|
|
Read it, Pete, just read it ... I'm not saying that everything Ladislav has written is rubbish, I've even upvoted him, but this article presents material that is plain wrong.
|
|
|
|
|
Espenn - I tried to read it, honestly I did, but the formatting threw me so much that I gave up. I appreciate that English isn't his first language but, even taking that into account, I couldn't follow it because paragraph breaks happened in odd places.
|
|
|
|
|
Find the section starting with I call it Hurray stl vector<everything> trend.
Do you want destructors to be called on an object when a pointer to that object goes out of scope?
shared_ptr provides something resembling this functionality at the library level.
|
|
|
|
|
Espen Harlinn wrote: Do you want destructors to be called on an object when a pointer to that object
goes out of scope?
I'm still confused by this article. Is he stating that he wants the language to cope with this by default? Is he asking for some form of reference counted/garbage collected tidy up?
When I was coding C++ (a long time ago I grant you), I rarely saw a need for the language to do this - there are too many edge cases where a non-determenistic delete could cause you problems. With a little bit of discipline, it's very easy to ensure that you tidy up (that and a couple of handy macros to write out the address of the object when it was created, and the address when it was deleted - if there's a mismatch in the count, you know that you may have a leak and you know which object it is).
|
|
|
|
|
Pete O'Hanlon wrote: Is he stating that he wants the language to cope with this by default? Is he
asking for some form of reference counted/garbage collected tidy up?
That seems to be the case.
The next release of the C++ standard is scheduled for 2017, and I don't think we'll see many changes to the language. We're probably going to add constructs supporting parallel programming and perhaps concepts. I also expect that we'll see some clarifications related to the current standard.
|
|
|
|
|
Hmmm, I would be against seeing that in the language by default. As you say, why add to the language when there are other ways that you could do this - and more importantly, as these methods don't rely on the standards ratification process, you can ship them ahead of the standard.
|
|
|
|
|
Pete O'Hanlon wrote: these methods don't rely on the standards ratification process, you can ship them ahead of the standard
This is how stuff usually makes it into the standard ...
|
|
|
|
|
I see it's been deleted, again.
|
|
|
|
|
|
I've just asked him why he keeps deleting and reposting the same article. We all know the reason why, but it will be interesting to see what he has to say.
From what I read there, it's not an improvement on the original version.
|
|
|
|
|
Looks like he thinks I'm trying "get" him, so guess I've wasted a bit of time
|
|
|
|
|
I must admit, his tactic of deleting and recreating articles to get rid of unwanted low votes is a new one on me.
|
|
|
|
|
Pete O'Hanlon wrote: I must admit, his tactic of deleting and recreating articles to get rid of unwanted low votes is a new one on me.
If it gets prevalent, somebody will probably step in and curb the practice ...
|
|
|
|
|
Richard MacCutchan wrote: why Espen has singled this one out
The title "The new C++ operator && and why should you start using it" is OK, it's something I'd like to see more about.
When you write Right now I am trying to test those changes in CLang, you are implying that you know what you are talking about at an expert level. What Ladislav suggests ahould be implemented at the language level can, and the are many examples of this, easily be implemented at the library level.
Find the section starting with I call it Hurray stl vector<everything> trend.
Do you want destructors to be called on an object when a pointer to that object goes out of scope?
The functionality he is looking for here is already provided by shared_ptr.
We're also looking at a proposal for C++ dynamic arrays, by Lawrence Crowl and Matt Austern - which will proably get my vote.
It's highly unlikely that the language will be altered to provide for features that can reasonably be implemented at the library level.
|
|
|
|
|
I agree with what you are saying, but I still feel that the best way to get the article improved is to enter into a discussion with the author. I have tried it today and he has made a minor change as a result.
Espen Harlinn wrote: When you write Right now I am trying to test those changes in CLang, you are implying that you know what you are talking about at an expert level. I guess anyone who posts an article is implying that they know what they are talking about, but the proof of that lies in the article itself. If it does not live up to its promise then it will be either voted into oblivion or ignored by the majority of readers.
One of these days I'm going to think of a really clever signature.
|
|
|
|
|
Richard MacCutchan wrote: it will be either voted into oblivion
I would much rather see an article getting a huge number of upvotes, and it would be nice if that article was written by Ladislav.
Richard MacCutchan wrote: enter into a discussion with the author
I believe I tried just that when he posted his initial article.
|
|
|
|
|
Hi,
So recently I've been noticing that posts in any forum that receive any vote value lower than 5 end up with the silver Low Vote colour. This must be a bug. It only used to be for values < 3, wasn't it? Now it's < 5. Interestingly, if a message has no votes, it just uses the normal, un-voted colour.
Ed
(P.s. I did search this forum to try and find something else highlighting this issue but didn't find anything.)
Edit: Example here at the moment: http://www.codeproject.com/Messages/4356216/Every-cloud-has-a-silver-lining.aspx[^]
Rating is 3.88 yet is has the silver low vote colour - this happens for 4 and 4.5 etc. Only absolute 5 seems to get the correct colour.
Or this one:
http://www.codeproject.com/Messages/4355262/So-coooool.aspx[^]
Rating 4.75 yet displayed in Silver.
|
|
|
|
|
You need to adjust the 'noise' setting on the orange bar at the top of the forum. I suggest you change this to Medium.
|
|
|
|
|
Erm...right... Okay so does the Noise setting set what value classes as a low vote then?
Sorry for my confusion I just didn't think I had changed any setting like this recently...perhaps I did without realising.
Anyway, thanks,
Ed
|
|
|
|
|
Ed Nutting wrote: I just didn't think I had changed any setting like this recently...perhaps I did without realising
I doubt if you did that. My settings too got reset, like: "notify me if someone replies my message..." OR Show only publicly available posts
I guess some change triggered these for us. OK for now, so no issues.
|
|
|
|
|
I wasn't sure where to put this so Site Bugs/Suggestions seemed like a likely place to get an answer. It's been "bugging" me for a while this one so here goes:
When reporting/approving an article, why when an article has had 5 reports and doesn't get published, does it not show all five member's names who reported it? It only seems to show the names of people who are Platinum in something (or some other selection system?). Can anyone explain what is actually happening and what the logic behind it is?
Thanks,
Ed
|
|
|
|
|
|
No I know how the approvals work - something is published once 5 Editors approve it (not necessarily Platinum - look up the rankings).
What I was wondering was why, when an article is NOT approved, it puts a message at the top saying "This item was closed....by [Editors who reported names]", except that the editors name only usually include 2 or 3 members and only Platinum ones, yet to have been closed/rejected it must have had at least 5 reports. It seems very odd to me...
Thanks anyway,
Ed
|
|
|
|
|
Quote: not necessarily Platinum - look up the rankings
How can you find out who approved your article?
|
|
|
|