smolder-devel Mailing List for Smolder (Page 3)
Brought to you by:
michael_peters
You can subscribe to this list here.
2006 |
Jan
|
Feb
|
Mar
(14) |
Apr
(1) |
May
|
Jun
(23) |
Jul
(11) |
Aug
(15) |
Sep
(10) |
Oct
(1) |
Nov
|
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2007 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
(2) |
Nov
|
Dec
|
2008 |
Jan
|
Feb
|
Mar
(2) |
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
(6) |
Sep
|
Oct
|
Nov
|
Dec
|
2009 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
(7) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2010 |
Jan
|
Feb
(3) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Mark S. <ma...@su...> - 2006-07-10 02:33:05
|
Michael Peters wrote: > The only reason we bundle the full sqlite binary with smolder is because we use > the .dump command to get the full database dump. I asked Matt Sargeant if he > could expose this functionality in DBD::SQLite. He pointed out that it's only > implemented in the CLI binary, but suggested I look at it's code and translate > it to Perl. So I did. Cool. Now that you've done it, it would be neat to see that functionality rolled back into DBD::SQLite. Now for a bug report of sorts. My "tmp" directory has become full of lots of randomly named subdirectories. About 263M worth. When I run "smolder_install", it copies this 263M meg worth of temp files to the target.... and then deletes them there. So it would be nice to not copy the temp files in the first place. Currently "cp" is being used to copy the files, which doesn't have an easy way to exclude things, unless you are using fancy shell tricks. Assuming the structure of folder is fairly static, perhaps the "cp" could be changed to specifically /include/ the files and or directories to copy. Mark -- http://mark.stosberg.com/ |
From: Michael P. <mp...@pl...> - 2006-07-10 00:55:35
|
The only reason we bundle the full sqlite binary with smolder is because we use the .dump command to get the full database dump. I asked Matt Sargeant if he could expose this functionality in DBD::SQLite. He pointed out that it's only implemented in the CLI binary, but suggested I look at it's code and translate it to Perl. So I did. It was pretty easy once I figured out that there's a table named sqlite_master that holds all the meta data for the database. It even holds the full table creation SQL, which makes extracting that really easy. So I've committed those changes to svn and completely removed the sqlite src only using DBD::SQLite. All tests are passing and the dump in the backup looks good too. Please try it out. Mark, Sorry about the recent work you did to make the sqlite binary build on freebsd. It's now obsolete :) -- Michael Peters Developer Plus Three, LP |
From: Michael P. <mp...@pl...> - 2006-07-06 19:55:31
|
Sam Tregar wrote: > I'm seeing weird grey-ed out bar-graphs when I choose a long enough > time-range. Screenshot attached. I figured out what this is. When using a 3d bar graph, each bar is separated by a little 1 px gray line. When there is a lot of data and each bar becomes less than 1 px wide to fit in the same dimensions, all you see are the 1px gray lines. With that much data, you can switch to a line, area or points graph to see the same data better. -- Michael Peters Developer Plus Three, LP |
From: Michael P. <mp...@pl...> - 2006-07-05 17:13:46
|
Sam Tregar wrote: > On Wed, 5 Jul 2006, Sam Tregar wrote: > >> I'm seeing weird grey-ed out bar-graphs when I choose a long enough >> time-range. Screenshot attached. > > As a side-note, I was thinking about playing with using a Canvas-based > chart with Smolder. This one looks pretty sweet: > > http://webfx.eae.net/dhtml/chart/chart.html > > I don't know if it would deal with mountains of data any better than > GD::Graph, but it would fit with the overall Web 2.0 feel of Smolder. > It also opens up the possibility of doing animations in the graphs, > like morphing from one graph to another when the settings change. > None of that is in the WebFX code now but it doesn't look like it > would be hard to do. That would be cool. I'd definitely welcome that addition. GD* is so web 1.0 -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@tr...> - 2006-07-05 17:09:27
|
On Wed, 5 Jul 2006, Sam Tregar wrote: > I'm seeing weird grey-ed out bar-graphs when I choose a long enough > time-range. Screenshot attached. As a side-note, I was thinking about playing with using a Canvas-based chart with Smolder. This one looks pretty sweet: http://webfx.eae.net/dhtml/chart/chart.html I don't know if it would deal with mountains of data any better than GD::Graph, but it would fit with the overall Web 2.0 feel of Smolder. It also opens up the possibility of doing animations in the graphs, like morphing from one graph to another when the settings change. None of that is in the WebFX code now but it doesn't look like it would be hard to do. -sam |
From: Sam T. <sa...@tr...> - 2006-07-05 17:04:08
|
I'm seeing weird grey-ed out bar-graphs when I choose a long enough time-range. Screenshot attached (this time less than 40k). -sam |
From: Sam T. <sa...@tr...> - 2006-07-05 17:00:19
|
I'm seeing weird grey-ed out bar-graphs when I choose a long enough time-range. Screenshot attached. -sam |
From: Michael P. <mp...@pl...> - 2006-07-05 03:25:03
