[NEW] deve/codeworker

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

[NEW] deve/codeworker

Auclair Vincent
Information for inst:CodeWorker-4.5.4

Comment:
A universal parsing tool & a source code generator

Description:
CodeWorker is a versatile Open Source parsing tool and a source code
generator devoted to generative programming. Generative programming is a
software engineering approach interested in automating the production of
reusable, tailor-made, adaptable and reliable IT systems.

Maintainer: Vincent Auclair <[hidden email]>

WWW: http://codeworker.free.fr/

Tested on i386
Comments ?

--
Vincent Auclair        -      auclair.vincent[ at ]gmail.com
(+33) 6 80 77 59 67

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Auclair Vincent
On Fri, May 7, 2010 at 11:47 AM, Auclair Vincent
<[hidden email]> wrote:

> Information for inst:CodeWorker-4.5.4
>
> Comment:
> A universal parsing tool & a source code generator
>
> Description:
> CodeWorker is a versatile Open Source parsing tool and a source code
> generator devoted to generative programming. Generative programming is a
> software engineering approach interested in automating the production of
> reusable, tailor-made, adaptable and reliable IT systems.
>
> Maintainer: Vincent Auclair <[hidden email]>
>
> WWW: http://codeworker.free.fr/
>
> Tested on i386
> Comments ?
>
And I forgot the tarball...

--
Vincent Auclair        -      auclair.vincent[ at ]gmail.com
(+33) 6 80 77 59 67

codeworker.tar.gz (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Landry Breuil-6
On Fri, May 07, 2010 at 12:46:55PM +0200, Auclair Vincent wrote:

> On Fri, May 7, 2010 at 11:47 AM, Auclair Vincent
> <[hidden email]> wrote:
> > Information for inst:CodeWorker-4.5.4
> >
> > Comment:
> > A universal parsing tool & a source code generator
> >
> > Description:
> > CodeWorker is a versatile Open Source parsing tool and a source code
> > generator devoted to generative programming. Generative programming is a
> > software engineering approach interested in automating the production of
> > reusable, tailor-made, adaptable and reliable IT systems.
> >
> > Maintainer: Vincent Auclair <[hidden email]>
> >
> > WWW: http://codeworker.free.fr/
> >
> > Tested on i386
> > Comments ?
> >
>
> And I forgot the tarball...

do-configure:
        @true
oh my.. CONFIGURE_STYLE is here for something, set it to an empty value
and it will just work :)
- COMMENT shouldn't start by a capital article, remove it
- why installing the static lib ? is it needed by the binary ?
- the DISTNAME/DISTFILES dance is a bit wrong, NAME is useless here as
  used only once. Instead, we usually set DISTNAME to the name of the
zip/tarball (minus EXTRACT_SUFX) and set the real PKGNAME by hand (ie
codeworkers-${V}.

Landry

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Auclair Vincent
On Fri, May 7, 2010 at 1:56 PM, Landry Breuil <[hidden email]> wrote:

> do-configure:
>        @true
> oh my.. CONFIGURE_STYLE is here for something, set it to an empty value
> and it will just work :)
> - COMMENT shouldn't start by a capital article, remove it
> - why installing the static lib ? is it needed by the binary ?
> - the DISTNAME/DISTFILES dance is a bit wrong, NAME is useless here as
>  used only once. Instead, we usually set DISTNAME to the name of the
> zip/tarball (minus EXTRACT_SUFX) and set the real PKGNAME by hand (ie
> codeworkers-${V}.
Thought I had sent this on Friday. Anyways here is the updated version.

I corrected the issues you mentioned.

You can generate code with codeworker, the generated code needs the
static library.
There aren't any dynamic library shipped yet.
The static library is also used when you want to use codeworker in a
c++ program.


