Hi guys, I made changes to roundcube CVS Sources to request, when composing emails, "Notification Requests".
I've added a checkbox to handle requests.
I send patch to M/L and Forum :)
Regards, Salvatore
I'm experiencing an infuriating "mostly repeatable" bug where when I first login to RC, and/or intermittently when I click on the Round Cube logo to go back to the Inbox (rather than clicking on Inbox) I can't toggle read/unread status ( I can toggle the icon but the mark function fails).
I've narrowed it down to the fact that it's storing and subsequently pulling the wrong mailbox name from the _SESSION. The mailbox that get's assigned in the SESSION is whichever one is stored last in the array $rcmail_config['default_imap_folders'] defined in the config file. If I set the Sent to last in that list, Sent is stored as mbox in the Session. If Trash is last, that becomes mbox.
Usually when I'm debugging, finding out what the bad data is, and where it is is 99% of the debugging. In this case I can't for the life of me figure out what is setting the mbox Session variable.
Here is my snippet from main.inc where I can prove that things first appear to go haywire:
error_log('main.inc 89: ' . time() . "\t" . $_SESSION['mbox'] . "\n",3,$logfile); error_log('main.inc 89 mailbox: ' . time() . "\t" . $IMAP->mailbox . "\n",3,$logfile);
// init session session_start(); $sess_id = session_id();
error_log('main.inc 91: ' . time() . "\t" . $_SESSION['mbox'] . "\n",3,$logfile);
Is anyone else experiencing the bug? Can anyone help me figure out how/where the session variable mbox is being assigned the last array value of $rcmail_config['default_imap_folders']
-Charles
I cannot reproduce this behavior. However, I added the following to program/steps/mail/func.inc to make sure that the correct mailbox is stored within the session:
if (empty($_SESSION['mbox'])) $_SESSION['mbox'] = $IMAP->get_mailbox_name();
Regards, Thomas
Charles McNulty wrote:
I'm experiencing an infuriating "mostly repeatable" bug where when I first login to RC, and/or intermittently when I click on the Round Cube logo to go back to the Inbox (rather than clicking on Inbox) I can't toggle read/unread status ( I can toggle the icon but the mark function fails). I've narrowed it down to the fact that it's storing and subsequently pulling the wrong mailbox name from the _SESSION. The mailbox that get's assigned in the SESSION is whichever one is stored last in the array $rcmail_config['default_imap_folders'] defined in the config file. If I set the Sent to last in that list, Sent is stored as mbox in the Session. If Trash is last, that becomes mbox. Usually when I'm debugging, finding out what the bad data is, and where it is is 99% of the debugging. In this case I can't for the life of me figure out what is setting the mbox Session variable. Here is my snippet from main.inc where I can prove that things first appear to go haywire:
error_log('main.inc 89: ' . time() . "\t" . $_SESSION['mbox'] . "\n",3,$logfile); error_log('main.inc 89 mailbox: ' . time() . "\t" . $IMAP->mailbox . "\n",3,$logfile);
// init session session_start(); $sess_id = session_id();
error_log('main.inc 91: ' . time() . "\t" . $_SESSION['mbox'] . "\n",3,$logfile);
Is anyone else experiencing the bug? Can anyone help me figure out how/where the session variable mbox is being assigned the last array value of $rcmail_config['default_imap_folders']
-Charles
Thomas Bruederli wrote:
I cannot reproduce this behavior. However, I added the following to
Okay, here are the precise instructions I can follow for reproducing the problem everytime on the most recent CVS:
not change in the folder list 4) Navigate to another folder by clicking on it on the left column 5) Navigate back to the Inbox - Observe that the Unread status has not been saved 6) repeat step 3. Observe that the Unread number now *does* change and that the status is correctly saved.
Can anyone else reproduce these steps and results?
-Charles
Hi Salvatore
Many thanks for your patch. I just applied it and will commit the changed file to the CVS soon. I guess now we also have to handle incoming mails requesting a receipt... any help on this would be appreciated.
Regards, Thomas
Salvatore Ansani wrote:
Hi guys, I made changes to roundcube CVS Sources to request, when composing emails, "Notification Requests".
I've added a checkbox to handle requests.
I send patch to M/L and Forum :)
Regards, Salvatore
On Mon, 01 May 2006 09:17:20 -0500, Charles McNulty charles@charlesmcnulty.com wrote:
Can anyone else reproduce these steps and results?
