| Print-Server - DE | Print-Server - EN | Linux-Firewall | |
| OS-Dev:Multitasking | BlueIllusion OS | Vibrant Soundbridge | PC BOX |
| Neuigkeiten |
|---|
| 04.12.2011: Hail, fellow internetters, how you're doing? I hope, you're fine, for I experience mixed stuff these days. At my workplace, things been goin' pell-mell the last few days, and I 've had to endure quite some bashing do to taking a set of seemingly undesired decisions. Heck, since what day, architecture decisions have to be desirable. They have to work, that is all there is to it. Be it as is, the dear colleagues have decided to meet in my absence and after that they been a-stomping straight to their boss. Oh my. And then, things been a-goin' bust big time. As if I'm a moron, four people at the same time started talking to me about how my architecture has to be talked throu' for they are but so not content with it. That been the time when I lost my temper a bit and told them off for doing this kinda stuff in my back. I simply felt like being talked over and exposed as Mr. Big Moron. Nevertheless, their claims are valid, despite the way they 've delivered them. I have thought about this before and I have thought about it over the weekend in a thread-like way, just to get the obvious solution delivered by /dev/brain again. Even if we are using a kind of rich-client solution, these rich-client components have to use back-end-components located at the Oracle WebCenter Server in a J2EE-Container. These Components offer all sorts of backend services to the clients - as well as communication to SOA-Services in a well-ordered and trace-able way. No Rich-Client will ever talk to any SOA-Service directly. As for "Rich-Client-Components talking to SOA-Services is not SOA" - this is Nonsense big time. SOA means to split up an application into reusable loosely coupled components whose interfaces are exposed in form of web-services - or something different. By no means is a SOA Service Landscape to be solely called by the but so famous Process Engine. Well - we gonna find a solution for that. Apart from that - I have finally found it! BlueIllusionOS: Heck! Searching for THIS bug took me weeks. I merely got a page fault and no indication as to where it happens. Thus I've wasted lengths of time looking in all possible and inpossible places. Spent time debugging the pipefs driver. No. The bug has not been there. Whilst adjusting the file.c-file to adapt to the new terminalfs-structures, I have fixed around with the do_exit-method. And in this method, I've inadvertently deleted the pipe-nodes upon exit of any program invoked by the shell (or some other program-starting stub). I only got the hint to that today after I 've finally recognized a repeatable pattern: the page fault hit upon exit of the invoked program. I've modified ls to wait for keyboard input ere exit. I 've got a niiice directory listing. Then, upon key press, the system has crashed. Well, after that, it's been merely a matter of comparing the do_exit procedures of file.c in the workspace and the first revision (which I know is working) from my repository. et voila - there I've found the culprit. Some little fixing, and my sweetie works again as supposed. Now, what remains is the ongoing debugging and fixing of pty, which is not yet working as I want it to. Stay tuned  |
| 07.10.2011: Oh my gosh, how time flies. There is progress everywhere. The thing with my PhD-Programm is continuing, and in the meantime, I have not only worked further on the pty-stuff for my os but also - look at this! - created a painting. after about 12 years of merely keeping to little drawings on occasion. Doodles, little comics and sorta. Obviously this has kept me in training somehow. Here is the actual painting:  and here is, for comparison, a painting I have done ere starting with the HTL:  12 years between the first and the second. Well - there has been a lot going 'round all the time. And still is. ;) BlueIllusionOS Heck, yeah. I have implemented most parts of the pty-driver in the past months. I 've also debugged parts of it already. Especially the debugging of the on-delete-handling has revealed quite some bugs - which I've fixed on the go. Now, I'm taking a break, ere I delve into merging the pty-driver with the "xterm" of this OS. Stay tuned. It doesna look like it, but there's a lot going on behind the scene. |
| 28.02.2011: Oh my, how time flys, how one forgets to keep still sometimes and take a breath. Some things are too much. Love has strolled by and kicked my ass very hard. I jumped and looked around, utterly bamboozled, for I have already forgotten how this feels: to be in love. So, important moments have passed while I still been busy grasping the whole thing. Well - who knows what it is good for. But be it as it is, I'm still happy. Happy, that emotion has come back to me with a flow of inspiration and glowing images of how a possible future might look like. Emotions of such force and intensity that they leave me looking as if clubbed over the head with a bat of oak wood. There must be an end to misery. There must be an end to all the sadness because I've found my Star again. The doctorate study I 've mentioned has gone bust. The boss of the Fakultaet fuer Informatik at TU Wien has refused to approbate the claim for a doctorate my professor and I have already signed and prepared. That is it, then. Said boss has an issue with students who have attended a university of applied science. Mebbe we ain' students at all, heh? Especially those who do this in the evenings, after work, huh? Such an over-sophisticated, spineless pussy. *grrrr* But things haven't failed yet. There is still hope, and I do not intend to give up. On an other note: There has something, quite something happened with my toy operating system, so here is our long awaited BlueIllusion OS Item: BlueIllusionOS: Hey! You wouldna believe, what a whole heck of work it is just to create a pseudoterminal-driver. First of all, this thing needs two parallel paths: one for the screen, the other one for the keyboard-stuff. Second: this thing needs a part in the File System too, so I have implemented a terminalfs-component for the FS-Server. This smoothly blends into the overall-structure of the FS-Server and the VFS stuff. Maan, I'm so excited. Sure - this still needs heavy debugging. Still, some small nitty gritty is yet to be done. And for someone not into OS-Stuff this might look like a fucking midget: nothing to even fuzz about. But I'm into OS-Stuff and I'm lookig forth to these incredible, long missed YEAH!! stay tuned :) |
