[Qt-interest] QAbstractItemModel::match and MatchFlags

[Andre Somers]

Op 20-6-2011 18:59, Eric Clark schreef:
>
> Thank you Andre. I see what you are saying about how to test the flag, 
> but it still seems like an error in the code. I am truly surprised 
> that none of the Qt folks have responded to this post. I agree that 
> what you said MatchContains should not equal MatchStartsWith | 
> MatchEndsWith. The only reason I pointed this out is because it seems 
> as though that is what they wanted to do, but made an error when doing 
> it. It is true though that if you set the MatchContains flag, that 
> MatchStartsWith and MatchEndsWith is a valid solution to 
> MatchContains. They are not the only solutions, but they are valid for 
> being contained within the string.
>
> Any Qt folks out there have any comments about this? Or should I just 
> go ahead and report it as a bug and hope that it gets fixed?
>
I'd say: go ahead, and submit it as a bug. This looks like a candidate 
for API cleanup for Qt 5, if you ask me. The root of the problem is, of 
course, that there is no really sensible way to express the situation 
where some flags are mutually exclusive, some real flags, and yet others 
a superset of some flags. At least: I know of none.

André


> Thanks,
>
> Eric
>
> *From:*qt-interest-bounces+eclark=ara.com at qt.nokia.com 
> [mailto:qt-interest-bounces+eclark=ara.com at qt.nokia.com] *On Behalf Of 
> *Andre Somers
> *Sent:* Saturday, June 18, 2011 1:16 AM
> *To:* qt-interest at qt.nokia.com
> *Subject:* Re: [Qt-interest] QAbstractItemModel::match and MatchFlags
>
> Op 18-6-2011 0:51, Eric Clark schreef:
>
> Hello Everyone,
>
> I have been looking at this for quite some time trying to figure why 
> it is the way it is because the Qt guys are usually pretty darn good 
> and usually have a very good reason for why they do things the way 
> they do, but I cannot figure this one out. I am making a custom match 
> function in my model and I am trying to handle all of the flags that 
> can be passed into the function, but there are a few flags that are 
> really throwing me off:
>
> MatchFlag::MatchContains = 1
>
> MatchFlag::MatchStartsWith = 2
>
> MatchFlag::MatchEndsWIth = 3
>
> The flag for MatchEndsWith has two bits flipped, and the two bits 
> (position 0 and 1) match the bits for MatchContains and 
> MatchStartsWith. In other words:
>
> MatchEndsWith == MatchContains | MatchStartsWith;
>
> This seems very wrong to me. To me, it seems like it should really be:
>
> MatchFlag::MatchContains = 3
>
> MatchFlag::MatchStartsWith = 2
>
> MatchFlag::MatchEndsWIth = 1
>
> MatchContains == MatchEndsWith | MatchStartsWith;
>
> I do not see how setting the ends with flag should also be setting the 
> starts with and contains flag. I hope I explained this right. Please 
> can someone explain to me what I am seeing? Is this my imagination or 
> this an actual error in the code?
>
> Good question!
> If I look at the Qt::MatchFlags documentation, it seems like the flags 
> are a combination of a traditional enum with some flags functionality 
> added on top. I disagree that MatchContains == MatchEndsWith | 
> MatchStartsWith. That just isn't true. It seems to me, that you need 
> to use it like this:
>
> if ((matchValue && 0xF) == myRequiredMatchType) { ...
>
> and test for the modifier flags (with values 16 and up) using the normal
> matchValue.testFlag(Qt::MatchRecursive);
>
> I must say: it is not very pretty, because even though it is a Flags, 
> you can not use testFlag() on the lower end of the values.
>
> André
>
>
> _______________________________________________
> Qt-interest mailing list
> Qt-interest at qt.nokia.com
> http://lists.qt.nokia.com/mailman/listinfo/qt-interest

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.qt-project.org/pipermail/qt-interest-old/attachments/20110620/77337f3c/attachment.html