|
Sam Tregar wrote: > On Fri, 30 Jun 2006, Michael Peters wrote: > >> Weird, I've never seen that. Do you know what the URL was that it >> was trying to redirect you to? > > I think it was > http://airtrout.homelinux.com:8081/app/public_auth/timeout. I think > there was a query arg too, but I didn't think to write it down. Ok, I tracked this down. It was a weird interaction of using mod_auth_tkt's guest login feature to protect the timeout/unauthorized section with the same timeout setting as the normal application. So it just redirected over and over creating that huge URL. The solution that seems to work is to set a timeout of 1 year on the timeout/unauthorized section and have it set to not refresh the cookie so that you won't get weird things like a timeout refreshing your cookie. A cleaner solution would be to turn off Apache's auth handling for the timeout <Location>, but I don't think you can do that since the parent <Location> has it turned on. -- Michael Peters Developer Plus Three, LP |
From: Mark S. <ma...@su...> - 2006-07-01 21:05:52
|
There appears to be a SQLite bootstrapping problem. smolder_createdb is creating a database file which it then cannot read. I suspect this is because during part of the process a system-wide version is being used, and at another stage the bundled version is being used. Here's how it looks: /bin/smolder_createdb --destroy This program will drop and recreate Smolder the database for this installation. Are you sure you want to proceed? (y/n) [default y] y Could not update db_version for db named: smolder! DBD::SQLite::db do failed: unsupported file format(1) at dbdimp.c line 269 [for Statement "UPDATE db_version set db_version=0.2"] at ./bin/smolder_createdb line 245, <STDIN> line 1. #### I just committed some minor doc and test updates. Mark -- http://mark.stosberg.com/ |
From: Sam T. <sa...@tr...> - 2006-06-30 21:20:33
|
On Fri, 30 Jun 2006, Michael Peters wrote: > Weird, I've never seen that. Do you know what the URL was that it > was trying to redirect you to? I think it was http://airtrout.homelinux.com:8081/app/public_auth/timeout. I think there was a query arg too, but I didn't think to write it down. -sam |
From: Michael P. <mp...@pl...> - 2006-06-30 21:16:56
|
Sam Tregar wrote: > My session timed-out but instead of getting a login or a new cookie I > got a firefox error: > > The page isn't redirecting properly > > Firefox has detected that the server is redirecting the request for > this address in a way that will never complete. > > I refreshed that a couple times and then got a 414 error, Request-URI > Too Large. That's a new one for me, and apparently for firefox too - > it trashed the display of the address bar. Weird, I've never seen that. Do you know what the URL was that it was trying to redirect you to? It normally just tries to take you back to the page you were viewing. The url's are pretty short in smolder, so I'm not sure what it could have been. I'm pretty sure it doesn't do any POST=>GET conversion or anything like that. -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@tr...> - 2006-06-30 21:08:58
|
My session timed-out but instead of getting a login or a new cookie I got a firefox error: The page isn't redirecting properly Firefox has detected that the server is redirecting the request for this address in a way that will never complete. I refreshed that a couple times and then got a 414 error, Request-URI Too Large. That's a new one for me, and apparently for firefox too - it trashed the display of the address bar. Restarting my browser fixed it and I logged in successfully. -sam |
From: Sam T. <sa...@tr...> - 2006-06-30 18:09:36
|
The XML link was returning the raw compressed report contents rather than uncompressing it. -sam |
From: Sam T. <sa...@tr...> - 2006-06-30 17:12:33
|
The graphing code used a MySQL-specific SQL function to format dates. I moved the date formatting into Perl and now graphs work for SQLite. -sam |
From: Michael P. <mp...@pl...> - 2006-06-30 09:20:44
|
Sam Tregar wrote: > I found the bug that was breaking email for me. A missing close() > resulted in send_emails() being called while the XML file was only > partially written to disk. This gave XML::Parser heartburn, but left > me confused since when I looked at the file it was always perfect! Sorry about the headeache. Patch applied. -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@tr...> - 2006-06-29 22:02:53
|
I found the bug that was breaking email for me. A missing close() resulted in send_emails() being called while the XML file was only partially written to disk. This gave XML::Parser heartburn, but left me confused since when I looked at the file it was always perfect! -sam |
From: Michael P. <mp...@pl...> - 2006-06-27 21:44:23
|
