Significant Pure Assembler Application In MASM ?
in [ASM]
|
Prev: aeBIOS Test Request
Next: CMOVcc vs. Jcc
From: Betov on 8 Oct 2007 10:20 Herbert Kleebauer <klee(a)unibwm.de> �crivait news:470A2203.157A4E51(a)unibwm.de: > Maybe you should ask in a linux group. I did this a year ago, but > without much success: > > http://groups.google.com/group/comp.os.linux.development.apps/browse_fr > m/thread/2c17c983165f27d4/2de83fa1b524e4bb?#2de83fa1b524e4bb :)) Never ask such things to Linux guys !!! :)) Betov. < http://rosasm.org >
From: santosh on 8 Oct 2007 10:24 Betov wrote: > santosh <santosh.k83(a)gmail.com> �crivait news:fed95t$b8s$1(a)aioe.org: > >> Yes. The ptrace system call. This provides only "bare minimum" >> facility. Building something like the RosAsm debugger on top of this >> is likely to be a lot of work. See the link: >> >> <http://www.linuxmanpages.com/man2/ptrace.2.php> > > Many thanks. Yes, this seems to be OK. If i go for it, it will > be all needed for having, at least, something similar to what > i used for the first RosAsm Debugger. > > Additional question: > > Is this Function available at any time (like any API)? It's a kernel API so it's available all the time. > Or will something be required before simply calling for it? If you want to debug a process which is not your child, (i.e. you did not explicitly fork and possibly exec), then you need to know the process's PID. But otherwise no, apart from the usual code to handle signals etc. > Or will a port of the given ptrace.h file be required? Since your going to call the kernel's API from assembly you need to know to which native types the various "C" types resolve into. The only way to do this is to examine the kernel header files, (in this case /usr/include/sys/ptrace.h), for your installation. Of course you need to know which registers to place the various arguments into, or on the stack perhaps. These details are likely to be the same for all kernels on a given architecture like the x86-32. The exact binary format for any structures would differ depending on the machine and the compiler. For example on my machine, (Pentium Dual Core with a 32 bit kernel), the types in ptrace.h are: The return value is 'long int' which on 32-bit x86 can be accommodated in a GPR. The first parameter for ptrace is an enum which is an int and so, again, the same as a GPR. The 'pid_t' and the 'void *' types are also storeble in a GPR.
From: Betov on 8 Oct 2007 10:31 santosh <santosh.k83(a)gmail.com> �crivait news:fed9tg$cv3$2(a)aioe.org: > Herbert Kleebauer wrote: > >> Betov wrote: >>> >>> Do you think a GTK bet could be wise? > > GTK is not a graphics subsystem. It's a toolkit. Wait a minute, here !!! What are you saying? Saying that if i assemble: call 'libgtk-2.0-0.gtk_label_set_text' the same way i would do under Windows (this example *exists*): call 'libgtk-win32-2.0-0.gtk_label_set_text' I would go nowhere, without the static lib ?! Nothing like a "libgtk-2.0-0" Dynamic Library under Linux ?! Betov. < http://rosasm.org >
From: rhyde on 8 Oct 2007 10:37 On Oct 8, 1:00 am, Betov <be...(a)free.fr> wrote: > Frank Kotler <fbkot...(a)verizon.net> écrivaitnews:FKcOi.6093$br2.2298(a)trndny03: > > This is a layer scale problem. What are the OSes for, if not for > providing layers to be re-used by the Applications? Progamming > something real with int 80 would be like doing so with the BIOS > interrupts, in my opinion. A terrific task that, considering the > number of volunteers, would take longer than the life of x86. You might pick up a book on Operating System design sometime. You seem to have a fundamental misunderstanding about the purpose of operating systems. They do not exist to provide application programs with a "library" of usable functions they can call. No doubt, your opinions have be colored by spending too much time with Windows (which, and this is poor design on the Windows' developers part, does incorporate some generic library code). > > OSes are there for providing that layer of Libraries which makes > the developing a possible thing. Not calling for them would have > no effect but dramaticaly slowing down a process, which is already > slow and long enough. Libraries are there for providing that layer of libraries. Libraries like GTK, for example. Why should the libraries be tied to a particular OS? If you write your code to use the libraries properly, you'll be able to port your code to, say, FreeBSD when you're done with Linux. > > I am always estonished when seeing a guy like you under-estimating > the quantity of work to be done. This is why you use pre-written libraries -- to reduce the amount of work. This is why you don't write "specific" code, to make it easy to port the project to a different OS. > Something like RosAsm was 10 years > of work for a significative number of contributors. Only my part > was a full time job for, at least 7 years, and Ludwig's one was around > 4 or 5 years. Well, I guess "two" is a significant number :-) Seriously, though, the reason RosAsm took so long is because you: 1) didn't know what you were doing 2) spent an inordinate amount of time on side trips (like your disassembler) rather than working on the *assembler*. 3) spent far too much time posting to various boards all over the internet trying to promote yourself. > So, where will the OSes be in 10 years, even when making > our lifes easier with the OS libs Functionalities?... This question > puzzles me. Look back 10 years and that will give you a good clue. > > Another point about Ubuntu: The reason why there exists an actual > hope with *Ubuntu* (in no way with the *Linuxes*) is because several > majors among the gears sellers choosed it, for being sold "out of > the box". This is not us who choose the future. This is the Market. > Personaly, i think i will make my choice in one week or so, when > the new 7.10 release will be official. I will then be able to > estimate the progress(es) they do, in between two releases... Wow, an amazing revelation. It's not important that you put features into RosAsm that the "average Joe" wants, but this is important for the OS you choose. Do you see the dichotomy here? hLater, Randy Hyde
From: Betov on 8 Oct 2007 10:38
santosh <santosh.k83(a)gmail.com> �crivait news:fed8me$9ht$1(a)aioe.org: > Are there many examples of the "GPL community" taking PD works and > putting them under GPL/LGPL after trivial modifications? Better than this: There are many examples of the ReactOS guys who took non-PD works, and who released them under GPL without any modification but arasing the MicroSoft Legal notes and replacing with: Author: Alex Ionescu. Recall: The defender of Master Randall Hyde, right here, at ALA. :) Betov. < http://rosasm.org > |