----- Original Message ----- From: "David Abrahams" To: Sent: Monday, August 09, 2004 6:20 AM Subject: [Boost-users] Re: RTTI for VC 6

This isn't a Boost.Build specific issue. Maybe you could repost it to the Boost developers' list?

I am sorry to bother you. I won't make suggestions like that any more.

----- Original Message ----- From: "David Abrahams" To: Sent: Monday, August 09, 2004 11:21 AM Subject: [Boost-users] Re: RTTI for VC 6

Touchy, touchy!

It's no bother, and I love seeing that kind of suggestion. I'm just suggesting that you post it in a more appropriate place so that the person(s) who ought to respond to it can see it.

I am subscribed to too many mailing lists now, and I have been subscribed to many others that I can't keep up with. Most of the messages in high-volume mailing lists are not relevant to me so I waste a lot of time if I try to keep up with them. So yes, I am very reluctant to subscribe to (yet another) another mailing list just to post a message and then receive another flood of messages that don't concern me. The (main) Cygwin mailing list is an example of a high-volume list. I wrote a few paragraphs intended to help other understand how to install it and I put that in my web site. It was not my intent to leave it there permanently, but the only comment I got back was that they did not want to have the documentation spread out in many places. So experiences such as that has contributed to my attitude.

----- Original Message ----- From: "David Abrahams" To: Sent: Monday, August 09, 2004 6:55 PM Subject: [Boost-users] Re: RTTI for VC 6

If you want to make a difference, post where those who care and can respond will see it. Or don't, but don't expect results.

The Boost "What to do about Boost bugs" page says to submit "a bug report to either boost-users list, boost mailing list, or our bug tracking facility at SourceForge" and that "either of the mailing lists is a preferred way". Am I correct that this list is the boost-users list? I am not sure whether my suggestion qualifies as a bug, but the Boost page about suggestions has similar instructions. Results? Am I expecting results? No, I tried to (or should have?) make it clear in my message that I posted the message to help others. I tried to make it clear that I already know the answer; I was posting an answer. If a person must use a time-consuming procedure to make a simple suggestion, then it is not likely they will. If this applies to bugs also, then the software is more likely to remain buggy.

----- Original Message ----- From: "David Abrahams" To: Sent: Monday, August 16, 2004 3:33 AM Subject: [Boost-users] Re: RTTI for VC 6

I took this:

...

That problem is probably a known problem, but something such as the following can be added somewhere to not allow a VC 6 program to even ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ compile if the RTTI option is off. I think it could make things a little easier for some beginners.

# if !defined(_CPPRTTI) # error "Boost requires use of RTTI" # endif

to be a request that Boost be patched to prevent compilation without RTTI. It didn't seem like a suggestion for beginners to do themselves; a fteralltheycanturnonRTTIjustaseasily.DidI misinterpret your post?

Right, it is not something a beginner should expect to do. However my point, whether valid or not, is that I am a beginner, and it is much, much more convenient for me to send to this list and monitor this list for replies. As far as being able to turn on RTTI, yes it is easy; however if a person is not familiar with that option then they can get a strange and/or irrelevant runtime error, with no clue to the cause. It is the relative difficulty in understanding the problem that is the issue. If it is not an easy fix to the libraries to issue a warning or an error message, then I can understand that it is not worth doing. Something I did not mention that probably is easy and worthwhile is to update the relevant portion of the documentation that lists the idiosyncrasies of VC 6; I don't remember seeing it at least.

Samuel wrote:

----- Original Message ----- From: "Peter Dimov" To: Sent: Wednesday, August 11, 2004 3:40 AM Subject: Re: [Boost-users] RTTI for VC 6

...

Samuel wrote:

...

VC 6 has an option for Run-Time Type Information (RTTI) that is normally off. I was able to compile and execute a sample program that had a problem. I noticed that Boost uses RTTI (such as the dynamic_cast operator), so I checked to see if the option was on; it was not. I turned the option on and the program worked.

That problem is probably a known problem, but something such as the following can be added somewhere to not allow a VC 6 program to even compile if the RTTI option is off. I think it could make things a little easier for some beginners.

# if !defined(_CPPRTTI) # error "Boost requires use of RTTI" # endif

Some parts of Boost need RTTI, some do not. It is not reasonable to reject a program that doesn't ever invoke a dynamic_cast on a polymorphic type (which is a level 1 warning IIRC) just because it happens to include a boost header.

I don't understand this. I did not reply earlier since I assumed it does not matter, and I still assume that whoever might be concerned about this can be the one to decide the relevance. However if you are saying that a warning will result from not using dynamic_cast then you did not read my message. The problem is that the compiled program did not work; it produced unexpected results (I forget the details). The unexpected results could easily cause someone else more work to solve.

What Mr. Dimov is saying is that it is impossible to place the code you would like to use above in Boost because it would affect other Boost libraries which do not need RTTI to work. Furthermore if it were placed in the header file of a library which did need RTTI to work, just including that header file, without actually using code in it that needed RTTI such as dynamic_cast, would cause an immediate compile time error.

