Makefile for plain C source, how to? (IMAP proxy)

[M66B]

It seems you are right. I just looked at the statusbar, but apparently if it shows logging in, actually a cache connection could be reused.

 

[er13]

It seems you are right.

Then (as Ralf actually already said in 31) there is absolutely no reason for using an imap proxy with thunderbird... and I assume with evolution too...

 

[M66B]

Except then for using Yahoo! IMAP, which requires a special command before logging in, although it is possible to use a modified Thunderbird version too (available here).
Actually I mostly use the Yahoo! web interface, but there is no easy way to backup messages. Using Thunderbird (or another client) I could download all messages local by synchronizing all folders.

 

[M66B]

Attached a tested patch for ImapProxy, including a little Web-IF.
Most English texts should be translated to German.
SSL/TLS is not tested yet (no support in the Web-IF too yet).
Comments and suggestions are of course welcome.

 

[M66B]

Can somebody tell me if and how I can use inet for ImapProxy?
Of course it would be best if there was a paragraph about it in the wiki ...

 

[M66B]

Attached a new patch with more configuration options and a better Yahoo! IMAP patch.
Not tested/working are inet and ssl support.
ImapProxy seems to be quite stable, synchronizing all my e-mail of the last years worked without any problem (in contrast to yahooproxy).

I have a few questions:

1. Should the daemon be inet ready to be used by inet (I don't get it working, but maybe I have made errors, because I am not sure about how to do this)
   

   mail.warn in.imapproxyd[13238]: main(): bind() failed: Address already in use

2. How should the execute permission be set in a patch? Is this überhaupt possible?

 

[olistudent]

1. Perhaps it's started already (ps)
2. It's not possible to set these bits in patches.

Regards
Oliver

 

[M66B]

1. Perhaps it's started already (ps)

I checked and it was not started already. What else could be wrong?

 

[RalfFriedl]

It probably needs an inetd mode (note inetd and not inet). Some programs detect this automatically, some require an option and some don't support this mode.

 

[M66B]

2. It's not possible to set these bits in patches

Is it possible/desirable to set the execute persmission in the package .mk file?

(/etc/default.xxx/xxx_conf, /etc/init.d/rc.xxx, /usr/lib/cgi-bin/zzz.cgi)

 

[Silent-Tears]

The time its only used in patches, the execute-bit is not set. Before checked in a developer will set the bits correctly.

 

[er13]

Frist of all the application must support inetd-mode, e.g. it should not try to bind the listen port by itself (is that the reason for "already in use")...
Besides that, what sense does it make to run an application which main purpose is to cache a connection in inetd mode?

 

[M66B]

I couldn't find anything about inetd support, so you are probably right. Attached a patch with inetd support removed.

Connections are cached only for a limited time (default 300 seconds), so using inetd could save some resources.

 

[RalfFriedl]

It is not possible to do cached connections in inetd mode.

 

[M66B]

I am completely new to the inetd stuff (although I have a general understanding about what it does), so I would appreciate it if you could explain this in short.

 

[M66B]

A new patch has been attached to ticket #847.

It is not possible to do cached connections in inetd mode

I think this applies to incoming connections, but not to outgoing connections.