| comp.lang.c++@googlegroups.com | Google Groups |  |
| Unsure why you received this message? You previously subscribed to digests from this group, but we haven't been sending them for a while. We fixed that, but if you don't want to get these messages, send an email to comp.lang.c+++unsubscribe@googlegroups.com. |
- Visual C++ Version 6 (Visual Studio 98) - 25 Updates
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 13 04:31PM -0700 On Saturday, September 13, 2014 7:05:23 PM UTC-4, Ian Collins wrote: > Waxing lyrical about a notoriously awful compiler on a forum frequented > by people such as me who had to suffer its awfulness tends to give the > opposite impression! What is it that makes it awful? I recognize it has limited C++ features, and from a standards-looking point of view it is not up to snuff ... but it does work, is very fast, and supports edit-and-continue in near real-time. If someone could explain to me how to make the more modern toolsets work faster with edit-and-continue I would stick with the current tools. It currently takes Visual Studio 2008 between 5 and 15 seconds to apply code changes using Windows Server 2003 in a VM on Linux Mint 64-bit machines with 4- or 6-cores with the guest given 2 GB of RAM. On my native Win7 machine using the same VS2008 version, it takes between 3 and 8 seconds. If I could get that down to 1 or 2 seconds I would continue using my existing tools. But the slowdown is very counter-productive as it puts pauses in my workflow, trains of thought. I wind up waiting for a tool to complete its task, rather than me continuing on once I've completed mine. This winds up being not only the slowest part of the equation during those types of debugging changes, but one which decreases my own performance as I get continually annoyed at the tool I'm needing to use. > If I'd known there was someone willing to pay for a copy, mine wouldn't > have gone to landfill a decade ago... Mine either. :-) I had a copy at some point, and also got rid of it in favor of newer tools. But over time I've seen the performance of the more modern tools decrease even as our processor speeds, RAM and hard disk speeds have increased. And whereas I enjoy working with more modern tools, because significant speed has been lost in its edit-and-continue abilities, there is a tradeoff I'm willing to endure to gain that speed back. There are a couple things I've never understood in certain developers: (1) That edit-and-continue is undesirable. It was the response I got when back in 2010?? I posted about adding edit-and-continue to GCC. Someone pointed out that Apple had added fix-and-continue to a version, but it was not put back in the main line?? And I've seen on more modern GCC roadmaps where it's been mentioned ... but I haven't seen it yet. (2) That older tools are inherently less than newer ones. In my experience this is often the case, but not always, and on some tools it's not even close. The CodeView 3.x debugger I used to use allowed for two-monitor debugging. It was amazingly fast. I miss that debugger, and actually strove with my Debi Debugger to have some of its functionality and speed. I actually wrote a text-based version of Debi at one point so I could debug my 512 byte floppy bootup code in a real debugger. :-) In my experience, there are certain tools that were just great. And while there are many facets of newer things which are also great, they don't negate, or aren't always sufficiently better than the old to justify the switch (especially on software, which doesn't get old and wear down as material things do). I appreciate your response, Ian. I don't understand why people view some things so harshly. If the type of programming I do, for example, is mostly C, with only a very few C++ things thrown in, and I gain in debugging speed the almost real-time application of code changes ... how is this is any way a bad thing for me? And whereas it may not be something you would like to use with the way you code, is your opinion of something formed exclusively by your personal use of the product? Or is there room for recognizing that other people may use tools in different ways, and therefore see value in products that others don't, and vice-versa? Best regards, Rick C. Hodgin |
