From yoni at rabkins.net Thu Jun 5 18:46:46 2008 From: yoni at rabkins.net (Yoni Rabkin) Date: Thu, 05 Jun 2008 21:46:46 +0300 Subject: [movitz-devel] [patch] fix and clean movitz-dump-image Message-ID: <87lk1j7m95.fsf@rabkins.net> Hello, I'm using SBCL 1.0.5, The attached patch does the following: * Fix "C-c d" (formerly movitz-dump-image) * Replace both `movitz-dump-image' and `movitz-dump-image-and-qemu' with a single function, namely `movitz-dump'. The only backward-breakage I know of is that now neither "C-c d" or "C-c C-d" accept a filename prefix argument. -------------- next part -------------- A non-text attachment was scrubbed... Name: movitz-slime-cleaner-dump.patch Type: text/x-diff Size: 3834 bytes Desc: fix and clean movitz-dump* URL: -------------- next part -------------- -- "Cut your own wood and it will warm you twice" From ffjeld at common-lisp.net Sat Jun 7 12:58:31 2008 From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld) Date: 07 Jun 2008 14:58:31 +0200 Subject: [movitz-devel] Re: [patch] fix and clean movitz-dump-image References: <87lk1j7m95.fsf@rabkins.net> Message-ID: <2h4p85ieq0.fsf@vserver.cs.uit.no> Yoni Rabkin writes: > Hello, I'm using SBCL 1.0.5, > > The attached patch does the following: [..] Checked in now. Thanks. -- Frode Vatvedt Fjeld From antoniosantoniadis at gmail.com Tue Jun 10 16:59:45 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Tue, 10 Jun 2008 18:59:45 +0200 Subject: [movitz-devel] My gsoc project Message-ID: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> Hello everyone, Yesterday I started work on my google summer of code project. After 1.5 days fighting with various emulators, i think bochs is working now as it should, so I have a testing environment. Running function ip4-test though leads to some error, that seems to lie in decode-keyargs-default in primitive-functions.lisp. As this function is written in assembly and I have no idea about assembly at all, I would be really happy if someone could help me out with this, or at least give me a hint. http://www.wetsand.org:8080/pictures/screen.png <--- here I uploaded a screenshot with the error messages. Regards, Antonis From ffjeld at common-lisp.net Thu Jun 12 12:55:15 2008 From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld) Date: 12 Jun 2008 14:55:15 +0200 Subject: [movitz-devel] Re: My gsoc project References: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> Message-ID: <2hzlpqhky4.fsf@vserver.cs.uit.no> "Antonis Antoniadis" writes: > http://www.wetsand.org:8080/pictures/screen.png <--- here I uploaded a > screenshot with the error messages. Hi Antonis, sorry about the delay. The error means that the function (setf ether-source) received bad keyword arguments (from format-ethernet-packet, the backtrace says). And indeed the keywords were missing in format-ethernet-packet. I've checked in a fix, but I don't have time to do any testing until this evening. Frode -- Frode Vatvedt Fjeld From antoniosantoniadis at gmail.com Thu Jun 12 13:39:33 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Thu, 12 Jun 2008 15:39:33 +0200 Subject: [movitz-devel] Re: My gsoc project In-Reply-To: <2hzlpqhky4.fsf@vserver.cs.uit.no> References: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> <2hzlpqhky4.fsf@vserver.cs.uit.no> Message-ID: <36271ae80806120639m64112fefu617d960c86cce61f@mail.gmail.com> Hi Frode, and thanks for the reply. I found this out today, so i changed the source, also i changed (setf (ether-source source start :source-start source-start) source... to (setf (ether-source packet start :source-start source-start) source... and don't get any more error messages. It seems though that it gets in an infinite loop in the last assert of ip4-init. Currently I'm trying to find out why this happens. I would also like to ask, is there some keyboard shortcut that causes an interrupt in movitz, like C-c C-c in slime? So I could perhaps backtrace and find out where it gets stuck/infinite-loops (right now my movitz mode isn't working as it should and i have to add print lines and create-image every time). Thanks, Antonis On 12 Jun 2008 14:55:15 +0200, Frode Vatvedt Fjeld wrote: > "Antonis Antoniadis" writes: > > > http://www.wetsand.org:8080/pictures/screen.png <--- here I uploaded a > > screenshot with the error messages. > > > Hi Antonis, sorry about the delay. The error means that the function > (setf ether-source) received bad keyword arguments (from > format-ethernet-packet, the