From: SourceForge.net <no...@so...> - 2010-03-20 18:09:43
|
Feature Requests item #1782896, was opened at 2007-08-27 17:53 Message generated for change (Comment added) made by egavilan You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=360894&aid=1782896&group_id=10894 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: 24. Channel Commands Group: None Status: Open Resolution: None Priority: 9 Private: No Submitted By: Jeremy Gibson (silpheed) Assigned to: Reinhard Max (rmax) Summary: Wanted: IPv6 Initial Comment: I work on some government-related code and we have a mandate to convert to IPv6 this next coming year. I was wondering if IPv6 could make it into 8.6 ??? ---------------------------------------------------------------------- Comment By: Emiliano (egavilan) Date: 2010-03-20 15:09 Message: It seems AI_ADDRCONFIG is an specific glibc optimization; with the obvious #ifdef the build completes. Whowever, there are failing socket test: ==== socket-2.2 tcp connection with client port specified FAILED ==== Contents of test case: # $x == "ready" at this point global port set sock [socket -myport $port 127.0.0.1 $listen] puts $sock hello flush $sock lappend x [gets $f] close $sock return $x ---- Test generated error; Return code was: 1 ---- Return code should have been one of: 0 2 ---- errorInfo: couldn't open socket: invalid argument while executing "socket -myport $port 127.0.0.1 $listen" ("uplevel" body line 4) invoked from within "uplevel 1 $script" ---- errorCode: POSIX EINVAL {invalid argument} ---- Test cleanup failed: error writing "stdout": broken pipe while executing "puts "[gets $file] $port"" (procedure "accept" line 3) invoked from within "accept sock5 127.0.0.1 63767" ---- errorInfo(cleanup): error writing "stdout": broken pipe while executing "puts "[gets $file] $port"" (procedure "accept" line 3) invoked from within "accept sock5 127.0.0.1 63767" while executing "close $f" ("uplevel" body line 3) invoked from within "uplevel 1 $cleanup" ---- errorCode(cleanup): NONE ==== socket-2.2 FAILED ==== socket-11.2 client specifies its port FAILED ==== Contents of test case: set s [socket -myport $port $remoteServerIP 2835] set r [gets $s] expr {$r==$port ? "ok" : "broken: $r != $port"} ---- Test generated error; Return code was: 1 ---- Return code should have been one of: 0 2 ---- errorInfo: couldn't open socket: invalid argument while executing "socket -myport $port $remoteServerIP 2835" ("uplevel" body line 2) invoked from within "uplevel 1 $script" ---- errorCode: POSIX EINVAL {invalid argument} ---- Test cleanup failed: can not find channel named "sock6" ---- errorInfo(cleanup): can not find channel named "sock6" while executing "close $s" ("uplevel" body line 2) invoked from within "uplevel 1 $cleanup" ---- errorCode(cleanup): TCL LOOKUP CHANNEL sock6 ==== socket-11.2 FAILED ---------------------------------------------------------------------- Comment By: Donal K. Fellows (dkf) Date: 2010-03-20 13:36 Message: {Copying in from that URL} [snip] /home/emiliano/src/tcl/unix/tclUnixSock.c: In function 'CreateSocketAddress': /home/emiliano/src/tcl/unix/tclUnixSock.c:1036: error: 'AI_ADDRCONFIG' undeclared (first use in this function) /home/emiliano/src/tcl/unix/tclUnixSock.c:1036: error: (Each undeclared identifier is reported only once /home/emiliano/src/tcl/unix/tclUnixSock.c:1036: error: for each function it appears in.) *** Error code 1 Stop. make: stopped in /home/emiliano/src/tcl/unix ---------------------------------------------------------------------- Comment By: Emiliano (egavilan) Date: 2010-03-20 13:33 Message: The rmax-ipv6-branch doesn't build on NetBSD. For the error message see http://paste.tclers.tk/2019 ---------------------------------------------------------------------- Comment By: Donal K. Fellows (dkf) Date: 2010-03-20 13:15 Message: c.f. rmax-ipv6-branch in CVS; assigning to person in charge of that ---------------------------------------------------------------------- Comment By: Donal K. Fellows (dkf) Date: 2009-06-08 18:32 Message: This was discussed in depth between myself and Reinhard Max at EuroTcl last week. We believe that it is possible to do it without significant changes *provided* people are content to let the OS pick which base protocol to use to communicate with the remote server; that can be done with no change to the scripted API (other than modifications to code that assumes that addresses are never in v6 format, of course...) We don't know if this is acceptable as an interim measure. (More control is desirable, but we couldn't get that sorted out in a reasonable timescale for 8.6) ---------------------------------------------------------------------- Comment By: Christoph (cero) Date: 2009-06-08 16:09 Message: Is there anyting new about IPv6 implementation to core tcl? I've seen a patch #713194 for that! ---------------------------------------------------------------------- Comment By: Alexandre Ferrieux (ferrieux) Date: 2008-09-23 19:19 Message: In the meantime to this implementation, what about [open |netcat6 r+] ? ---------------------------------------------------------------------- Comment By: Dennis Solari (dsolari) Date: 2008-09-23 18:08 Message: Needed IPv6: The client side is being requested by many of our customers since they are converting/converted to IP6. As an ISV using tcl we need to support it or convert to a different language. So why not release the client support in 8.6 even if the server support will not be ready or agreed to...? Using tcl socket calls as a client would meet our requirements for now from customers. ---------------------------------------------------------------------- Comment By: Donal K. Fellows (dkf) Date: 2007-08-31 10:34 Message: Logged In: YES user_id=79902 Originator: NO Additional notes: Use inet_ntop() instead of inet_ntoa() Use inet_pton() instead of inet_aton() Use getaddrinfo() instead of gethostbyname() and gethostbyaddr() ---------------------------------------------------------------------- Comment By: Donal K. Fellows (dkf) Date: 2007-08-31 10:14 Message: Logged In: YES user_id=79902 Originator: NO IPv6 client sockets look to be fairly straight-forward on Unix (all changes localized to tclUnixChan.c) but I'm less sure about server sockets. How practical is it to have a single socket listening for connects on both IPv4 and IPv6 sockets? (Thankfully, they both share the same port-space, so at least *that* won't get in our way!) ---------------------------------------------------------------------- Comment By: Jeremy Gibson (silpheed) Date: 2007-08-30 14:03 Message: Logged In: YES user_id=963204 Originator: YES Of course if it made it as a patch into 8.5, that would be very nice. And to clarify, the mandate is from OMB for government networks and apps to IPv6 compliant by sometime next year. (Can't remember the specific date.) ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=360894&aid=1782896&group_id=10894 |