| Ian Collins <ian-news@hotmail.com>: Sep 14 01:16PM +1200 Rick C. Hodgin wrote: > and from a standards-looking point of view it is not up to snuff ... but > it does work, is very fast, and supports edit-and-continue in near > real-time. It was notoriously buggy, fixes didn't appear and it had limited support for the C++98 standard. > mine. This winds up being not only the slowest part of the equation > during those types of debugging changes, but one which decreases my own > performance as I get continually annoyed at the tool I'm needing to use. I guess if you are working with unoptimised C98 code, the smaller memory footprint of an older tool would probably suit your needs. I remember complaining vociferously but unsuccessfully to the developers of my tools when they replaced the native GUI for a (spit) Java monstrosity which used an order or magnitude more RAM. >> If I'd known there was someone willing to pay for a copy, mine wouldn't >> have gone to landfill a decade ago... <snip> > pointed out that Apple had added fix-and-continue to a version, but it > was not put back in the main line?? And I've seen on more modern GCC > roadmaps where it's been mentioned ... but I haven't seen it yet. Fix and continue has limited utility. It is only useful for unoptimised C. As soon as there's hit of optimisation, especially inline code, or a template it becomes very difficult to write and cumbersome to use. > strove with my Debi Debugger to have some of its functionality and speed. > I actually wrote a text-based version of Debi at one point so I could > debug my 512 byte floppy bootup code in a real debugger. :-) Older tools only support language standards current when they were written. Even last year's version of the debugger I use (dbx) won't support applications built for C++11 due to ABI incompatibilities. > don't negate, or aren't always sufficiently better than the old to justify > the switch (especially on software, which doesn't get old and wear down > as material things do). There are, if you are willing to live within their constraints. > is any way a bad thing for me? And whereas it may not be something you > would like to use with the way you code, is your opinion of something > formed exclusively by your personal use of the product? I guess you are happy to work with code that older tools can handle. But have you ever asked yourself if your tools and way of working are holding you back? I would hate to be stuck with C89 plus snippets of C++, I much prefer to have the full language toolkit at my disposal. I sometimes come across some of my old code from the 90s and think "now that would have been so much easier with modern C++". I even find the same with last year's pre-C++11 code! -- Ian Collins |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 13 07:16PM -0700 On Saturday, September 13, 2014 9:16:30 PM UTC-4, Ian Collins wrote: > complaining vociferously but unsuccessfully to the developers of my > tools when they replaced the native GUI for a (spit) Java monstrosity > which used an order or magnitude more RAM. I wouldn't mind having even a console-based debugger, so long as it was feature-rich and similar to a GUI (meaning you have defined windows, drag-and-drop, could mouse over variables and have their values shown, click on windows, edit code, etc.). I had originally planned for my debugger for Visual FreePro to be text-based in a GUI framework ... but have since moved it to use the underlying Visual FoxPro-like forms and classes, so it can be maintained by the community more easily. > Fix and continue has limited utility. It is only useful for unoptimised > C. As soon as there's hit of optimisation, especially inline code, or a > template it becomes very difficult to write and cumbersome to use. Agreed. But the purpose of edit-and-continue (or fix-and-continue) is when you are writing new code, trying to get it to work. In fact, I've been able to enter the debugger with an algorithm only defined, and then code the entire body while debugging. The lifecycle is edit-and-continue during development and initial testing, then more traditional models once you reach initial maturity on the algorithm's development and testing. > But have you ever asked yourself if your tools and way of working are > holding you back? I would hate to be stuck with C89 plus snippets of > C++, I much prefer to have the full language toolkit at my disposal. I don't think I know enough of the newer C++ features to have any sort of idea about what I might be missing out on. But in my experience with C++ to date, I have not seen great advantage in those features I have looked at. I have had to learn many new things to get what I've viewed as only mild conveniences in