23:07:34
      
     | 
| * John Sequeira (js...@me...) [011204 17:59]: > I'm having trouble getting OI installed on window using the latest > tarball. > > When I try to visit the first page in the browser, OI complains that > it can't obtain the configuration information from my Stash class. > In the Stash class sourcecode, %ITEMS is defined as an empty hash. > I must have skipped the step where this gets filled in (unless it's > at startup time?). This is an issue with initialization -- I borked the Win32 initialization. A patch for OpenInteract/ApacheStartup.pm is attached -- it was a dumb mistake and easily fixed, so it should work ok even though I'm not near a Win32 machine. This update is in CVS as well. Thanks for reporting this, Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: John S. <js...@me...> - 2001-12-04 22:49:33
      
     | 
| I'm having trouble getting OI installed on window using the latest tarball. When I try to visit the first page in the browser, OI complains that it can't obtain the configuration information from my Stash class. In the Stash class sourcecode, %ITEMS is defined as an empty hash. I must have skipped the step where this gets filled in (unless it's at startup time?). Any suggestions? John Sequeira http://www.pobox.com/~johnseq jo...@po... | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-04 21:53:51
      
     | 
| * Bryce Harrington (br...@ne...) [011203 18:55]: > Has anyone created a package for doing task management? I've created something for personal use, more of a time tracking thing. But it was cut short of finished and I haven't picked it back up again. I know a fairly substantial user has implemented a trouble ticket sort of system, but that's for internal use. One thing to look at is perhaps wrapping around the RT package. RT is being used for bug tracking on perl modules (see http://rt.cpan.org/ for that, and http://www.fsck.com/projects/rt/index.html on RT in general). I haven't used it, but it seems pretty robust -- hey, it uses Template Toolkit! I'm not sure a wrapper would even be appropriate -- maybe for user information/authentication integration, plus fitting its info into your overall site design. Hope this helps, Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: Bryce H. <br...@ne...> - 2001-12-03 23:45:34
      
     | 
| Has anyone created a package for doing task management? One thought I've been dinging around with is integrating task / request / bug tracking in a sort of integrated manner with particular packages. So for example, if you had a setup where you had 'projects', 'people', and 'hardware', you could associate tasks that involve projects (putting out product releases, doing testing runs, etc.), or people (setting up a person's account, assigning him resources, etc.), or hardware (installing a new memory card, figuring out why it crashes on full moons). Anyway, has anyone already done something like this (or even vaguely close)? Background: We've got a collection of technical and non-technical users, with pretty disparate needs and desires. We've tried Bugzilla and Request Tracker, and have considered a few others, but they were all either too complicated to use (for the non-technical) or too hard to customize and extend (for the technical), and I'd like to give a shot at seeing if something could be done with OpenInteract that might provide the optimum solution. Bryce | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-03 13:41:30
      
     | 
| * Bryce Harrington (br...@ne...) [011202 23:31]:
> On Windows you may see references to GUID's.  Dunno if UUID == GUID
> or if there are incompatibilities.  I've seen uuidgen source code
> packages and they're small enough you could reasonably consider
> including it with the app.  I suspect though that you're going to
> have luck finding something UUID-ish on any modern platform you
> might consider.
AFAICT, It's the same thing -- the docs for Data::UUID say:
    This module provides a framework for generating UUIDs (Universally
    Unique Identifiers, also known as GUIDs (Globally Unique
    Identifiers)...
> > Still, I think a SPOPS::Key::UUID is in order. More soon.
> 
> Cool.  :-)
This is in CVS now -- and it's even got a test! It will be in the next
version.
> For the record, here was some of my thinking vis a vis UUID's:
> 
> 1.  Sorting by ID # is really handy in some situations.  For instance,
>     it's a quick and dirty way to check the most recent additions.  An
>     alternative is to record the creation date.
Very true. The time issue should take care of it.
> 2.  UUID's are gonna take up more space in the tables since they're
>     strings.  If you have a lot of foreign keys and junction tables and
>     such, this can mean a significant size increase. They also risk
>     slowing down the application since string compares instead of fast
>     int compares are required.  In practice, I don't know if either of
>     these issues are worth worrying about or not.  Making it a
>     selectable option whether to use ints or uuid's would let the admin
>     determine the best approach according to their particular needs.
Also very true, but IMO space isn't a constraining factor for
anyone. The speed issue might be something to consider, but if it's a
primary key then it's indexed anyway and really doesn't matter.
> 3.  Often it's handier to refer to an item in the db by it's id # than
>     anything else.  I dare anyone to try to memorize 10 uuids, though.
>     ;-)  Not sure of the best solution here, but there's always cut and
>     paste...  ;-)
Absolutely.
> Anyway, I'd love to hear other people's thoughts on these...  (Most
> of the people I usually work with don't really know the pros and
> cons of UUID's vs. autoincrements and view UUID's as nigh-magical,
> which hasn't been very helpful.  *grin*)
The docs refer to the paper with the algorithm for generating these,
which should probably take the magic out of it for your coworkers :-)
Google brings up:
   http://www.webdav.org/specs/draft-leach-uuids-guids-01.txt
One of the benefits of UUIDs can be that people won't refer to a
record by its unique ID, which can be misleading, IMO. Making the UUID
inscrutable reinforces that the fact that it doesn't mean anything. 
it can also be useful for apps that need to use a datastore (possibly
legacy) that has the awkward multiple keys per record.
Thanks for bringing this up, it's very useful.
Chris
-- 
Chris Winters (ch...@cw...)
Building enterprise-capable snack solutions since 1988.
 | 
| 
      
