MySQL++ on MinGW

#1

--0-606096962-1126616961=:36129
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

i am working on Dev C++ to write a simple mysql++ program. can any one tell me about the issues regarding the dll files. and how to get one the dll that will wrk for dev c+?

---------------------------------
Yahoo! India Matrimony: Find your partner now.
--0-606096962-1126616961=:36129--

#2

boka man wrote:

Quote:

i am working on Dev C++ to write a simple mysql++ program.

The next version of MySQL++ will not even have MinGW support, because it
doesn't make sense to use it right now.

The primary problem is that MinGW cannot make a working DLL version of
MySQL++. You can find a detailed problem report by searching the
mailing list for posts I've made on MinGW previously. I cannot tell you
anything beyond that. I don't use MinGW myself, so I have little
personal interest in spending more time on the problem. If someone gets
it working, my next argument will collapse, so I will probably put MinGW
support back into MySQL++.

The secondary problem is that keeping MinGW support in MySQL++ means it
is only good for LGPL or GPL programs. Since the primary reason to use
MinGW is to avoid the GPL problems with Cygwin, I don't see any rational
reason to use MinGW and MySQL++ as long as the DLL problem remains.
Even if there is some rational reason to use MinGW for GPL or LGPL
programs instead of Cygwin, it's still a tiny minority of MySQL++ users.
I will not keep a broken feature in MySQL++ just to support a tiny
minority.

Is there some reason you cannot use one of the free versions of Visual
C++ mentioned in the MySQL++ FAQ?

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#3

------=_Part_15893_10485022.1126627175987
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Warren,
I'm not sure why you say that you can't make a working dll of MySQL++ with=
=20
MinGW. I've been doing this for some time now and I've posted some of the=
=20
mods that need to be done to make this work. I haven't been working with=20
MySQL++ for several months now due to some other stuff I'm busy with - so I=
=20
haven't downloaded and compiled version 2.0 or later versions of MySQL++.=
=20
I'd worked with the latest version as of February this year and was able to=
=20
make a dll that worked fine with my program. I did have problems linking=20
multiple files with MySQL library. But that was because I needed to compile=
=20
the MySQL distribution using MinGW.
I hope you continue to provide support for MinGW as my program depends on=
=20
it.
I'm afraid I'm only a very part time programmer and am not familiar with=
=20
the Unix tools that you use to make and configure project files. I don't=20
understand some of the lingo that you all use regarding open source, cvs,=
=20
tar-balls etc. It's a little bit difficult to contribute under those=20
circumstances and I apologize for that.=20
Best regards,
Zahroof Mohamed
Senior Engineer, R&D
Cooper Compression

On 9/13/05, Warren Young <mysqlpp (AT) etr-usa (DOT) com> wrote:=20

Quote:

=20
boka man wrote:
i am working on Dev C++ to write a simple mysql++ program.
=20
The next version of MySQL++ will not even have MinGW support, because it
doesn't make sense to use it right now.
=20
The primary problem is that MinGW cannot make a working DLL version of
MySQL++. You can find a detailed problem report by searching the
mailing list for posts I've made on MinGW previously. I cannot tell you
anything beyond that. I don't use MinGW myself, so I have little
personal interest in spending more time on the problem. If someone gets
it working, my next argument will collapse, so I will probably put MinGW
support back into MySQL++.
=20
The secondary problem is that keeping MinGW support in MySQL++ means it
is only good for LGPL or GPL programs. Since the primary reason to use
MinGW is to avoid the GPL problems with Cygwin, I don't see any rational
reason to use MinGW and MySQL++ as long as the DLL problem remains.
Even if there is some rational reason to use MinGW for GPL or LGPL
programs instead of Cygwin, it's still a tiny minority of MySQL++ users.
I will not keep a broken feature in MySQL++ just to support a tiny
minority.
=20
Is there some reason you cannot use one of the free versions of Visual
C++ mentioned in the MySQL++ FAQ?
=20
--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...mail (DOT) co=
m
=20

------=_Part_15893_10485022.1126627175987--

#4

Zahroof,
I'm sure there is a bigger community than Warren believes who would
appreciate having MySQL++ work in MinGW. I would also be very interested in
hearing if you have been able to compile anything in the version 2.0 or
later branches in MinGW into a working DLL, and if so, what steps you took
to do that. I worked on this project for about a month trying to accomplish
just that, and have recently given up; realizing that it will take someone
quite beyond my skillset and more adept in dealing with "DLL hell".=20=20
I, too, am like you in that I don't always understand all that they
talk about on this mailing-list, but I've learned that's okay. If I can get
a fix working, Warren is always very willing to help and try to incorporate
it into the trunk. We just need a fix!
My need for MySQL++ to compile with MinGW into a workable DLL is
that the rest of my code uses MinGW and I can't make MinGW link to any msvc
compiled DLLs. Again, DLL hell.
Warren, have you considered moving to a product like Bakefile? I
know that it is still in Beta, but I thought it seemed like it could solve a
lot of problems you've been discussing with dealing with different native
makefiles (it can even generate those VS project files, to keep Chris Frey
happy). It worked well for me in MinGW in compiling and linking wxWidgets
(via DLL).

