Re: [Evolution-hackers] Calling construct() multiply times?
Figured out why I kept getting called back into the construct() callback whenever I select a folder. In the connect() callback, I checked to see if service-url-port is 0, if so, set it to the port to the default port number. I did this because I wanted to know what the port number whereever I'm at in the code. That port number is messing it up and I haven't figured out what Camel is doing but I just #ifdef the code out for now and hard coded all the port number to what I need to use. On 5/2/2002 at 10:32 AM Not Zed wrote: It would probably be easy to point out the problems if you gave us a pointer to the source, like we do. :) On Thu, 2002-05-02 at 05:12, Dan Winship wrote: On Wed, 2002-05-01 at 13:02, Shawn Walker wrote: I constructed my base_url as: provider://username@server/ In get_folder_info_online() I created a URL for each folder as: provider://username@server/folder1 That URL is being created by camel_url_to_string(). In hash_folder_name(), I get: provider://username@server/folder1;noselect=yes So you want to make sure that your hash and equals functions don't consider the path to be part of the URL. but, I don't get ;noselect=yes for all the folders, just some (I haven't looked into what ;noselect=yes means). It means evolution-mail thinks the folder is not selectable, which has to do with what fields you did and didn't fill in in the CamelFolderInfo in get_folder_info. -- Dan ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers
[Evolution-hackers] Calling construct() multiply times?
Why is Camel keep calling my construct() routine multiple times? I have the callback setup to call into my connect() function and I do my own connection, return TRUE. Camel then calls my get_folder_info() and I return the folder tree. I select a folder that I want to show the messages, Camel then calls my construct() and connect() again when I already have a connection. What do I need to tell Camel since I don't need another connection, I already have it. Note, I'm not using any of Camel connection calls to establish the connection to the server, I'm doing it. ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers
Re: [Evolution-hackers] Calling construct() multiply times?
I constructed my base_url as: provider://username@server/ In get_folder_info_online() I created a URL for each folder as: provider://username@server/folder1 That URL is being created by camel_url_to_string(). In hash_folder_name(), I get: provider://username@server/folder1;noselect=yes but, I don't get ;noselect=yes for all the folders, just some (I haven't looked into what ;noselect=yes means). I'm not including any port number for the URL. I should also mention that when I get called into my construct() the first time, the path is set to /, but when I click on a folder and it called construct(), the path is now /folder1. On 5/1/2002 at 10:53 AM Dan Winship wrote: On Wed, 2002-05-01 at 09:28, Shawn Walker wrote: Why is Camel keep calling my construct() routine multiple times? I have the callback setup to call into my connect() function and I do my own connection, return TRUE. Camel then calls my get_folder_info() and I return the folder tree. I select a folder that I want to show the messages, Camel then calls my construct() and connect() again when I already have a connection. Most likely, the URLs you're returning from get_folder_info do not compare as equal to the URL that was used to build the folder tree, and so it's creating a new store object. Eg, if the URL you configured in the config dialog is foobar://server:999, and get_folder_info returns foobar://server/folder1, foobar://server/folder2, etc, and your url hash and compare functions look at the port number, then it would create a new store foobar://server to fetch the folders from. Likewise with the other URL components. -- Dan ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers
Re: [Evolution-hackers] Calling construct() multiply times?
On Wed, 2002-05-01 at 13:02, Shawn Walker wrote: I constructed my base_url as: provider://username@server/ In get_folder_info_online() I created a URL for each folder as: provider://username@server/folder1 That URL is being created by camel_url_to_string(). In hash_folder_name(), I get: provider://username@server/folder1;noselect=yes So you want to make sure that your hash and equals functions don't consider the path to be part of the URL. but, I don't get ;noselect=yes for all the folders, just some (I haven't looked into what ;noselect=yes means). It means evolution-mail thinks the folder is not selectable, which has to do with what fields you did and didn't fill in in the CamelFolderInfo in get_folder_info. -- Dan ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers
Re: [Evolution-hackers] Calling construct() multiply times?
You realise this doesn't excuse you from licensing GPL, right? Just making sure :-) Jeff On Wed, 2002-05-01 at 09:28, Shawn Walker wrote: [snip] Note, I'm not using any of Camel connection calls to establish the connection to the server, I'm doing it. ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers ___ evolution-hackers maillist - [EMAIL PROTECTED] http://lists.ximian.com/mailman/listinfo/evolution-hackers