backtrace says). And indeed the keywords > were missing in format-ethernet-packet. I've checked in a fix, but I > don't have time to do any testing until this evening. > > Frode > > > -- > Frode Vatvedt Fjeld > > _______________________________________________ > movitz-devel site list > movitz-devel at common-lisp.net > http://common-lisp.net/mailman/listinfo/movitz-devel > From antoniosantoniadis at gmail.com Thu Jun 12 14:23:39 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Thu, 12 Jun 2008 16:23:39 +0200 Subject: [movitz-devel] Re: My gsoc project In-Reply-To: <36271ae80806120639m64112fefu617d960c86cce61f@mail.gmail.com> References: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> <2hzlpqhky4.fsf@vserver.cs.uit.no> <36271ae80806120639m64112fefu617d960c86cce61f@mail.gmail.com> Message-ID: <36271ae80806120723m5b66aa0frcc6a9f0320476724@mail.gmail.com> I think I know why the infinite loop happens. I found out that in this specific case i can break the loop by hitting space, and that it loops because it's unable to resolve the ip. On 6/12/08, Antonis Antoniadis wrote: > Hi Frode, and thanks for the reply. I found this out today, so i > changed the source, also i changed (setf (ether-source source start > :source-start source-start) source... > to (setf (ether-source packet start :source-start source-start) > source... and don't get any more error messages. > It seems though that it gets in an infinite loop in the last assert of > ip4-init. Currently I'm trying to find out why this happens. > I would also like to ask, is there some keyboard shortcut that causes > an interrupt in movitz, like C-c C-c in slime? So I could perhaps > backtrace and find out where it gets stuck/infinite-loops (right now > my movitz mode isn't working as it should and i have to add print > lines and create-image every time). > > Thanks, > > Antonis > > > On 12 Jun 2008 14:55:15 +0200, Frode Vatvedt Fjeld > > wrote: > > "Antonis Antoniadis" writes: > > > > > http://www.wetsand.org:8080/pictures/screen.png <--- here I uploaded a > > > screenshot with the error messages. > > > > > > Hi Antonis, sorry about the delay. The error means that the function > > (setf ether-source) received bad keyword arguments (from > > format-ethernet-packet, the backtrace says). And indeed the keywords > > were missing in format-ethernet-packet. I've checked in a fix, but I > > don't have time to do any testing until this evening. > > > > Frode > > > > > > -- > > Frode Vatvedt Fjeld > > > > _______________________________________________ > > movitz-devel site list > > movitz-devel at common-lisp.net > > http://common-lisp.net/mailman/listinfo/movitz-devel > > > From ffjeld at common-lisp.net Thu Jun 12 16:29:09 2008 From: ffjeld at common-lisp.net (Frode Vatvedt Fjeld) Date: 12 Jun 2008 18:29:09 +0200 Subject: [movitz-devel] Re: My gsoc project References: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> <2hzlpqhky4.fsf@vserver.cs.uit.no> <36271ae80806120639m64112fefu617d960c86cce61f@mail.gmail.com> <36271ae80806120723m5b66aa0frcc6a9f0320476724@mail.gmail.com> Message-ID: <2hve0ehb1m.fsf@vserver.cs.uit.no> "Antonis Antoniadis" writes: > I think I know why the infinite loop happens. I found out that in > this specific case i can break the loop by hitting space, and that > it loops because it's unable to resolve the ip. Without an interrupt-driven keyboard driver it's not possible to have a reliable C-c thingy, but this is definitely something I want to have soon too. There is one mechanism in there: the GC will poll for ESC keypresses and break if there is one. Actually, there's one more possibility: the :more top-level command (try i.e. ":more (loop (print 'hello))" at the REPL) will allow you to break at each screenful of output. -- Frode Vatvedt Fjeld From yoni at rabkins.net Thu Jun 12 17:01:37 2008 From: yoni at rabkins.net (Yoni Rabkin) Date: Thu, 12 Jun 2008 20:01:37 +0300 Subject: [movitz-devel] My gsoc project In-Reply-To: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> (Antonis Antoniadis's message of "Tue, 10 Jun 2008 18:59:45 +0200") References: <36271ae80806100959q1af46ce9wbc978f5e952cf5d6@mail.gmail.com> Message-ID: <87bq268u4u.fsf@rabkins.net> "Antonis Antoniadis" writes: > Hello everyone, > > Yesterday I started work on my google summer of code project. After > 1.5 days fighting with various emulators, i think bochs is working now > as it should, so I have a testing environment. Could you send me the command line you use for