syntax, yet use of such features increases the knowledge requirements of developers who will come along after and maintain the code. My view may be atypical. > I sometimes come across some of my old code from the 90s and think "now > that would have been so much easier with modern C++". I even find the > same with last year's pre-C++11 code! There are some C++ features I intend to code into my RDC compiler, but they are very few beyond the basic class and some helper extensions through the use of casks. Over the long term I've found that coding things more explicitly, with simpler algorithms, proper comments, and other basic things, it makes the maintenance an order of magnitude easier and faster than code which requires a greater degree of knowledge and/or expertise to wield. Best regards, Rick C. Hodgin |
| Ian Collins <ian-news@hotmail.com>: Sep 14 03:25PM +1200 Rick C. Hodgin wrote: > On Saturday, September 13, 2014 9:16:30 PM UTC-4, Ian Collins wrote: No I didn't, I wrote it on Sunday afternoon! > The lifecycle is edit-and-continue during development and initial > testing, then more traditional models once you reach initial maturity > on the algorithm's development and testing. I think I've suggested to you before that test driven development would be a natural progression for you to try. I used to rely heavily on fix and continue in dbx, but now I just use the debugger for production core files or when I'm sufficiently intrigued as to why a test fails to debug it rather than simply back out the change that broke it. > viewed as only mild conveniences in syntax, yet use of such features > increases the knowledge requirements of developers who will come along > after and maintain the code. I've had a concentrated dose of C++11 in the past couple of months due to beta testing a new compiler and working on a new project with like minded colleagues. I can't see myself going back. -- Ian Collins |
| Christian Gollwitzer <auriocus@gmx.de>: Sep 14 09:10AM +0200 Am 13.09.14 19:20, schrieb Rick C. Hodgin: > [...] > The old tools were freeer. I am choosing them for those reasons, even if > they don't live up to modern standards in coding. It would be interesting to see you argue with RMS SCNR, Christian |
| Jorgen Grahn <grahn+nntp@snipabacken.se>: Sep 14 07:26AM On Sat, 2014-09-13, Jorgen Grahn wrote: >> this one aspect: nearly instantaneous edit-and-continue debugging. > Fine, but please note that this pretty alienates you and > comp.lang.c++. We cannot help you, and you cannot help us. To clarify that a bit: as I understand it, the language supported by VS6 is sufficiently different from C++98 that most of us will have forgotten how to use it -- or are trying to forget. Especially if it didn't have the standard library containers -- all my code relies heavily on those. Then eventually I suppose it will also become harder to discuss C++98, as more and more people here get used to C++11. Never used VS6 myself, but I remember that gcc for C++ was in a bad shape in the late 1990s. And I remember Windows users searching for decent STL implmentations (STLPort?) well into the past decade. /Jorgen -- // Jorgen Grahn <grahn@ Oo o. . . \X/ snipabacken.se> O o . |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Sep 14 10:14AM +0100 On 14/09/2014 03:16, Rick C. Hodgin wrote: > The lifecycle is edit-and-continue during development and initial > testing, then more traditional models once you reach initial maturity > on the algorithm's development and testing. Bullshit mate; edit-and-continue should only be occassionally used whilst debugging an issue not used all the time during the creation of new code or refactoring old code. Personnally I have NEVER used edit-and-continue and have had no problems and can develop new code quickly. Basically YOU ARE DOING IT WRONG and YOUR DEVELOPMENT PROCESS "LIFECYCLE" SUCKS. /Flibble |