Sam Tregar wrote: > On Tue, 27 Jun 2006, Michael Peters wrote: > >> I've added a new version of Dispatch. This should fix some of these woes. > > Yup, now I get a real ISE, with the expected error: > > [Tue Jun 27 14:25:09 2006] [error] Error executing run mode > 'process_add_report': undef error - \nnot well-formed (invalid token) > at line 1, column 0, byte 0 at > /home/sam/smolder/lib/i686-linux/XML/Parser.pm line 187\n at > /home/sam/smolder/lib/CGI/Application/Dispatch.pm line 506\n Well, that's good right :) > This is against a fresh database after running the tests, all of which > passed. Are there tests for the email code? It seems like there > might not be judging by the TODO in t/db-smoke-test.t: > > local $TODO = "not testing email sending"; > > What do you think of the test-SMTP-server stuff I did in Arcos? I > don't think it would be too hard to port to Smolder. Yeah, there's no tests for that. It was done test-smtp arcos stuff. I think that would be a great stuff to add to the test suite. I'll put it on the TODO. If you feel like tackling it, let me know and I'll give you svn right access. -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@pl...> - 2006-06-27 18:29:52
|
On Tue, 27 Jun 2006, Michael Peters wrote: > I've added a new version of Dispatch. This should fix some of these woes. Yup, now I get a real ISE, with the expected error: [Tue Jun 27 14:25:09 2006] [error] Error executing run mode 'process_add_report': undef error - \nnot well-formed (invalid token) at line 1, column 0, byte 0 at /home/sam/smolder/lib/i686-linux/XML/Parser.pm line 187\n at /home/sam/smolder/lib/CGI/Application/Dispatch.pm line 506\n This is against a fresh database after running the tests, all of which passed. Are there tests for the email code? It seems like there might not be judging by the TODO in t/db-smoke-test.t: local $TODO = "not testing email sending"; What do you think of the test-SMTP-server stuff I did in Arcos? I don't think it would be too hard to port to Smolder. -sam |
From: Michael P. <mp...@pl...> - 2006-06-27 04:34:09
|
Michael Peters wrote: > > Sam Tregar wrote: > >> Something must be eating $@ up the chain, but I don't see it. > > I think it's the latest (developer) version of CGI::Application::Dispatch. I've > seen this in other places too. I'm poking into it. I've added a new version of Dispatch. This should fix some of these woes. -- Michael Peters Developer Plus Three, LP |
From: Michael P. <mp...@pl...> - 2006-06-27 01:47:12
|
Sam Tregar wrote: > Something must be eating $@ up the chain, but I don't see it. I think it's the latest (developer) version of CGI::Application::Dispatch. I've seen this in other places too. I'm poking into it. -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@tr...> - 2006-06-26 21:17:12
|
On Mon, 26 Jun 2006, Sam Tregar wrote: > I see the warn() but not the croak(): > > TEMPLATE ERROR: undef error - > not well-formed (invalid token) at line 1, column 0, byte 0 at > /home/sam/smolder/lib/i686-linux/XML/Parser.pm line 185 > > Something must be eating $@ up the chain, but I don't see it. Oh - I also don't know what XML is choking XML::Parser. The XML for the report I'm uploading is well-formed and the email templates look fine. -sam |
From: Sam T. <sa...@tr...> - 2006-06-26 21:14:28
|
Email stopped going out after I updated Smolder this morning. All I'm getting is a blank screen after uploading a new report and nothing's in the logs. I've tracked it down to the TT run in Smolder::Email::send_mime_mail: $TEMPLATE->process( "Email/$name.tmpl", $tt_params, \$html ) || croak $TEMPLATE->error(); That's failing silently somehow. If I replace it with: unless($TEMPLATE->process( "Email/$name.tmpl", $tt_params, \$html)) { warn("TEMPLATE ERROR: " . $TEMPLATE->error()); croak $TEMPLATE->error(); } I see the warn() but not the croak(): TEMPLATE ERROR: undef error - not well-formed (invalid token) at line 1, column 0, byte 0 at /home/sam/smolder/lib/i686-linux/XML/Parser.pm line 185 Something must be eating $@ up the chain, but I don't see it. -sam |
From: Michael P. <mp...@pl...> - 2006-06-25 14:32:43
|
This email is announcing the release of Smolder 0.2 to add some enhancements and bugfixes on the way to 1.0. The biggest addition in this release is the built-in SQLite support, controlled by the DBPlatform configuration var. This also opens the way for other DB's to be supported in the future. Changes for this release: * Turn on mod_auth_tkt for all dynamic requests so that Apache auth stages are run and valid credentials produced - M. Peters * Added smolder_prove, to run any normal Perl tests for smolder - M. Peters * Send the port number in URLs in emails - Sam Tregar * Skip the authinfo test if Apache is not running - Sam Tregar * Added missing modules to src (XML::Parser, XML::SAX::Expat, Exception::Class and Exception::Class::TryCatch) - Sam Tregar * Using a smaller number of Apache children and Apache::SizeLimit to limit any memory hogging - M. Peters * Added built-in SQLite support - M. Peters * Added multi-db support via Smolder::DBPlatform - M. Peters * Upgraded DBI to 1.50 - M. Peters * Added notification to the user of failed AJAX requests - M. Peters -- Michael Peters Developer Plus Three, LP |
From: Sam T. <sa...@tr...> - 2006-06-24 01:30:05
|
I setup my Smolder on port 8081 but the links sent out in email didn't include the port. The attached patch fixes this. -sam |
From: Michael P. <mp...@pl...> - 2006-06-23 16:33:05
|
Sam Tregar wrote: > Cool. I also added XML::Parser and XML::SAX::Expat. Without them I > was getting warnings from XML::SAX about not finding > ParserDetails.ini, which is usually a sign that XML::SAX can't find a > real parser and will fall back to XML::SAX::PurePerl, which is buggy > and slow. also added. Thanks again. -- Michael Peters Developer Plus Three, LP |