bochs? I might use it as the default for built-in bochs support for movitz-slime.el. -- "Cut your own wood and it will warm you twice" From antoniosantoniadis at gmail.com Mon Jun 16 21:46:02 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Tue, 17 Jun 2008 00:46:02 +0300 Subject: [movitz-devel] very strange problem Message-ID: <36271ae80806161446o7d1f8ce9y51a733c6c0be6c8@mail.gmail.com> Hello everyone, I suspect that ip4 is working now. At least I don't get any errors, and it continues looping. I wanted to print some stuff during the loops, to get reassured that everything works, but I have some very strange problem: Compiling the file ip4.lisp (either as part of (create-image) or as part of (movitz-compile-file)), turns the file into binary (or at least it looks like binary), and it seems that the old file gets compiled (as none of my messages get printed). I have no idea what could cause this, does anyone of you? I suspect an emacs problem, I can post the binary-file if that helps. thanks a lot, Antonis From antoniosantoniadis at gmail.com Sun Jun 22 07:57:17 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Sun, 22 Jun 2008 10:57:17 +0300 Subject: [movitz-devel] 2 short, ip/udp related questions Message-ID: <36271ae80806220057k492ab005t228000f42b451b66@mail.gmail.com> Hello everybody, I have two short questions: 1. What is the way to send packets of some given protocol to movitz? 2. I also asked this in the blog, do you think that after udp is working, it's worth the time/effort to create some way to transfer changes to a running movitz using udp? On one hand it will make development easier, but on the other hand I'll (try to) write tcp anyway also. Regards, Antonis From antoniosantoniadis at gmail.com Sun Jun 22 14:04:36 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Sun, 22 Jun 2008 17:04:36 +0300 Subject: [movitz-devel] 2 short, ip/udp related questions In-Reply-To: <20080622103523.429F62A3C@fdy2.demon.co.uk> References: <36271ae80806220057k492ab005t228000f42b451b66@mail.gmail.com> <20080622103523.429F62A3C@fdy2.demon.co.uk> Message-ID: <36271ae80806220704u46d3cbf4k75997ba9ecba687d@mail.gmail.com> On 6/22/08, Robert Swindells wrote: > > You wrote: > >1. What is the way to send packets of some given protocol to movitz? > > > There isn't really a single way to generate protocol traffic. > > Which protocol do you want to try ? > I wanted to try icmp, udp and tcp, to see if function ip input works for all cases. Most important though for now would be udp. Should I implement tftp before udp? > > >2. I also asked this in the blog, do you think that after udp is > >working, it's worth the time/effort to create some way to transfer > >changes to a running movitz using udp? On one hand it will make > >development easier, but on the other hand I'll (try to) write tcp > >anyway also. > > > Where is your blog ? The link you sent me just gives a blank page. > I am not sure whether I had to change the port for some reason after sending the link, although I don't think so. Anyway, going to www.recursively.org, and pressing the blog link from there, always works, except in the seldom cases that the server is down. > I would think about making movitz be in control of any transfer > protocol and to use something that already exists so that you only > need to write one side of it. > > The simplest protocol to implement would be TFTP. > Do I understand this correct, that you mean implement tftp to transfer the changes, and don't bother more with udp (for this purpose of course)? > > Robert > Thanks a lot for the fast reply, Antonis From rjs at fdy2.demon.co.uk Sun Jun 22 16:07:18 2008 From: rjs at fdy2.demon.co.uk (Robert Swindells) Date: Sun, 22 Jun 2008 17:07:18 +0100 (BST) Subject: [movitz-devel] 2 short, ip/udp related questions In-Reply-To: <36271ae80806220704u46d3cbf4k75997ba9ecba687d@mail.gmail.com> (antoniosantoniadis@gmail.com) References: <36271ae80806220057k492ab005t228000f42b451b66@mail.gmail.com> <20080622103523.429F62A3C@fdy2.demon.co.uk> <36271ae80806220704u46d3cbf4k75997ba9ecba687d@mail.gmail.com> Message-ID: <20080622160718.4DC522B00@fdy2.demon.co.uk> Antonios Antoniadis wrote: >On 6/22/08, Robert Swindells wrote: >> >> You wrote: >> >1. What is the way to send packets of some given protocol to movitz? >> >> >> There isn't really a single way to generate protocol traffic. >> >> Which protocol do you want to try ? >> >I wanted to try icmp, udp and tcp, to see if function ip input works >for