| Paavo Helde <myfirstname@osa.pri.ee>: Sep 14 04:15AM -0500 "Rick C. Hodgin" <rick.c.hodgin@gmail.com> wrote in > features, and from a standards-looking point of view it is not up to > snuff ... but it does work, is very fast, and supports > edit-and-continue in near real-time. I all depends on priorities. One of my first priorities is to write portable code which can be compiled on different platforms. I guess there is a subset of MSVC6 which is more or less portable, but this is too limited for my taste (cannot use for loop scope variables, cannot access std::map across DLL borders, etc, etc). The edit-and-continue thing probably worked well enough only for pretty C-style code. I recall vaguely that half of the time it did not work for my C++ programs and I gave up on using it because I could never be sure what code it actually produced. In the current MSVC++ versions there are the "minimal recompile" and "incremental linking" features which could be used for speeding up the debugging process. Cheers Paavo |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 05:46AM -0700 On Sunday, September 14, 2014 3:10:51 AM UTC-4, Christian Gollwitzer wrote: > > The old tools were freeer. I am choosing them for those reasons, even > > if they don't live up to modern standards in coding. > It would be interesting to see you argue with RMS RMS and I have had discussions over various things. He is the main reason I founded the Liberty Software Foundation. I also desire to give to the people "free software" (what I call "liberty software"), but to do so upon a foundation built around the Lord Jesus Christ. I offer up my skills, talents, abilities to Him, and unto men, and I do so specifically by name because He is the true rock of our (believer's) lives, the bedrock foundation of everything. For example, I came very close in May/June/July 2012 to continuing development on the HURD kernel. I discussed several aspects of this with RMS, and he actually advised me not to put my efforts into completing the kernel because they already had a kernel in Linux. He advised me to work on a few GNU projects that he saw as fundamental in moving forward, namely Gnash (GNU's Flash alternative). Through those discussions I discovered some things about RMS and his personal beliefs which made me take a completely different path with my life, and to abandon any work I would've done on the HURD, for GNU, or for the FSF. I've been working on Visual FreePro ever since. In any event ... I continue to pray for Richard. I would like to see him come to faith and believe so we could work together as brothers. I have admired his skills and purpose. In the movie Revolution OS you see a lot of his philosophy conveyed very succinctly. It was that movie which first brought me understanding of the differences between open source and copyleft protected free software. But until his goals are founded upon serving the Lord, and in that way serving other men, it is something I cannot be a part of. > SCNR, > Christian What does "SCNR" mean? My brain keeps interpreting it as "SCRN" (short for SCREEN). :-) Best regards, Rick C. Hodgin |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 05:51AM -0700 On Sunday, September 14, 2014 3:26:21 AM UTC-4, Jorgen Grahn wrote: > Never used VS6 myself, but I remember that gcc for C++ was in a bad > shape in the late 1990s. And I remember Windows users searching for > decent STL implmentations (STLPort?) well into the past decade. I understand. I haven't done much C++ programming. The bulk of the programming I've done in C++ has used the class, and very little more than that. And the class was only used for encapsulation and ease of portability. I have not seen advantages in C++ code which make it sufficiently worthwhile to learn the extra bits required to use them, wield the extra bits continuously for maintenance, and require the extra complexity in the compilation tools to compile them. I find the straight-forward conveyance of C to be more desirable, and especially so once the baseline algorithms are created which handle the dozen or so routine things required for iteration, navigation through data, etc. But, I recognize this is a C++ group so I will leave that discussion there. :-) Best regards, Rick C. Hodgin |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 05:59AM -0700 On Sunday, September 14, 2014 5:14:36 AM UTC-4, Mr Flibble wrote: > quickly. Basically YOU ARE DOING IT WRONG and YOUR DEVELOPMENT PROCESS > "LIFECYCLE" SUCKS. > /Flibble (1) If you would like me to respond to you in the future, avoid using profanity. Your choice. (2) You are one of the most hostile people I've ever encountered on a newsgroup. I feel sorry for you dwelling in all that hate and rage. (3) If you'd like to be set free of that hate and rage, please contact me and I will teach you about Someone who can fill you with love, joy, peace, and hope. You'll be amazed at what you're missing. And what the enemy is robbing you of in this life. (4) I've used the software lifecycle I indicate successfully since the late 1990s. Many developers see no value in edit-and-continue. I could