-Patrick

-----Original Message-----
From: Zahroof Mohamed [mailto:zahroofm (AT) gmail (DOT) com]=20
Sent: Tuesday, September 13, 2005 12:00 PM
To: MySQL++ Mailing List
Subject: Re: MySQL++ on MinGW

Warren,
I'm not sure why you say that you can't make a working dll of MySQL++ with=
=20
MinGW. I've been doing this for some time now and I've posted some of the=
=20
mods that need to be done to make this work. I haven't been working with=20
MySQL++ for several months now due to some other stuff I'm busy with - so I=
=20
haven't downloaded and compiled version 2.0 or later versions of MySQL++.=
=20
I'd worked with the latest version as of February this year and was able to=
=20
make a dll that worked fine with my program. I did have problems linking=20
multiple files with MySQL library. But that was because I needed to compile=
=20
the MySQL distribution using MinGW.
I hope you continue to provide support for MinGW as my program depends on=
=20
it.
I'm afraid I'm only a very part time programmer and am not familiar with=
=20
the Unix tools that you use to make and configure project files. I don't=20
understand some of the lingo that you all use regarding open source, cvs,=
=20
tar-balls etc. It's a little bit difficult to contribute under those=20
circumstances and I apologize for that.=20
Best regards,
Zahroof Mohamed
Senior Engineer, R&D
Cooper Compression

On 9/13/05, Warren Young <mysqlpp (AT) etr-usa (DOT) com> wrote:=20

Quote:

=20
boka man wrote:
i am working on Dev C++ to write a simple mysql++ program.
=20
The next version of MySQL++ will not even have MinGW support, because it
doesn't make sense to use it right now.
=20
The primary problem is that MinGW cannot make a working DLL version of
MySQL++. You can find a detailed problem report by searching the
mailing list for posts I've made on MinGW previously. I cannot tell you
anything beyond that. I don't use MinGW myself, so I have little
personal interest in spending more time on the problem. If someone gets
it working, my next argument will collapse, so I will probably put MinGW
support back into MySQL++.
=20
The secondary problem is that keeping MinGW support in MySQL++ means it
is only good for LGPL or GPL programs. Since the primary reason to use
MinGW is to avoid the GPL problems with Cygwin, I don't see any rational
reason to use MinGW and MySQL++ as long as the DLL problem remains.
Even if there is some rational reason to use MinGW for GPL or LGPL
programs instead of Cygwin, it's still a tiny minority of MySQL++ users.
I will not keep a broken feature in MySQL++ just to support a tiny
minority.
=20
Is there some reason you cannot use one of the free versions of Visual
C++ mentioned in the MySQL++ FAQ?
=20
--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...mail (DOT) com
=20

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#5

------=_Part_15986_5198551.1126629016333
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

On 9/13/05, Patrick McMichael <p.s.mcmichael (AT) gats-inc (DOT) com> wrote:=20

Quote:

=20
Zahroof,
I'm sure there is a bigger community than Warren believes who would
appreciate having MySQL++ work in MinGW. I would also be very interested=
=20
in
hearing if you have been able to compile anything in the version 2.0 or
later branches in MinGW into a working DLL, and if so, what steps you too=
k
to do that. I worked on this project for about a month trying to=20
accomplish
just that, and have recently given up; realizing that it will take someon=
e
quite beyond my skillset and more adept in dealing with "DLL hell".

I plan to download the version 2.0x software soon and see if I can work=20
with it. Problem is that Warren is so active that every week almost there i=
s=20
a new release. When I'm ready to work on my project, I'd like to download=
=20
the stable release of MySQL++ and work with that. If I'm successful I'll le=
t=20
you know what I had to do.
=20
My need for MySQL++ to compile with MinGW into a workable DLL is

Quote:

that the rest of my code uses MinGW and I can't make MinGW link to any=20
msvc
compiled DLLs. Again, DLL hell.

By the way, I'm using Parinya software's IDE to do my work. I tried out=
=20
Dev C++ at one point and had some problems. I've found that I need to=20
recompile everything that I'd like to link to on the version of MinGW's gcc=
=20
compilers for it to link properly. I am not trying to mix dll's created wit=
h=20
msvc++ with MinGW c++ dlls. I'm also using FORTRAN dlls compiled with the=
=20
same version of gcc in my project. I'm sure not many people here have to=20
deal with FORTRAN...
Zahroof Mohamed
Senior Engineer, R&D
Cooper Compression

------=_Part_15986_5198551.1126629016333--

#6

Zahroof Mohamed wrote:

Quote:

I'm not sure why you say that you can't make a working dll of MySQL++ with
MinGW.

I started the following thread a month and a half ago, and got no useful
replies:

http://lists.mysql.com/plusplus/4702

The situation is very simple: I do not use all platforms. I only
actively support platforms I personally use, or for which there is
community support. Any other platform is not supported, and special
code to support it is likely to be removed at my earliest convenience.
Dead or useless code causes maintenance hassles, so this is simply an
act of self-preservation. If you have the ability to turn useless code
into useful code, step up and provide a patch. Then the code stays.