To put it bluntly, comments such as this accomplish nothing except dump emotions at the expense of others. In the Netbeans mailing list, there are a few people that enjoy dumping on Windows in the manner that people dump on VC here. Please, please will someone please tell these people that comments such as this don't accomplish anything useful for Boost? Or am I being unreasonable? ----- Original Message ----- From: "Victor A. Wagner Jr." To: Sent: Sunday, August 15, 2004 12:40 AM Subject: Re: [Boost-users] RTTI for VC 6

to put it bluntly, it's not boost, it VC++6.0 (the plethora of errors) upgrading would save you and a lot of others a lot of problems

Victor A. Wagner Jr. wrote:

I'm on record here as saying we should save the world a fortune in time and quit supporting such a poor compiler. (now THAT is flame bait, and I don't give a <insert favorite expletive>.... Microsoft VC++6.0 is NOT a reasonable compiler..... within months of it's release, the ISO standard

I agree to this point, however ....

In my opinion, the only reason that boost hasn't officially told everyone that VC++6.0 is no longer supported is because Dave Abrahams apparently has a customer that "cannot update". It costs a small

... as I have checked in archives our old discussion (thread "1.31.0 regressions" started on boost developer list on 2004-02-10), David was not the only one opting for continuing support for old compilers. I also want to remind you that Borland compilers are at the same (or even lower) level of "conformity" like MSVC6, and they even did not released any newer version to replace BCB6, while Microsoft released two and is pushing its customers to upgrade (MSVC6 mainstream support is ending Sept 30 this year, while extended paid support is available for one more year, see http://support.microsoft.com/default.aspx?scid=fh;[ln];LifeDevToolFam ) B.

At Sunday 2004-08-15 05:31, you wrote:

Note that the original response was that I should post my comments to the other forum yet the Boost web site says that this forum is one of two forums to post bugs and suggestions in.

Another response indicated that my suggestion is not possible. I did not say that it is possible, but I doubt that it can be said that it definitely is not possible.

You are definitely being unreasonable about saying that I "dumps on the warning messages". Try to show us where I did that; you can't.

from the message I originally responded to: "There might have been a warning, but Boost produces so many other warnings that I missed the warning if there was one."

When you say "why hasn't the OP updated??" you ask that as if you don't really want to know the answer. I get the impression that whatever I say, you will still insist that I am being unreasonable.

I think it's unreasonable to expect anyone to continue to have to write code for a crippled product. It's been over 2 years since vs.net was released. How long do _you_ think it's reasonable for have to be restricted to a once existing product? I note that you haven't said why you haven't/can't/won't upgrade, so it's hard to tell whether I'll think _you_ are being unreasonable. It's very likely that I'll consider the management who made the decision to not upgrade are the culprits.

Is that the official position?

I'm not an "official"

Is it likely that support of VC 6 will be dropped?

According to another message in this thread, Microsoft will drop support in 46 days.

If that is not true, then the comments below that are critical of me and indicating that support should be expected to be stopped seems unreasonable to me.

I have no idea when the library authors will quit supporting an old compiler

Again, look to see where I dumped on the warning messages. You can't find that and therefore you need to state that it is incorrect to say I did.

I posted where I thought you "dumped" on the warnings. the exact substring to which I refer is: "but Boost produces so many other warnings" Now if you're not using vc6, I apologize. If you _are_, then VC6 has been known for some time to be particularly inept at handling templates (for example it publishes warnings about fully expanded typenames exceeding 256 characters). I was attempting to point that out and then I was chided for "dumping on VC". I have in the past, (VC6) but I don't believe I've complained much about 7, 7.1, or 8. I'm currently running regression tests on 7.1 and 8 so I _do_ have VC, it's just 6.0 that I (rightfully, IMO) think is a PoS.

----- Original Message ----- From: "Victor A. Wagner Jr." To: Sent: Sunday, August 15, 2004 3:42 AM Subject: Re: [Boost-users] RTTI for VC 6

...

you're being unreasonable. some guy came in here with a KNOWN poor and out of date compiler and dumps on the warning messages he gets and implies that the folks a boost should/could do something about it (Microsoft HAS done something about it, they're issued TWO releases since then.....why hasn't the OP updated??) . I'm on record here as saying we should save the world a fortune in time and quit supporting such a poor compiler. (now THAT is flame bait, and I don't give a <insert favorite expletive>.... Microsoft VC++6.0 is NOT a reasonable compiler..... within months of it's release, the ISO standard was voted (of course they didn't conform, I believe that any rational person looking at ALL the behavior of Microsoft back in 1997-1998 can ONLY come to the conclusion that they were trying to "kill" C++). Microsoft STILL has NOT updated the STL that they ship with VC++6.0...do you want me to go on?)

