ELF loading
in [ASM]
|
From: //o//annabee on 30 Aug 2007 04:08 P� Thu, 30 Aug 2007 01:14:55 +0100, skrev JDavison <jd(a)swoosh.net>: > //\\o//\\annabee wrote: > >> Dear Sevag. You could at least try to sound convincing. > > Is that another one of your lame jokes, boy? ooooh oooh ohh yes that was a friendy as well. I wish Sevag would take a course in markedting so that there would be something to fight. :)
From: //o//annabee on 30 Aug 2007 04:36 P� Thu, 30 Aug 2007 05:36:01 +0100, skrev rhyde(a)cs.ucr.edu <rhyde(a)cs.ucr.edu>: > On Aug 28, 9:08 am, CodeMonk <jas...(a)yahoo.com> wrote: >> Betov wrote: >> >> >> Then why not add static linking to your assembler? You may have no >> >> reason to use it, but it will certainly please some >> >> > Because i am not used to work for implementing stupidities, >> > even when when people want it, clown. >> >> This deficiency could be addressed by wrapping the requisite LIB in a >> DLL and using the LoadLibrary and GetProcAddress API functions. An >> annoyance to be sure, but it's the only viable work-around I see. > > There are many bigger problems than just the annoyance. > First of all, there is the version problem (also known as "DLL Hell"). > Second, there is the issue of size: loading a DLL gets you *all* the > routines, not just the ones you actually call. Wrong. Only code that runs stays in memory. Rest is pushed to disk, if loaded at all. Paging works for code as well as for other memory. Otherwise it is a lame argument that is dead 10 years ago. Windows itself loads a LOT more dlls for your program in anyway. Nomatter the program, many DLLs not related to the app is loaded in the background. I wish it did not, because this is one reason why some asm programs arent as instant as they should be when executing. You can verify what I am saying using a good PE spy like for instance processexplorer. Many asm program under windows, runs with FAR more code and data that is due to unneeded OS stupidity, that the actuall asm program ever needs or called for. Added to that is that a DLL with asm code, will be FAR smaller then a DLL for a HLL. The ratio will be around 1/200 in the average case. 2 mega HLL dll >> 90kb asm dll. And this ration seems to double each 5 years or so. > Of course, there is also the issue that functions in a DLL must be > called indirectly. Often, this is done by linking in an extra JMP > instruction (admittedly, not necessary, but many tools wind up doing > this). This is a NON issue. A call takes, (Paging ignored) next to no time. At most a few cycles. This is simply irrelevant when considering the task that do take time. If you ever saw this argument somewhere, be assured it is a selling argument only. It has absolutly no impact on the perceived speed of running code at all. Typical tasks that takes lot of time is drawing, sorting seeking, heavy I/O and stuff like that. > Also, DLLs carve up your address space inconveniently. A problem if > you have huge memory allocations or you use (large) memory-mapped > files. True, it steals some of the adresses. But again, not in any way like what the OS does, behind our back. The OS is now, since NT2000 brutally stealing lots of space for unwanted, uncalled for DLLs. This problem has exploded much worse in later OSes. > Static linking avoids all of these problems. > hLater, > Randy Hyde
From: //o//annabee on 30 Aug 2007 04:38 P� Thu, 30 Aug 2007 05:44:51 +0100, skrev rhyde(a)cs.ucr.edu <rhyde(a)cs.ucr.edu>: > On Aug 28, 8:22 pm, Evenbit <nbaker2...(a)charter.net> wrote: >> On Aug 28, 12:16 pm, JDavison <j...(a)swoosh.net> wrote: >> >> >> >> > Everyone knows you are guilty of stealing the name for your relatively >> > user-less (i.e. "useless") product. You're not fooling anyone, clown. >> >> You are wrong on that note. Nobody stays around when a child is >> crying about how another kid stole his favorite toy. >> >> So, if you are confident in your claims to the "RosAsm" trademark, why >> don't you take Rene to court? You would get the chance to prove to a >> jury how Rene's distribution of RosAsm somehow delutes your income >> from a product carrying the same name. > > Joel doesn't own the name. So he couldn't take Rene to court. > OTOH, the ReactOS team *has* asked Rene to stop using their project's > name in his product name. The ethical thing for Rene to do would be to > switch to a different name. Idiotic argument. Those guys who said that needs to have a brain transplant. > > Maybe this Linux thing of Rene's is a sea change. Perhaps we'll start > hearing about a "Wine-Asm" before too much longer. :-) > hLater, > Randy Hyde >
From: Evenbit on 30 Aug 2007 03:55 On Aug 30, 12:50 am, "rh...(a)cs.ucr.edu" <rh...(a)cs.ucr.edu> wrote: > > > If Rene didn't > > "steal" the name from *you*, then *who*, *where*, *what*, and *when* > > did it come from? ' > > who: The ReactOS team Where is their competing "assembler?" > where: here, as well as on the ReactOS board. > what: that Rene has misappropriated their project name as a mechanism > to suck some of their "glory" for his own self. Is this "glory" something that can be quantified? This claim looks *very* weak. > when: for the past couple of years. > > >Who is being hurt? > > Rene has caused *considerable* grief to the ReactOS project. What is the old saying? "If it doesn't kill you, then it makes you stronger." Ya know, there once was a dude by the name of Torvalds who cloned an OS. Perhaps the ReactOS team should seek his advice on how to quickly deal with such issues so they can go back to focusing on the items that are really important? > > > Where is the crime? > > Crime? > None at all, really. Just a matter of inethical behavior and hypocrasy > on Rene's part. "inethical behavior" -- bring him before the Ethics Board! "hypocrasy" -- that's more common than the air we breath! What's new?? > Why keep bringing it up? Because Rene is the guy who > goes around claiming I "stole" FASM (with Tomasz' permission, I might > add), all the while refusing to change the name he stole from ReactOS. > If Rene is going to bring up this mythical "theft" of FASM, it's only > fair to point out the *actual* misappropriation of the ReactOS name > that he has made. Would "WindowsAsm" also be considered a "misappropriation?" > > > Why should > > this story be of interest to any of a.l.a. readers? > > Why are you not all over Rene for the "theft of FASM issue?" He would just label me "minion" and be done with it. He only plays the games that move toward *his* goals. Nathan.
From: Evenbit on 30 Aug 2007 04:12
On Aug 30, 12:44 am, "rh...(a)cs.ucr.edu" <rh...(a)cs.ucr.edu> wrote: > > Joel doesn't own the name. So he couldn't take Rene to court. > OTOH, the ReactOS team *has* asked Rene to stop using their project's > name in his product name. The ethical thing for Rene to do would be to > switch to a different name. Well that is extremely easy to comply with: RosAsm = Rene's Open Source Assembler If you've payed attention to the "About" and "FAQ" pages on websites, you'd know that acronyms change their meaning quite a bit in this business. I'd estimate maybe 2 or 3 dozen projects have done so in the span of the last fifteen years. > > Maybe this Linux thing of Rene's is a sea change. Perhaps we'll start > hearing about a "Wine-Asm" before too much longer. :-) Well he *does* whine a good deal about your work. But that goes back to my original point to Joel -- when a child won't stop crying, people tend to leave the room. So I do not believe that NG readers are really paying attention to what Rene says. The only ones that do are those who are looking for an argument... or they are fellow whiners themselves -- "misery loves company!!" Nathan. |