Mailing List Archive

1 2 3  View All
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Well Actually mysql is in there, but it is linked dynamically. If you
look at the package i posted up, there is a libmysql.dll that is
referenced when a QT app is run. I did have a bit of a hard time to
get everything in QT linked right, so actually QT is linked to a
static mysql library which references the dynamic mysql library I
provided.


On Tue, 5 Oct 2004 13:01:47 -0400, Buechler, Mark R
<mark_r_buechler@bausch.com> wrote:
> I'm compiling Win32-QT3 now with MinGW. Question - did you manage to get
> MySQL plugin compiled with it? I'd like to stay completely away from Cygwin
> if possible but I don't think that'll be possible with MySQL.
>
> - Mark.
>
> -----Original Message-----
> From: Mario L [mailto:superm1@gmail.com ]
> Sent: Tuesday, October 05, 2004 1:47 AM
> To: Ian Caulfield; Development of mythtv
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> I managed to finally get a build of QT3.3 compiled. I had to get it done
> with mingw to avoid all of the linking issues in cygwin, and then rebuild
> qmake for cygwin to allow myth to be built in cygwin. For some reason
> libmyth is having a really hard time linking against it however, and starts
> referring to all of this x11 stuff for some reason. I did check, and this
> build works with all of the standard qt stuff, ex designer, linguist, and
> all of the tutorials. I posted a rar including all of the necessities to QT
> (libs, binaries, and
> includes) if you want to try to link libmyth against it and see the link
> errors for yourself.
>
> http://clue.eng.iastate.edu/~superm1/qt-3-win32-mingw.rar
>
> Also, everything I used to build QT3 is at the root of that directory.
> I built against the 3_3_branch of CVS, and had to make a couple of my own
> changes. I used mingw provided with devcpp.
>
> I do have to say though, WOW there is a difference in speed between cygwin
> compiled QTWin32 ( for what worked at least) and QTWin32 compiled under
> mingw. If designer wasn't so huge (or if I had more file space provided by
> school) then I would post so you could see the difference. Let me know if
> any progress is made by anyone to get this link working
>
>
> Bausch & Lomb
> 150 Years of Perfecting Vision, Enhancing Life (TM)
>
> EMAIL DISCLAIMER
>
> Please Note: The information contained in this message may be privileged and
> confidential, protected from disclosure, and/or intended only for the use of
> the individual or entity named above. If the reader of this message is not
> the intended recipient, or an employee or agent responsible for delivering
> this message to the intended recipient, you are hereby notified that any
> disclosure, distribution, copying or other dissemination of this
> communication is strictly prohibited. If you received this communication in
> error, please immediately reply to the sender, delete the message and
> destroy all copies of it.
>
> Thank You
>
>
Re: compiling mythtv frontend for cygwin [ In reply to ]
Well not exactly, mingw gcc seems to do things alot differently then
the cygwin gcc does, and this makes for lots of troubles linking
libraries between the two. So here is where the unfortunateness (if
thats even a word) stands:

Options:
1) Compile QT 3_3_branch with mingw (Compiles fine), link with SQL
from undefined source (Links fine)
A) Can't link myth compiled with cygwin to QT with mingw,
errors about how mingw causes a jump into the QT code
B) Can't compile myth with mingw runtime within cygwin (-mno-cygwin),
Compiles and Links (with changes up to my old patch):
libavcodec, libavfomat
Doesn't finish compilation of libmyth, originally i was
getting errors about missing sys headers, but recently i have errors
in mythcontext.cpp about a undefined sleep function, which i haven't
looked too much into yet.

2) Compile QT 3_3_branch with cygwin, can't compile - no support yet

3) Compile QT3_1_branch with cygwin (Compiles fine), link with SQL from tarball
A) QT3_1 branch has link problems with current cygwin ld, and
causes apps to crash

4) Compile QT3_1 branch with mingw (Compiles fine), but no support to link SQL

Of these options, 1B is the most promissing, and would run the
quickest, but is still turning out to be pain. I have a QT binary up
if you want to give this a partial shot. Just make sure you declare
your QMAKESPEC to be win32-g++ and your QTDIR to wherever you put this
binary. The SQL file is a dynamic library and if you actually get
myth to compile will need to be copied along with the qt-mt file.