Quote:

I've been doing this for some time now and I've posted some of the
mods that need to be done to make this work.

I looked back through all of your old posts, and the only thing I can
see is the suggestion to move all code from row.cpp into row.h. I
objected to that, and that was the end of it, I thought. Is this what
you're referring to?

Quote:

I did have problems linking
multiple files with MySQL library. But that was because I needed to compile
the MySQL distribution using MinGW.

I don't recall reading this before now. This is the sort of thing you
need to post to the list if you want me to consider MinGW a supported
platform. Also, I want to know how exactly you got MySQL to build.
From my experience with the Cygwin port, I know it is not trivial.

Quote:

I hope you continue to provide support for MinGW as my program depends on
it.

Then you need to continue to support the MySQL++ project, by helping out
when I post messages like the one I linked to above. I can't do it alone.

Quote:

I'm afraid I'm only a very part time programmer and am not familiar with
the Unix tools that you use to make and configure project files.

I was wrong to suggest using this in the first place, in our original
email exchanges. I didn't know enough about MinGW at the time. The
proper solution is to get the makemake files working.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#7

Patrick McMichael wrote:

Quote:

I'm sure there is a bigger community than Warren believes who would
appreciate having MySQL++ work in MinGW.

If there is such a large community, why do my posts for help with MinGW
go without response? Why did it take me threatening to remove it to get
even this much response?

I don't give a tinker's damn about MinGW. But MySQL++ works on lots of
other platforms that I don't care about, simply because someone else
submitted the info I needed to make it work. So you see, my platform
biases do not matter. The only thing that matters is whether there are
active contributors in the community to support it.

If you want it to happen, _you_ need to make it happen.

Quote:

I, too, am like you in that I don't always understand all that they
talk about on this mailing-list, but I've learned that's okay.

Yes, this attitude is okay, as long as you're willing to wait for
someone else to fix the problem. Until a fix appears, why does it
matter if I remove MinGW support? Such a decision is not permanent.

Quote:

Warren, have you considered moving to a product like Bakefile?

Chris Frey would pitch a fit.

You might be asking yourself why that matters. It's simple: Chris
contributes, so I care what he thinks. I may be Ultra Supreme Dictator
of MySQL++, but I listen to such people. If you or Zahroof starts
contributing MinGW help, you get a seat on the advisory panel, too.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#8

Zahroof Mohamed wrote:

Quote:

Problem is that Warren is so active that every week almost there is
a new release.

I call that a feature. "Release early, release often." -- E.S.R.

Quote:

I'd like to download
the stable release of MySQL++ and work with that.

Have you considered using the Subversion tree? In a sense, this will
put you even more on the bleeding edge, but it will also relieve you of
having to integrate each new version into your tree. When you're done
making changes, svn can be asked to generate a patch file. (See the
HACKERS file.) I'm far more likely to apply a tested patch than react
to a prose description of needed changes.

Quote:

By the way, I'm using Parinya software's IDE to do my work. I tried out
Dev C++ at one point and had some problems. I've found that I need to
recompile everything that I'd like to link to on the version of MinGW's gcc
compilers for it to link properly.

Just a friendly warning: if your solution involves making people use the
same IDE you do, I'll reject it. In the specific case of MinGW, you
need to figure out some way that anyone can build it, regardless of IDE.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#9

boka man wrote:

Quote:

reason i am working on is because of its ide Dev C++. i personally find
ide's better than command lines(cause i am new to development).

On Windows, IDEs generally are indeed better than command lines, but
that's mainly because the Windows command line sucks, and the
alternatives are also imperfect.

The problem with IDEs, though, is that everyone has his own choice, and
they all use incompatible project files. I cannot support them all, so
my current solution is to support none of them. I realize this isn't
ideal, either, but it's the position that works best for me.

Quote:

i think using dlltool should fix it!

Ah, no. Read this: http://lists.mysql.com/plusplus/4702

Quote:

now that u are saying that making mysql++ wrk on mingw is not wise,

I'm not saying that it's unwise to use MinGW. I'm just saying that,
until the DLL problem is fixed, MinGW loses its primary advantage over
Cygwin.

Since the Cygwin port works, I have no incentive to support MinGW in
this situation. Change the situation, and I will change my mind.

Quote:

can u give me a more detailed look at it or places to get the
info? and how i can contribute to make this possible?

You need to learn how to get MinGW to make a MySQL++ DLL, and then how
to contribute a tasteful patch. Ugly patches get rejected.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw

#10

Warren Young wrote:

Quote:

In the specific case of MinGW

Typo: I meant "In the specific case of MySQL".

And keep in mind, you only need to get the MySQL client library
building. There's no need to get the rest of the MySQL distribution --
the network and embeddable servers, the command-line tools, the test
suite... -- working.

--
MySQL++ Mailing List
For list archives: http://lists.mysql.com/plusplus
To unsubscribe: http://lists.mysql.com/plusplus?unsu...ie.nctu.edu.tw