In my opinion, the only reason that boost hasn't officially told everyone that VC++6.0 is no longer supported is because Dave Abrahams apparently has a customer that "cannot update". It costs a small fortune to try to make everything work with VC++6.0 and it is NOT worth the effort).

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"

whatever ALL I said was that the plethora of errors was VC++6.0 NOT boost and you go make a "federal case" over it. get OVER it.... VC6 is a very bad compiler it was upgraded by the manufacturer 2 years ago your continued use of it is foolish now THERE is a personal insult...just for you...original my message was solely about the compiler why you had to bitch about it is certainly beyond MY comprehension. Oh, you _still_ haven't told me WHY you stick with such a poor tool. At Sunday 2004-08-15 16:24, you wrote:

----- Original Message ----- From: "Victor A. Wagner Jr." To: Sent: Sunday, August 15, 2004 6:41 AM Subject: Re: [Boost-users] RTTI for VC 6

...

from the message I originally responded to: "There might have been a warning, but Boost produces so many other warnings that I missed the warning if there was one."

That is a statement of fact. That is not what I call dumping. What I call dumping is being critical in a manner that expresses emotions.

exactly WHAT is emotional about this: "to put it bluntly, it's not boost, it VC++6.0 (the plethora of errors) upgrading would save you and a lot of others a lot of problems"

...

I think it's unreasonable to expect anyone to continue to have to write code for a crippled product.

Note what was said about Borland. I think there are other examples also, but I can't find the relevant page that rates the various compilers. Do you know for a fact that VC 6 is worse than the others, or are you discriminating against VC 6 for some other reason? How well does the IBM compiler perform?

I don't know, nor is it relevant to THIS discussion

Another relevant question is how well does the Gnu compiler rate when tested for support of Windows? As far as I know, it does not support Windows (runtime) as well as VC does. So before you criticize use of VC, please be sure that there are reasonable alternatives.

there are, and I even mentioned them vs.net and vs.net2003 (the latter being preferred)

I think that if the Gnu compiler was as useful for Windows development as VC is, then there would be much more use of the Gnu compiler.

irrelevant

...

It's very likely that I'll consider the management who made the decision to not upgrade are the culprits.

You are making an assumption that is incorrect.

You haven't answered by question. You ass/u/med that I didn't care, and if you want to quote out of context you can make it sound like anything.

...

According to another message in this thread, Microsoft will drop support in 46 days.

I said nothing about Microsoft support of VC 6; someone else did, but I did not.

that IS relevant to whether VC6 is worth continuing to use!!

...

I posted where I thought you "dumped" on the warnings. the exact substring to which I refer is: "but Boost produces so many other warnings"

Again, that is not what I call dumping. I should have used another term or method of saying what I meant to say. I really do think that your previous comments expressed your emotions about Microsoft and unfairly discriminated against Microsoft in a manner that is not relevant here.

You're the one that started this... I made a statement of fact...see the above (COMPLETE, not out of context) original reply

If the Microsoft VC 6 complier is the most difficult one for the Boost libraries to support, then it is reasonable to discriminate against it. If not, then it is unproductive to discriminate against just Microsoft.

you just don't get it, do you. You bitch about a problem, I tell you that upgrading would save everyone a lot of time and you go ballistic.

There is plenty to criticize Microsoft for, so please don't think I consider them to be perfect or totally innocent. One big problem is that in order to upgrade to .Net we must purchase a Visual Studio update, which is a significantly greater cost. Microsoft often churns us for updates that have much more than we really need. Further discussion of that could be done in another message or forum or whatever. The important thing is that you can't accomplish anything useful by criticizing my use of VC 6.

Like hell I can't accomplish anything. your COMPILER is the source of your difficulties. End of story. and in 46 days, even Microsoft wont' support you.

Such comments can't improve the situation. Whatever you are trying to accomplish, there are more productive ways to do it.

apparently there aren't any more useful ways with you. You seem stuck in a rut of using a 7 year old compiler and won't upgrade. I'll say it again upgrading will save everyone a lot of time and trouble. statement of fact no emotion

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

Victor A. Wagner Jr. http://rudbek.com The five most dangerous words in the English language: "There oughta be a law"

----- Original Message ----- From: "Peter Dimov" To: Sent: Monday, August 16, 2004 3:49 AM Subject: Re: [Boost-users] RTTI for VC 6

Samuel wrote:

...

I don't understand this. I did not reply earlier since I assumed it does not matter, and I still assume that whoever might be concerned about this can be the one to decide the relevance. However if you are saying that a warning will result from not using dynamic_cast then you did not read my message. The problem is that the compiled program did not work; it produced unexpected results (I forget the details). The unexpected results could easily cause someone else more work to solve.

We can't address the issue without knowing about the details you've forgotten.

I think there is a more productive way to say what you want to say. Instead of wording it as a criticism, it is just as easy to make it a more direct request. I did not bother to remind myself of the details when I wrote the above comment because it was not sufficiently relevant. I can provide the details of course. I will do that in a separate message.