all cases. Most important though for now would be udp. Should I >implement tftp before udp? The IP layer doesn't care what is inside a message, it just needs to have a correct checksum. As long as you can see correct data being received then I would move on. TFTP runs on top of UDP. See RFC 1350 and movitz/losp/lib/net/tftp.lisp. >> >2. I also asked this in the blog, do you think that after udp is >> >working, it's worth the time/effort to create some way to transfer >> >changes to a running movitz using udp? On one hand it will make >> >development easier, but on the other hand I'll (try to) write tcp >> >anyway also. >> >> Where is your blog ? The link you sent me just gives a blank page. > >I am not sure whether I had to change the port for some reason after >sending the link, although I don't think so. Anyway, going to >www.recursively.org, and pressing the blog link from there, always >works, except in the seldom cases that the server is down. Ok, I can view it now. >> I would think about making movitz be in control of any transfer >> protocol and to use something that already exists so that you only >> need to write one side of it. >> >> The simplest protocol to implement would be TFTP. >> >Do I understand this correct, that you mean implement tftp to transfer >the changes, and don't bother more with udp (for this purpose of >course)? You need to start thinking about the layers that will be in a full stack. There are pieces of most of them already within Movitz, but they are not all linked together, and may not be currently written in a way which allows them to work together. Maybe think about how a user would want to make use of a feature such as reading a file from a remote system. From them running '(load "foo")' on a lisp source file what would need to take place for it to work ? Which streams would you implement and link together ? One version of this example could use TFTP as the protocol to talk to a file server. Robert Swindells From antoniosantoniadis at gmail.com Fri Jun 27 21:24:35 2008 From: antoniosantoniadis at gmail.com (Antonis Antoniadis) Date: Sat, 28 Jun 2008 00:24:35 +0300 Subject: [movitz-devel] tftp Message-ID: <36271ae80806271424wa92e87v382335e4b158cff3@mail.gmail.com> Hello again. The last couple of days I'm working on tftp, in order to become able to transfer files from the host machine to movitz and evaluate them. As can be seen in the attached screenshot, I can only send write requests, but don't get responses. I'm not sure I understand the reason exactly but in my way of thinking: To write on the host you need write permissions, which means you have to login there as a user with write permissions. Though tftp (for simplicity reasons i guess) doesn't support that functionality. Am I wrong about all this? I've now tried sending to my developing machine, to my brothers computer, and to lispnyc (where heow kindly provided me with a login, and tcpdump access for this reason), the result is always the same. What is the way to get an ACK as a response for my WRQ from the host, so that I can actually get a port number and start sending data? Antonis -------------- next part -------------- A non-text attachment was scrubbed... Name: tftp-new.png Type: image/png Size: 62387 bytes Desc: not available URL: From pjb at informatimago.com Sat Jun 28 21:25:29 2008 From: pjb at informatimago.com (Pascal J. Bourguignon) Date: Sat, 28 Jun 2008 23:25:29 +0200 Subject: [movitz-devel] tftp In-Reply-To: <36271ae80806271424wa92e87v382335e4b158cff3@mail.gmail.com> References: <36271ae80806271424wa92e87v382335e4b158cff3@mail.gmail.com> Message-ID: <18534.44105.653508.291197@hubble.informatimago.com> Antonis Antoniadis writes: > Hello again. > The last couple of days I'm working on tftp, in order to become able > to transfer files from the host machine to movitz and evaluate them. > As can be seen in the attached screenshot, I can only send write > requests, but don't get responses. I'm not sure I understand the > reason exactly but in my way of thinking: > To write on the host you need write permissions, which means you have > to login there as a user with write permissions. Though tftp (for > simplicity reasons i guess) doesn't support that functionality. Am I > wrong about all this? It depends on the tftp server. Usually, it's enough to chmod 777 the directory where tftp keeps its files. -- __Pascal Bourguignon__ http://www.informatimago.com/ Un chat errant se soulage dans le jardin d'hiver Shiki