--
Vincent Auclair        -      auclair.vincent[ at ]gmail.com
(+33) 6 80 77 59 67

codeworker.tar.gz (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Landry Breuil-6
On Tue, May 11, 2010 at 12:13:46PM +0200, Auclair Vincent wrote:

> On Fri, May 7, 2010 at 1:56 PM, Landry Breuil <[hidden email]> wrote:
> > do-configure:
> >        @true
> > oh my.. CONFIGURE_STYLE is here for something, set it to an empty value
> > and it will just work :)
> > - COMMENT shouldn't start by a capital article, remove it
> > - why installing the static lib ? is it needed by the binary ?
> > - the DISTNAME/DISTFILES dance is a bit wrong, NAME is useless here as
> >  used only once. Instead, we usually set DISTNAME to the name of the
> > zip/tarball (minus EXTRACT_SUFX) and set the real PKGNAME by hand (ie
> > codeworkers-${V}.
>
> Thought I had sent this on Friday. Anyways here is the updated version.
>
> I corrected the issues you mentioned.
>
> You can generate code with codeworker, the generated code needs the
> static library.
> There aren't any dynamic library shipped yet.
> The static library is also used when you want to use codeworker in a
> c++ program.

Strip the  'a' from COMMENT, they're useless..
It doesn't respect CXXFLAGS (ie try $CFLAGS=-DFOO CXXFLAGS=-DBAR
make). it will turn on -Wall which will show tons of interesting
warnings.. MAKE_FLAGS is not ok, you add $(INCDIRS) as bindly done in the
Makefile but it is not set at this point (and it's useless). CC/CXX should
be honored too, as it should be c++ and not g++ as gmakes uses CXX
for the objects that don't have an explicit target.

Atm, i have
MAKE_FLAGS = CXXFLAGS='${CXXFLAGS}' LFLAGS='-lm' CC='$CXX}' CXX='${CXX}'
which correctly uses CXXFLAGS and makes a correct use of CC/CXX without
patching Makefile. Well, it overrides CC with CXX as it's used in the
Makefile when building generator.cpp/codeworker binary.. so in the end
the makefile shouldbe patched too so that it respects
CXXFLAGS/LDFLAGS for the explicit codeworker target.

Landry

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Auclair Vincent
On Thu, May 13, 2010 at 1:04 PM, Landry Breuil <[hidden email]> wrote:

> On Tue, May 11, 2010 at 12:13:46PM +0200, Auclair Vincent wrote:
>> On Fri, May 7, 2010 at 1:56 PM, Landry Breuil <[hidden email]> wrote:
>> > do-configure:
>> >        @true
>> > oh my.. CONFIGURE_STYLE is here for something, set it to an empty value
>> > and it will just work :)
>> > - COMMENT shouldn't start by a capital article, remove it
>> > - why installing the static lib ? is it needed by the binary ?
>> > - the DISTNAME/DISTFILES dance is a bit wrong, NAME is useless here as
>> >  used only once. Instead, we usually set DISTNAME to the name of the
>> > zip/tarball (minus EXTRACT_SUFX) and set the real PKGNAME by hand (ie
>> > codeworkers-${V}.
>>
>> Thought I had sent this on Friday. Anyways here is the updated version.
>>
>> I corrected the issues you mentioned.
>>
>> You can generate code with codeworker, the generated code needs the
>> static library.
>> There aren't any dynamic library shipped yet.
>> The static library is also used when you want to use codeworker in a
>> c++ program.
>
> Strip the  'a' from COMMENT, they're useless..
> It doesn't respect CXXFLAGS (ie try $CFLAGS=-DFOO CXXFLAGS=-DBAR
> make). it will turn on -Wall which will show tons of interesting
> warnings.. MAKE_FLAGS is not ok, you add $(INCDIRS) as bindly done in the
> Makefile but it is not set at this point (and it's useless). CC/CXX should
> be honored too, as it should be c++ and not g++ as gmakes uses CXX
> for the objects that don't have an explicit target.
>
> Atm, i have
> MAKE_FLAGS =    CXXFLAGS='${CXXFLAGS}' LFLAGS='-lm' CC='$CXX}' CXX='${CXX}'
> which correctly uses CXXFLAGS and makes a correct use of CC/CXX without
> patching Makefile. Well, it overrides CC with CXX as it's used in the
> Makefile when building generator.cpp/codeworker binary.. so in the end
> the makefile shouldbe patched too so that it respects
> CXXFLAGS/LDFLAGS for the explicit codeworker target.
>