Hi Charles,
I can confirm that I can reproduce the same behavior every time I try. Its annoying, but I'd always worked around it as you suggested.
Happy to help test solutions or get you any information I can to help with debugging.
Andrew
On Mon, 01 May 2006 16:50:09 +0200, Thomas Bruederli roundcube@gmail.com wrote:
Hi Salvatore
Many thanks for your patch. I just applied it and will commit the changed file to the CVS soon. I guess now we also have to handle incoming mails requesting a receipt... any help on this would be appreciated.
But this should be an optional. I personally don't like to have mails come out from return receipt, as that's one of the ways spammers confirm reception of mails.
Martin Marques wrote:
On Mon, 01 May 2006 16:50:09 +0200, Thomas Bruederli roundcube@gmail.com wrote:
Hi Salvatore
Many thanks for your patch. I just applied it and will commit the changed file to the CVS soon. I guess now we also have to handle incoming mails requesting a receipt... any help on this would be appreciated.
But this should be an optional. I personally don't like to have mails come out from return receipt, as that's one of the ways spammers confirm reception of mails.
I don't thin of an automatic confirmation but pupping up an alert or whatever asking the user if he/she wants to send the confirmation notice. Like Thunderbird does.
Thomas
On Mon, 01 May 2006 19:35:30 +0200, Thomas Bruederli roundcube@gmail.com wrote:
Martin Marques wrote:
But this should be an optional. I personally don't like to have mails
come out from return receipt, as that's one of the ways spammers confirm reception of mails.
I don't thin of an automatic confirmation but pupping up an alert or whatever asking the user if he/she wants to send the confirmation notice. Like Thunderbird does.
That would be OK. :-)
On Mon, 1 May 2006 14:48:15 -0300, Martin Marques martin@bugs.unl.edu.ar wrote:
I don't thin of an automatic confirmation but pupping up an alert or whatever asking the user if he/she wants to send the confirmation notice. Like Thunderbird does.
That would be OK. :-)
I'm working on it :) By now I've created code to handle receipt request and display confirmation POPUP (or an AJAX dialog).
Tomorrow I think I can complete code to send back receipt, so I can release PATCH :)
Bau, Salvatore
Thomas Bruederli wrote:
I cannot reproduce this behavior. However, I added the following to program/steps/mail/func.inc to make sure that the correct mailbox is stored within the session:
if (empty($_SESSION['mbox'])) $_SESSION['mbox'] = $IMAP->get_mailbox_name();
Regards, Thomas
I couldn't get this patch to work, and in fact I don't believe from looking at it that it would address the issue. The only place anywhere in the code that $_SESSION['mbox'] is set is in func.inc:
// set imap properties and session vars if (strlen($_GET['_mbox'])) { $IMAP->set_mailbox($_GET['_mbox']); $_SESSION['mbox'] = $_GET['_mbox']; }
If I put your snippet before this it would just be overwritten, and if I put it after it would overwrite it with an incorrect value.
The really weird thing about this bug is that when I log $_GET['_mbox'], I'm never able to observe _GET as the wrong value (the last value in the folders array). I logged dozens of lines of code to try to pinpoint where else $_SESSION['mbox'] was being assigned and I just couldn't find it.
To summarize the issues, the bug can be observed by following these steps:
not change in the folder list 4) Navigate to another folder by clicking on it on the left column 5) Navigate back to the Inbox - Observe that the Unread status has not been saved 6) repeat step 3. Observe that the Unread number now *does* change and that the status is correctly saved.
The problem is that the program is attempting to flag a message in a mailbox that doesn't exist because it's looking in the wrong mailbox. It's looking in the mailbox that's stored in the SESSION['mbox'] and the SESSION['mbox'] is inexplicably set to the last value in the array $rcmail_config['default_imap_folders'] defined in main.inc.php.
SESSION['mbox'] is only ever explicitly set in func.inc where it is assigned to a $_GET value.
I know this is hard to follow, but if anyone who can reproduce the error has any time to look at it or a decent way to work around it, I'd really appreciate it!
-Charles
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I'm thinking more along the lines of a NSIS (http://nsis.sourceforge.net/Main_Page) for PHP scripts. It doesn't look to me like anything like it exists, so I thought I'd take a crack at it using RC as my guinea pig.
-Charles
I would check out something like gallery2s install script i used it just the other day and was really impressed with the ease of it.
Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I'm thinking more along the lines of a NSIS (http://nsis.sourceforge.net/Main_Page) for PHP scripts. It doesn't look to me like anything like it exists, so I thought I'd take a crack at it using RC as my guinea pig.