      
      From: Bryce H. <br...@ne...> - 2001-12-03 04:22:24
      
     | 
| On Sun, 2 Dec 2001, Chris Winters wrote:
> * Bryce Harrington (br...@ne...) [011202 23:03]:
> > One alternative I've been considering to auto-increments (but for a
> > somewhat different reason: combining data sources from multiple
> > installations) is UUID's via something like `uuidgen`.  UUID's have
> > their own host of problems but the approach might be worth
> > pondering.
>
> This is a great idea -- the Data::UUID module on CPAN should in theory
> take care of this transparently, although I don't know how well it
> does in non-unix environments. (The author tested under cygwin, but I
> didn't see anything about Win32...)
On Windows you may see references to GUID's.  Dunno if UUID == GUID or
if there are incompatibilities.  I've seen uuidgen source code packages
and they're small enough you could reasonably consider including it with
the app.  I suspect though that you're going to have luck finding
something UUID-ish on any modern platform you might consider.
> Still, I think a SPOPS::Key::UUID is in order. More soon.
Cool.  :-)
For the record, here was some of my thinking vis a vis UUID's:
1.  Sorting by ID # is really handy in some situations.  For instance,
    it's a quick and dirty way to check the most recent additions.  An
    alternative is to record the creation date.
2.  UUID's are gonna take up more space in the tables since they're
    strings.  If you have a lot of foreign keys and junction tables and
    such, this can mean a significant size increase. They also risk
    slowing down the application since string compares instead of fast
    int compares are required.  In practice, I don't know if either of
    these issues are worth worrying about or not.  Making it a
    selectable option whether to use ints or uuid's would let the admin
    determine the best approach according to their particular needs.
3.  Often it's handier to refer to an item in the db by it's id # than
    anything else.  I dare anyone to try to memorize 10 uuids, though.
    ;-)  Not sure of the best solution here, but there's always cut and
    paste...  ;-)
Anyway, I'd love to hear other people's thoughts on these...  (Most of
the people I usually work with don't really know the pros and cons of
UUID's vs. autoincrements and view UUID's as nigh-magical, which hasn't
been very helpful.  *grin*)
Bryce
 | 
| 
      
      
      From: Chris W. <ch...@cw...> - 2001-12-03 04:02:30
      
     | 
| * Bryce Harrington (br...@ne...) [011202 23:03]: > One alternative I've been considering to auto-increments (but for a > somewhat different reason: combining data sources from multiple > installations) is UUID's via something like `uuidgen`. UUID's have > their own host of problems but the approach might be worth > pondering. This is a great idea -- the Data::UUID module on CPAN should in theory take care of this transparently, although I don't know how well it does in non-unix environments. (The author tested under cygwin, but I didn't see anything about Win32...) Still, I think a SPOPS::Key::UUID is in order. More soon. Chris -- Chris Winters (ch...@cw...) Building enterprise-capable snack solutions since 1988. | 
| 
      
      
      From: Bryce H. <br...@ne...> - 2001-12-03 03:54:45
      
     | 
| On Sun, 2 Dec 2001, Chris Winters wrote: > * Andreas Nolte (And...@Be...) [011201 16:34]: > > has anyone of you used OI with mysql and the GEMINI table type? We > > have some concurrency problems, since we have a lot of updates and > > inserts with auto_increment columns. The docs of mysql say, that > > MyISAM talbes ( the default ) lock the whole table, before an > > auto_increment insert can be made - this hurts performance badly in > > our environment, where multiple users do updates / inserts at the > > same time. > > I'd never heard this before, that it locks the whole table. Sheesh. > > Another alternative is dumping auto-incrementing fields entirely and > using something else -- it probably wouldn't be too difficult to > create a socket server (in Perl even!) to generate unique IDs. But > this might have bigger problems that using the Gemini/InnoDB tables -- > another point of failure, etc. One alternative I've been considering to auto-increments (but for a somewhat different reason: combining data sources from multiple installations) is UUID's via something like `uuidgen`. UUID's have their own host of problems but the approach might be worth pondering. Bryce |