Okay, so I patched the makefile, removed the a in the comment and
changed the MAKE_FLAGS.
But, there are way to many warnings with -W and -Wall. I already have
a few dozen patches which is too much. I've send some patches upstream
already. For now, I only saw initialisation list order warnings and
unused variables. But I haven't checked everything yet.

Should I repost the tarball without the patches for the warnings or
wait until it has been dealt upstream. (Should be fast enough)

Thanks for your help!

--
Vincent Auclair        -      auclair.vincent[ at ]gmail.com
(+33) 6 80 77 59 67

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Landry Breuil-6
On Fri, May 14, 2010 at 04:49:42PM +0200, Auclair Vincent wrote:

> On Thu, May 13, 2010 at 1:04 PM, Landry Breuil <[hidden email]> wrote:
> > On Tue, May 11, 2010 at 12:13:46PM +0200, Auclair Vincent wrote:
> >> On Fri, May 7, 2010 at 1:56 PM, Landry Breuil <[hidden email]> wrote:
> >> > do-configure:
> >> >        @true
> >> > oh my.. CONFIGURE_STYLE is here for something, set it to an empty value
> >> > and it will just work :)
> >> > - COMMENT shouldn't start by a capital article, remove it
> >> > - why installing the static lib ? is it needed by the binary ?
> >> > - the DISTNAME/DISTFILES dance is a bit wrong, NAME is useless here as
> >> >  used only once. Instead, we usually set DISTNAME to the name of the
> >> > zip/tarball (minus EXTRACT_SUFX) and set the real PKGNAME by hand (ie
> >> > codeworkers-${V}.
> >>
> >> Thought I had sent this on Friday. Anyways here is the updated version.
> >>
> >> I corrected the issues you mentioned.
> >>
> >> You can generate code with codeworker, the generated code needs the
> >> static library.
> >> There aren't any dynamic library shipped yet.
> >> The static library is also used when you want to use codeworker in a
> >> c++ program.
> >
> > Strip the  'a' from COMMENT, they're useless..
> > It doesn't respect CXXFLAGS (ie try $CFLAGS=-DFOO CXXFLAGS=-DBAR
> > make). it will turn on -Wall which will show tons of interesting
> > warnings.. MAKE_FLAGS is not ok, you add $(INCDIRS) as bindly done in the
> > Makefile but it is not set at this point (and it's useless). CC/CXX should
> > be honored too, as it should be c++ and not g++ as gmakes uses CXX
> > for the objects that don't have an explicit target.
> >
> > Atm, i have
> > MAKE_FLAGS =    CXXFLAGS='${CXXFLAGS}' LFLAGS='-lm' CC='$CXX}' CXX='${CXX}'
> > which correctly uses CXXFLAGS and makes a correct use of CC/CXX without
> > patching Makefile. Well, it overrides CC with CXX as it's used in the
> > Makefile when building generator.cpp/codeworker binary.. so in the end
> > the makefile shouldbe patched too so that it respects
> > CXXFLAGS/LDFLAGS for the explicit codeworker target.
> >
>
> Okay, so I patched the makefile, removed the a in the comment and
> changed the MAKE_FLAGS.
> But, there are way to many warnings with -W and -Wall. I already have
> a few dozen patches which is too much. I've send some patches upstream
> already. For now, I only saw initialisation list order warnings and
> unused variables. But I haven't checked everything yet.
>
> Should I repost the tarball without the patches for the warnings or
> wait until it has been dealt upstream. (Should be fast enough)

The warnings are not really important per se, lots of ports build fine
with TONS of warnings, but at least -Wall makes sure they are shown. We
can still import codeworker with the warnings (ie only patch for
Makefile), and when upstream releases another version update to it. At
least they are aware of it and working on fixing them.

