There are two Qt bindings for CL to choose from. The old one at
http://sourceforge.net/projects/lisp-cffi-qt4 and the new commonqt.
Can someone who used both compare them in terms of completeness,
readyness and ease of use. Is it possible to use Qt .ui files with one
of these, or all the gui has to be created programatically.
The old one seems more lispy, but seems unmaintained. The demos look
great and the code is clean and readable. The new one has some
momentum, but usage of #_ everywher is a put off. I couldn't make
smoke to work on my machine so I can'd judge for myself.
What do you think?
Regards,
Marko
On 2009-04-09, Marko Koci? <···········@gmail.com> wrote:
> There are two Qt bindings for CL to choose from. The old one at
> http://sourceforge.net/projects/lisp-cffi-qt4 and the new commonqt.
> Can someone who used both compare them in terms of completeness,
> readyness and ease of use. Is it possible to use Qt .ui files with one
> of these, or all the gui has to be created programatically.
>
> The old one seems more lispy, but seems unmaintained. The demos look
> great and the code is clean and readable. The new one has some
> momentum, but usage of #_ everywher is a put off. I couldn't make
> smoke to work on my machine so I can'd judge for myself.
> momentum, but usage of #_ everywher is a put off.
(I'm the author of CommonQt.)
My impression is that lisp-cffi-qt4 will not be developed any further.
The real competition is between CommonQt and another project you didn't
mention, called cl-smoke. Both very new and probably wouldn't exist as
seperate projects if we had known about each other's work.
Is it silly that there are two such similar projects? Yes.
Will we merge them or will one of them supersede the other? Don't know.
Personally I'm busy trying to get CommonQt to run on Windows at all,
because I don't have a need for Qt bindings at the moment if they don't
support Windows.
Assuming that works out, I'll consider porting over the auto-generation
features from one of the other projects. Right now, I have a generator
that works (see http://www.lichteblau.com/git/?p=qtclos.git;a=summary),
but it's too slow at compilation and load time, so I'll probably rewrite
that part from scratch in a different manner.
To my knowledge there is no .ui compiler for Lisp yet.
The normal uic for C++ is written in C++ itself, and there's a fork of
that compiler for QtRuby. I'm guessing that it would be fairly
straightforward to take one of those versions and tweak it until the
output is Common Lisp code rather than Ruby or C++ code.
Alternatively you'd have to rewrite uic from scratch in Lisp, which
might be nicer in the end, but would be more work.
> I couldn't make
> smoke to work on my machine so I can'd judge for myself.
What kinds of problems did you have getting smoke running?
d.
David Lichteblau wrote:
> On 2009-04-09, Marko Koci? <···········@gmail.com> wrote:
>> There are two Qt bindings for CL to choose from. The old one at
>> http://sourceforge.net/projects/lisp-cffi-qt4 and the new commonqt.
>> Can someone who used both compare them in terms of completeness,
>> readyness and ease of use. Is it possible to use Qt .ui files with one
>> of these, or all the gui has to be created programatically.
>>
>> The old one seems more lispy, but seems unmaintained. The demos look
>> great and the code is clean and readable. The new one has some
>> momentum, but usage of #_ everywher is a put off. I couldn't make
>> smoke to work on my machine so I can'd judge for myself.
>> momentum, but usage of #_ everywher is a put off.
>
> (I'm the author of CommonQt.)
>
> My impression is that lisp-cffi-qt4 will not be developed any further.
Same impression here. I started to help the author, but both of us
realized that the approach was a maintenance nightmare. Tk is a much
easier target. Smoke leverages someone else's maintenance.
> Personally I'm busy trying to get CommonQt to run on Windows at all,
> because I don't have a need for Qt bindings at the moment if they don't
> support Windows.
Smoke is very hard to build; its authors seem to assume you are also
building KDE, but there are ways around that. I got it built on my
macbook and was going to ask you some questions about it at ILC, but
never had time. I think it works... never actually tested that it ran
properly.
> To my knowledge there is no .ui compiler for Lisp yet.
Why bother? Wouldn't QUiLoader and QFormBuilder be sufficient?
- Daniel
On Apr 11, 5:48 am, D Herring <········@at.tentpost.dot.com> wrote:
> David Lichteblau wrote:
>
> > (I'm the author of CommonQt.)
Thanks for nice library, and for promoting competition. I have never
heard
about cl-smoke, will take a look as soon as I google it. Hope these
two
projects will merge eventually.
> > My impression is that lisp-cffi-qt4 will not be developed any further.
>
> Same impression here. I started to help the author, but both of us
> realized that the approach was a maintenance nightmare. Tk is a much
> easier target. Smoke leverages someone else's maintenance.
Ltk is nice, but qt just looks better.
> > Personally I'm busy trying to get CommonQt to run on Windows at all,
> > because I don't have a need for Qt bindings at the moment if they don't
> > support Windows.
That is the point of qt, to have cross platform (as in win/mac/linux)
ui.
> Smoke is very hard to build; its authors seem to assume you are also
> building KDE, but there are ways around that. I got it built on my
> macbook and was going to ask you some questions about it at ILC, but
> never had time. I think it works... never actually tested that it ran
> properly.
What about other bindings instead of smoke? Is it possible to use the
same bindings as for QtScript? I'm just brainstorming, of course.
> > To my knowledge there is no .ui compiler for Lisp yet.
>
> Why bother? Wouldn't QUiLoader and QFormBuilder be sufficient?
That is what I was asking.
