Assembler/Compiler testing
in [ASM]
|
From: rhyde on 11 Mar 2006 15:37 Evenbit wrote: > > That's not surprising. The RosAsm assembly language has been very > > static over the years. A few minor additions here and there, but > > nothing amazing. You spend the vast majority of your time working on > > things that are not part of the assembler itself. > > erm... Isn't this an 'apples to oranges' comparison? No. > > I believe that Intel, and now AMD, are the designers of the language. > If they didn't make any major 'code breaking' changes to the language, > then why should Rene? First of all, an assembler processes quite a few more elements than machine instructions. Pseudo-opcodes, directives, stuff like that. And should Intel and AMD decide to add new instructions, that tends to break programs written in the past that just happened to use those mnemonics as macros or other symbols. Rene claims this never happens with RosAsm programs, but that's just because he only uses the 25 or so instructions that he's comfortable with. > Since the chip designers don't often break the > op-codes (they just add new ones), then a simple assembler never needs > to be updated (except for adding support for new opcodes, features, > etc., or correcting a bug). And guess what? That's a good example of times when existing code breaks. > > Your HLA has so many advanced features that it is a distinct language > of its own -- not a simple assembler. And that doesn't make it *any* different than RosAsm with respect to new features trouncing on old programs. The only "advantage" RosAsm may have here is that it doesn't explicitly encode machine mnemonics as *reserved* words (just as keywords). But while this doesn't create problems when old programs use mnemonics for their own purposes (e.g., macros), it *does* create problems when a user-written macro conflicts with a new mnemonic that comes along. Again, the main reason the RosAsm user base may not have experienced this problem is because they tend to be a less sophisticated bunch and don't often use the newer instructions. > The changes you make are likely > to be in those 'advanced features' areas (library, macros, etc.). And what that has to do with breaking the *assembler*, and why this would be specific to HLA, I cannot fathom. The same problem exists with *any* language (not even assembly language). > It > seems clear to me that maintaining and testing the HLA toolchain is an > entirely different ballgame (and a much bigger job) than the minor > updates that Rene makes to his PE Editor (a very advanced hex-editor > featuring built-in assembler and some advanced plugins). Formal testing is pretty much the same *regardless* of the size of the tool. There is no question that HLA v1.x could use a *lot* more of it. But at least it has had some. RosAsm hasn't. And the fact that the assembler component of RosAsm is fairly simply does not imply at all that he could skip that step. Cheers, Randy Hyde
From: rhyde on 11 Mar 2006 15:43 o//annabee wrote: > >> > > > > Compiling the same thing over and over again doesn't prove anything. > > You seem to know as much about software testing as Rene. > > Which is to say, nothing at all. > > :)) > > Was it not you who said? : > > "Of course, for each release I usually (unless I'm in a hurry or forget) > compile every HLA sample source file I have on Webster to see if I've > broken anything major. While this doesn't test the *entire* language or > compiler by any stretch of the imagination, it does catch stupid stuff. > > And making sure that the example programs still compile is kind of > important. " > > You claim: "its important", yet "it proves nothing". Boy, are you really that stupid? Forgive me for being patronizing and explaining some as simple as this. I do not "recompile the example programs" 100s of times a day for the fun of it. That would prove nothing. What I *will* do is run the regression test suite I've put together *whenever I create a new release of HLA*. > > If it is important, then it must also be important to know that when using > RosAsm, all your code gets tested in this way _ALL_ the time. Many many > times each day on my PC. Complete waste of time. Many times a day doing the same exact thing is a complete waste of time if there aren't changes to RosAsm between each compile. Perhaps you have other side effects in mind (such as working on your code), but the second recompile proves nothing that the first didn't also prove with respect to testing RosAsm. > > And for you assertment that it "proves nothing".... "nothing" isn't quite correct. The n-1 times a day you recompile your program don't show anything more than the first compilation did. Unless (which is a distinct possibility) RosAsm is non-deterministic and randomly executes defective code. > > Well, it proves that the code compiles..... so that assertion is wrong. > > Just like you :) Maybe when you've finished learning how to program in assembly, several years from now, you might want to turn your attention to software testing and quality assurance. You obviously have a lot to learn about the subject. > > And you are wrong because such a successful compilation DOES NOT catch > "stupid stuff". It ONLY catches syntax errors. I'm sure from your perspective, this is what you think. > > You have no base. And the sooner you shut up the better. Because even a > beginner asm programmer can arrest you on half the things you parrot from > the scriptkiddies posting to slashdot. :-) Whatever you say. Your own posts insult you more than I ever could. So why should I try to compete with that? Cheers, Randy Hyde
From: Betov on 11 Mar 2006 16:59 rhyde(a)cs.ucr.edu ?crivait news:1142109441.601119.15190 @e56g2000cwe.googlegroups.com: > The only "advantage" RosAsm > may have here is that it doesn't explicitly encode machine mnemonics as > *reserved* words (just as keywords). Whereas, in the case of your shitty HLL Pre-Parser, we can keep sure that, even without any new introduction of Mnemonic conflicting with your no-end list of reserved words, it will still be incompatible from release to release, as shown since day one of your swindling attempt. :) Betov. < http://rosasm.org >
From: o//annabee on 11 Mar 2006 17:39 P? Sat, 11 Mar 2006 21:43:47 +0100, skrev <rhyde(a)cs.ucr.edu>: > > o//annabee wrote: >> You claim: "its important", yet "it proves nothing". > > Boy, are you really that stupid? :) Yes. My stupidity levels are about what they wore previously, and is likly to stay about the same levels for years to come. Unlike you, were you seem able to have a trajectory leading on. For me, its no problem. I am able to write fully fledged asm application, even at my level, and you, who are supposedly not yet fully stupid, are _unable_ to write full fledged asm applications. So if you are the measure of stupidity, which seems to be the case, then I am happy to be what ever I am, as long as it is diffrent from what you are. > Forgive me for being patronizing and explaining some as simple as this. [lol....rest of post deleted in altruism] > Cheers, > Randy Hyde
From: sevagK on 11 Mar 2006 23:52
Betov wrote: > rhyde(a)cs.ucr.edu écrivait news:1142109441.601119.15190 > @e56g2000cwe.googlegroups.com: > > > The only "advantage" RosAsm > > may have here is that it doesn't explicitly encode machine mnemonics as > > *reserved* words (just as keywords). > > Whereas, in the case of your shitty HLL Pre-Parser, we can > keep sure that, even without any new introduction of Mnemonic > conflicting with your no-end list of reserved words, it will > still be incompatible from release to release, as shown since > day one of your swindling attempt. > > :) > > Betov. > Absurd. I've updated at least 10 versions perhaps more without having to make any changes to my existing code base. The only release that some code may break is the next one coming up, but that's trivial and HLA's error detection will quickly isolate the few places where this needs change. You identified the wrong guy as the swindler, Rene. -sevag.k www.geocities.com/kahlinor |