Landry

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Stuart Henderson
On 2010/05/14 17:22, Landry Breuil wrote:
>
> The warnings are not really important per se, lots of ports build fine
> with TONS of warnings, but at least -Wall makes sure they are shown.

It depends on exactly what they are of course. For example, if there are
implicit declarations (for functions returning pointers) or warnings about
pointer conversions you can expect some problems on LP64 arch...

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Landry Breuil-6
On Fri, May 14, 2010 at 04:48:20PM +0100, Stuart Henderson wrote:
> On 2010/05/14 17:22, Landry Breuil wrote:
> >
> > The warnings are not really important per se, lots of ports build fine
> > with TONS of warnings, but at least -Wall makes sure they are shown.
>
> It depends on exactly what they are of course. For example, if there are
> implicit declarations (for functions returning pointers) or warnings about
> pointer conversions you can expect some problems on LP64 arch...

Just to make things clear.. _some_ warnings are not important, some like
you mention are to be fixed. Here for codeworker, it's tons of c++
warnings about variables initialized after/before a superclass member or
smth like that, and tons of #pragma warning.

Landry

Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Auclair Vincent
On Fri, May 14, 2010 at 6:45 PM, Landry Breuil <[hidden email]> wrote:

> On Fri, May 14, 2010 at 04:48:20PM +0100, Stuart Henderson wrote:
>> On 2010/05/14 17:22, Landry Breuil wrote:
>> >
>> > The warnings are not really important per se, lots of ports build fine
>> > with TONS of warnings, but at least -Wall makes sure they are shown.
>>
>> It depends on exactly what they are of course. For example, if there are
>> implicit declarations (for functions returning pointers) or warnings about
>> pointer conversions you can expect some problems on LP64 arch...
>
> Just to make things clear.. _some_ warnings are not important, some like
> you mention are to be fixed. Here for codeworker, it's tons of c++
> warnings about variables initialized after/before a superclass member or
> smth like that, and tons of #pragma warning.
>
New tarball with the Makefile patched.
I changed CC to CXX and LFLAGS to LDFLAGS
Patches for the warnings are being sent upstream.

--
Vincent Auclair        -      auclair.vincent[ at ]gmail.com
(+33) 6 80 77 59 67

codeworker.tar.gz (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [NEW] deve/codeworker

Landry Breuil-6
On Tue, May 18, 2010 at 11:42:21AM +0200, Auclair Vincent wrote:

> On Fri, May 14, 2010 at 6:45 PM, Landry Breuil <[hidden email]> wrote:
> > On Fri, May 14, 2010 at 04:48:20PM +0100, Stuart Henderson wrote:
> >> On 2010/05/14 17:22, Landry Breuil wrote:
> >> >
> >> > The warnings are not really important per se, lots of ports build fine
> >> > with TONS of warnings, but at least -Wall makes sure they are shown.
> >>
> >> It depends on exactly what they are of course. For example, if there are
> >> implicit declarations (for functions returning pointers) or warnings about
> >> pointer conversions you can expect some problems on LP64 arch...
> >
> > Just to make things clear.. _some_ warnings are not important, some like
> > you mention are to be fixed. Here for codeworker, it's tons of c++
> > warnings about variables initialized after/before a superclass member or
> > smth like that, and tons of #pragma warning.
> >
>
> New tarball with the Makefile patched.
> I changed CC to CXX and LFLAGS to LDFLAGS
> Patches for the warnings are being sent upstream.

This one is good to go, thanks!

Landry