not live without it. I am using Microsoft-based tools today because of it. I am building edit-and-continue into my own RDC tools. And once I get them sufficiently developed, I will migrate everything away from those tools to operate entirely in my own. (5) If edit-and-continue doesn't work for you, you do not need to use it. As I say, for my coding style, it is the single greatest asset I possess. At least an order of magnitude beyond any other debugger feature. Best regards, Rick C. Hodgin |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Sep 14 02:01PM +0100 On 14/09/2014 13:46, Rick C. Hodgin wrote: > software. But until his goals are founded upon serving the Lord, > and in that way serving other men, it is something I cannot be a > part of. So you are a troll then, a god bothering troll. Your god doesn't exist mate: evolution is proof of this. /Flibble |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Sep 14 02:04PM +0100 On 14/09/2014 13:59, Rick C. Hodgin wrote: > profanity. Your choice. > (2) You are one of the most hostile people I've ever encountered on a > newsgroup. I feel sorry for you dwelling in all that hate and rage. I am not angry mate and I only hate deliberate idiocy. > me and I will teach you about Someone who can fill you with love, joy, > peace, and hope. You'll be amazed at what you're missing. And what the > enemy is robbing you of in this life. To repeat my reply else-thread: you are a god bothering troll mate and evolution is proof that your god doesn't exist. /Flibble |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 06:06AM -0700 On Sunday, September 14, 2014 9:01:17 AM UTC-4, Mr Flibble wrote: > So you are a troll then, a god bothering troll. Your god doesn't exist > mate: evolution is proof of this. There is scientific proof which negates the existence of evolution. Look into the complexity of DNA, the information contained within, the variation of creation across the species, the diversity in application of the available fundamental components, the underlying biochemical hardware, the transport mechanisms, error correction, the various operating systems, the software protocols in place, the interaction between the disparate and specialized cells. Evolution does not exist. Creation does. Here's a movie which will help you understand: https://www.youtube.com/watch?v=00vBqYDBW5s I am not a troll. But rather I am a Christian. And as I go about my life, I bring things back into their correctly centered, God-focused nature. I found things upon the One who created all things. Best regards, Rick C. Hodgin |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Sep 14 03:10PM +0100 On 14/09/2014 14:06, Rick C. Hodgin wrote: >> So you are a troll then, a god bothering troll. Your god doesn't exist >> mate: evolution is proof of this. > There is scientific proof which negates the existence of evolution. What proof is that mate? > variation of creation across the species, the diversity in application > of the available fundamental components, the underlying biochemical > hardware, the transport mechanisms, error correction, the various Non of which contradicts evolution mate. > operating systems, the software protocols in place, the interaction > between the disparate and specialized cells. Gibberish mate. > Evolution does not exist. Creation does. The evidence for evolution is legion mate whilst the evidence for divine creation is totally absent. > Here's a movie which will help you understand: > https://www.youtube.com/watch?v=00vBqYDBW5s I don't watch god bothering videos mate; life is too short and there is no afterlife. > I am not a troll. But rather I am a Christian. And as I go about my > life, I bring things back into their correctly centered, God-focused > nature. I found things upon the One who created all things. Sure you are a troll mate; one of the many trolls who are also Christian. Your god is the god of Abraham mate and evolution is proof that that particular god does not exist. /Flibble |
| Paavo Helde <myfirstname@osa.pri.ee>: Sep 14 09:24AM -0500 "Rick C. Hodgin" <rick.c.hodgin@gmail.com> wrote in > personal beliefs which made me take a completely different path with > my life, and to abandon any work I would've done on the HURD, for GNU, > or for the FSF. Seems like a good decision. We definitely do not need HURD kernel developed with MSVC6 and in the edit-and-continue style programming. Cheers Paavo |