> - Daniel
Marko Kocić wrote:
> On Apr 11, 5:48 am, D Herring <········@at.tentpost.dot.com> wrote:
>> David Lichteblau wrote:
>>
>>> (I'm the author of CommonQt.)
>
> Thanks for nice library, and for promoting competition. I have never
> heard
> about cl-smoke, will take a look as soon as I google it. Hope these
> two
> projects will merge eventually.
>
>>> My impression is that lisp-cffi-qt4 will not be developed any further.
>> Same impression here. I started to help the author, but both of us
>> realized that the approach was a maintenance nightmare. Tk is a much
>> easier target. Smoke leverages someone else's maintenance.
>
> Ltk is nice, but qt just looks better.
>
>>> Personally I'm busy trying to get CommonQt to run on Windows at all,
>>> because I don't have a need for Qt bindings at the moment if they don't
>>> support Windows.
>
> That is the point of qt, to have cross platform (as in win/mac/linux)
> ui.
I recommend qooxdoo. desktops are so 20th century.
hth,kt
On 2009-04-11, Marko Koci? <···········@gmail.com> wrote:
>> Why bother? ?Wouldn't QUiLoader and QFormBuilder be sufficient?
>
> That is what I was asking.
Yes, QUiLoader works fine in CommonQt.
(QFormBuilder isn't in the bindings, not sure why.)
d.
David Lichteblau <···········@lichteblau.com> writes:
> Personally I'm busy trying to get CommonQt to run on Windows at all,
> because I don't have a need for Qt bindings at the moment if they don't
> support Windows.
I was wondering if lisp bindings exist for Qt on Windows. The answer
seems to be no. Thanks for the information. I was somewhat skeptical
how a lisp interface on top of C++ will look like. I will take a look
anyway.
--
Chetan
On 2009-04-12, Chetan <············@xspam.sbcglobal.net> wrote:
> David Lichteblau <···········@lichteblau.com> writes:
>
>> Personally I'm busy trying to get CommonQt to run on Windows at all,
>> because I don't have a need for Qt bindings at the moment if they don't
>> support Windows.
>
> I was wondering if lisp bindings exist for Qt on Windows. The answer
> seems to be no.
Well. They exist!
They compile, they load, they run, ... and then they crash.
But stay tuned for more. cll will be the first place to know when I
figure out the part with the crashing.
d.
On 2009-04-12, David Lichteblau <···········@lichteblau.com> wrote:
>> I was wondering if lisp bindings exist for Qt on Windows. The answer
>> seems to be no.
[...]
> But stay tuned for more. cll will be the first place to know when I
> figure out the part with the crashing.
As promised, here the results.
The good news is that CommonQt works great on Windows.
So what is the bad news? Of the implementations tested (SBCL, CCL,
Allegro alisp.exe and buildi.exe, LispWorks) only LispWorks manages to
run CommonQt without crashing.
But better one implementation than none, I guess.
d.
On 14 abr, 09:56, David Lichteblau <···········@lichteblau.com> wrote:
> On 2009-04-12, David Lichteblau <···········@lichteblau.com> wrote:
>
> >> I was wondering if lisp bindings exist for Qt on Windows. The answer
> >> seems to be no.
> [...]
> > But stay tuned for more. cll will be the first place to know when I
> > figure out the part with the crashing.
>
> As promised, here the results.
>
> The good news is that CommonQt works great on Windows.
>
> So what is the bad news? Of the implementations tested (SBCL, CCL,
> Allegro alisp.exe and buildi.exe, LispWorks) only LispWorks manages to
> run CommonQt without crashing.
>
> But better one implementation than none, I guess.
>
> d.
I am having some trouble and in CommonQT home page there is no contact
information. I can't clone the repository:
··········@gugamilare-desktop:~$ git clone git://repo.or.cz/commonqt.git
Initialized empty Git repository in /home/gugamilare/
commonqt/.git/
remote: Counting objects: 159, done.
Receiving objects: 100% (159/159), 55.41 KiB | 6 KiB/s, done.: 40%
(64/159), 52.00 KiB | 6 KiB/s
remote: Compressing objects: 100% (156/156), done.
remote: Total 159 (delta 103), reused 0 (delta 0)
Resolving deltas: 100% (103/103), done.
warning: remote HEAD refers to nonexistent ref, unable to checkout.
I couldn't checkout from http://repo.or.cz/r/commonqt.git as well
(same problem), but I managed to checkout from http://repo.or.cz/w/commonqt.git/
. I encountered more problems after doing this, but I will try to
figure it out myself first, it can be some problem related with the
libsmoke provided by kubuntu's package manager.
On 2009-04-14, gugamilare <··········@gmail.com> wrote:
> I am having some trouble and in CommonQT home page there is no contact
> information. I can't clone the repository:
··············@common-lisp.net would be the official place for bug reports.
> ··········@gugamilare-desktop:~$ git clone git://repo.or.cz/commonqt.git
[...]
> warning: remote HEAD refers to nonexistent ref, unable to checkout.
Hmm, that's not good. For now, try the repository at
http://www.lichteblau.com/git/commonqt.git
> I couldn't checkout from http://repo.or.cz/r/commonqt.git as well
> (same problem), but I managed to checkout from http://repo.or.cz/w/commonqt.git/
> . I encountered more problems after doing this, but I will try to
> figure it out myself first, it can be some problem related with the
> libsmoke provided by kubuntu's package manager.
Yes, a known problem. The smoke version included in Ubuntu is too old
for CommonQt and doesn't work out of the box. I developed CommonQt for
recent smoke from svn. As it turns out, smoke was changed incompatibly
last December.
d.