-Charles
The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete this material from any computer.
In accordance with industry regulations, all messages are retained and are subject to monitoring.
This message has been scanned for viruses and dangerous content and is believed to be clean.
Securities offered through Cantella & Co., Inc., Member NASD/SIPC. Home Office: 2 Oliver Street, 11th Floor, Boston, MA 02109 Telephone: (617)521-8630
Agora shopping cart has a nice installer. lifetype.net recently
(in our development version) added one page of installation checks, which hopefully will greatly reduce the troubles people have in successfully installing, and then getting failures on first use.
On Mon, 1 May 2006, Chris Richardson wrote:
I would check out something like gallery2s install script i used it just the other day and was really impressed with the ease of it.
Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I'm thinking more along the lines of a NSIS (http://nsis.sourceforge.net/Main_Page) for PHP scripts. It doesn't look to me like anything like it exists, so I thought I'd take a crack at it using RC as my guinea pig.
Jon Daley wrote:
Agora shopping cart has a nice installer. lifetype.net recently
(in our development version) added one page of installation checks, which hopefully will greatly reduce the troubles people have in successfully installing, and then getting failures on first use.
On Mon, 1 May 2006, Chris Richardson wrote:
I would check out something like gallery2s install script i used it just the other day and was really impressed with the ease of it.
Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I'm thinking more along the lines of a NSIS (http://nsis.sourceforge.net/Main_Page) for PHP scripts. It doesn't look to me like anything like it exists, so I thought I'd take a crack at it using RC as my guinea pig.
Oh why not... I'll take it on in its first form!!
~Brett
Charles,
Please do not reply on a post when opening a new thread. Because all references are kept within the replied message, my thread view gets messed up and so does the archive: http://lists.roundcube.net/mail-archive/roundcube.dev/2006/05/
Write a new message from scratch if you want to post a new topic.
Regards, Thomas
Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I'm thinking more along the lines of a NSIS (http://nsis.sourceforge.net/Main_Page) for PHP scripts. It doesn't look to me like anything like it exists, so I thought I'd take a crack at it using RC as my guinea pig.
-Charles
On Mon, 2006-05-01 at 17:33 -0500, Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I realize this is not PHP 1-click install, but several weeks ago I contributed GNU autotool support to perform installation. It does not seem to have been imported into CVS yet. It was a patch submitted in bug #1460202.
PHP 1-click install would be very nice, but there are other installation scenarios which will need a command line script mechanism. The two should compliment one another.
John Dennis wrote:
On Mon, 2006-05-01 at 17:33 -0500, Charles McNulty wrote:
On a completely different note, is anyone currently working on a "point-and-click" install, and if not, does anyone know of a "php script installer" project? I googled around for one but all I could find were one-click scripts for intended for web hosts and which have to be adapted for each script. Not really what I'm looking for.
I realize this is not PHP 1-click install, but several weeks ago I contributed GNU autotool support to perform installation. It does not seem to have been imported into CVS yet. It was a patch submitted in bug #1460202.
PHP 1-click install would be very nice, but there are other installation scenarios which will need a command line script mechanism. The two should compliment one another.
Gee.. if you've already done it.. then I guess I should stop working on mine (not that it's that good anyway).
~Brett
Charles McNulty wrote:
To summarize the issues, the bug can be observed by following these steps:
- Login
- Click on the 'Round Cube Webmail' icon on the top left corner
- Mark any Read message as Unread - Observe that the Unread number does
not change in the folder list 4) Navigate to another folder by clicking on it on the left column 5) Navigate back to the Inbox - Observe that the Unread status has not been saved 6) repeat step 3. Observe that the Unread number now *does* change and that the status is correctly saved.
..
SESSION['mbox'] is only ever explicitly set in func.inc where it is assigned to a $_GET value.
I know this is hard to follow, but if anyone who can reproduce the error has any time to look at it or a decent way to work around it, I'd really appreciate it!
-Charles
Well, I finally figured it out and comitted code to fix this. I should have known from the fact that I couldn't figure out where the Session var was being set that it was a register globals issue. Anyway the fix was simply going through and renaming a whole slew of $mbox to a different variable name (mostly $mbox_name except where something else was a little better (like $new_mbox or $mbox_row) Anyway, please refrain from using $mbox or any other session variable names. I know it's stupid that people still have register globals on, but sometimes people don't have a choice.
-Charles