| Christian Gollwitzer <auriocus@gmx.de>: Sep 14 05:08PM +0200 Am 14.09.14 16:24, schrieb Paavo Helde: >> or for the FSF. > Seems like a good decision. We definitely do not need HURD kernel developed > with MSVC6 and in the edit-and-continue style programming. LOL. Maybe, this is what RMS *really* wanted to express: "oh please stay away from my HURDy kernel! I'll propose some application level stuff to him that will better fit his skills." Christian PS: To each one his own belief - but no need to indulge it onto others. My name might be foolish;) |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 08:15AM -0700 On Sunday, September 14, 2014 10:10:23 AM UTC-4, Mr Flibble wrote: > On 14/09/2014 14:06, Rick C. Hodgin wrote: > > There is scientific proof which negates the existence of evolution. > What proof is that mate? A person must be willing to hear the truth ... to hear the truth. Best regards, Rick C. Hodgin |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 08:16AM -0700 On Sunday, September 14, 2014 10:24:58 AM UTC-4, Paavo Helde wrote: > with MSVC6 and in the edit-and-continue style programming. > Cheers > Paavo LOL! Good one. :-) Very funny and clever. Of course, I would've used "proper" tools for HURD development. That period was an interesting period in my life. It has sent me on a course which is different than anything I previously imagined. It is also still unfolding. Best regards, Rick C. Hodgin |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 08:19AM -0700 On Sunday, September 14, 2014 11:09:14 AM UTC-4, Christian Gollwitzer wrote: > LOL. Maybe, this is what RMS *really* wanted to express: "oh please stay > away from my HURDy kernel! I'll propose some application level stuff to > him that will better fit his skills." Possible. I never thought of that. However, it was his first response when I contacted him about the possibility of picking up development on the HURD kernel. > Christian > PS: To each one his own belief - but no need to indulge it onto others. > My name might be foolish;) The name itself isn't foolish, but only the application of the individual (with regards to the thing). Best regards, Rick C. Hodgin |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 08:22AM -0700 On Saturday, September 13, 2014 11:25:32 PM UTC-4, Ian Collins wrote: > > On Saturday, September 13, 2014 9:16:30 PM UTC-4, Ian Collins wrote: > No I didn't, I wrote it on Sunday afternoon! Ian, do you have any videos of presentations you've given? Or documents outlining classes or sessions you've taught which are available to the general public? And is your name pronounced eye or ee? Best regards, Rick C. Hodgin |
| Geoff <geoff@invalid.invalid>: Sep 14 08:58AM -0700 On Sun, 14 Sep 2014 08:15:10 -0700 (PDT), "Rick C. Hodgin" >> > There is scientific proof which negates the existence of evolution. >> What proof is that mate? >A person must be willing to hear the truth ... to hear the truth. Pot, Kettle, Black. |
| "Rick C. Hodgin" <rick.c.hodgin@gmail.com>: Sep 14 09:12AM -0700 On Sunday, September 14, 2014 11:58:38 AM UTC-4, Geoff wrote: > On Sun, 14 Sep 2014 08:15:10 -0700 (PDT), "Rick C. Hodgin" > >A person must be willing to hear the truth ... to hear the truth. > Pot, Kettle, Black. Truth speaks with one voice, and one voice only. It is why you must be willing to hear the truth to hear the truth. It is easy to listen to other voices because there are millions of them. But there is only one voice which is Truth. Best regards, Rick C. Hodgin |
| Scot Dellinger <scotyde42@gmail.com>: Sep 14 04:19PM Geoff wrote: >>> What proof is that mate? >>A person must be willing to hear the truth ... to hear the truth. > Pot, Kettle, Black. Sounding echo inhere? |
| Mr Flibble <flibbleREMOVETHISBIT@i42.co.uk>: Sep 14 05:29PM +0100 On 14/09/2014 16:15, Rick C. Hodgin wrote: >>> There is scientific proof which negates the existence of evolution. >> What proof is that mate? > A person must be willing to hear the truth ... to hear the truth. I thought you relied on faith? Faith is belief without evidence so how can you know what you believe is true? If your position is the truth then there must be mountains of evidence backing it up mate. How about you give me just ONE piece of evidence that shows that your god exists? /Flibble |
| You received this digest because you're subscribed to updates for this group. You can change your settings on the group membership page. To unsubscribe from this group and stop receiving emails from it send an email to comp.lang.c+++unsubscribe@googlegroups.com. |
No comments:
Post a Comment