On Fri, 8 Oct 2004 12:46:34 -0400, Eric A. Litman <elitman@yahoo.com> wrote:
> Have you managed to make any more progress w/ mingw+QT/Win32?
>
>
>
>
> ----- Original Message -----
> From: "Mario L" <superm1@gmail.com >
> To: "Buechler, Mark R" <mark_r_buechler@bausch.com >; "Development of mythtv"
> <mythtv-dev@mythtv.org >
> Sent: Tuesday, October 05, 2004 1:27 PM
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> > Well Actually mysql is in there, but it is linked dynamically. If you
> > look at the package i posted up, there is a libmysql.dll that is
> > referenced when a QT app is run. I did have a bit of a hard time to
> > get everything in QT linked right, so actually QT is linked to a
> > static mysql library which references the dynamic mysql library I
> > provided.
> >
> >
> > On Tue, 5 Oct 2004 13:01:47 -0400, Buechler, Mark R
> > <mark_r_buechler@bausch.com > wrote:
> >> I'm compiling Win32-QT3 now with MinGW. Question - did you manage to get
> >> MySQL plugin compiled with it? I'd like to stay completely away from
> >> Cygwin
> >> if possible but I don't think that'll be possible with MySQL.
> >>
> >> - Mark.
> >>
> >> -----Original Message-----
> >> From: Mario L [mailto:superm1@gmail.com ]
> >> Sent: Tuesday, October 05, 2004 1:47 AM
> >> To: Ian Caulfield; Development of mythtv
> >> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
> >>
> >> I managed to finally get a build of QT3.3 compiled. I had to get it done
> >> with mingw to avoid all of the linking issues in cygwin, and then rebuild
> >> qmake for cygwin to allow myth to be built in cygwin. For some reason
> >> libmyth is having a really hard time linking against it however, and
> >> starts
> >> referring to all of this x11 stuff for some reason. I did check, and
> >> this
> >> build works with all of the standard qt stuff, ex designer, linguist, and
> >> all of the tutorials. I posted a rar including all of the necessities to
> >> QT
> >> (libs, binaries, and
> >> includes) if you want to try to link libmyth against it and see the link
> >> errors for yourself.
> >>
> >> http://clue.eng.iastate.edu/~superm1/qt-3-win32-mingw.rar
> >>
> >> Also, everything I used to build QT3 is at the root of that directory.
> >> I built against the 3_3_branch of CVS, and had to make a couple of my
> >> own
> >> changes. I used mingw provided with devcpp.
> >>
> >> I do have to say though, WOW there is a difference in speed between
> >> cygwin
> >> compiled QTWin32 ( for what worked at least) and QTWin32 compiled under
> >> mingw. If designer wasn't so huge (or if I had more file space provided
> >> by
> >> school) then I would post so you could see the difference. Let me know
> >> if
> >> any progress is made by anyone to get this link working
> >>
> >>
> >> Bausch & Lomb
> >> 150 Years of Perfecting Vision, Enhancing Life (TM)
> >>
> >> EMAIL DISCLAIMER
> >>
> >> Please Note: The information contained in this message may be privileged
> >> and
> >> confidential, protected from disclosure, and/or intended only for the use
> >> of
> >> the individual or entity named above. If the reader of this message is
> >> not
> >> the intended recipient, or an employee or agent responsible for
> >> delivering
> >> this message to the intended recipient, you are hereby notified that any
> >> disclosure, distribution, copying or other dissemination of this
> >> communication is strictly prohibited. If you received this communication
> >> in
> >> error, please immediately reply to the sender, delete the message and
> >> destroy all copies of it.
> >>
> >> Thank You
> >>
> >>
> >
>
>
> --------------------------------------------------------------------------------
>
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev@mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
>
>
Re: compiling mythtv frontend for cygwin [ In reply to ]
Not exactly a complete patch, but here is what I have got in the order
of moving things over to mingw (assuming you got libavcodec and
libavformat and qt3_3 compiled, and my previous patch applied).

All for libmyth:
/**********Dont forget to add the -mno-cygwin switch when compiling,
you will regret it if you forget************/

1) Two new files, dlfcn_win32.h & usleep_win32.h

Mythcontext.cpp

#include "usleep_win32.h"

Remotefile.cpp

#include "usleep_win32.h"

//util.cpp

#include "usleep_win32.h"

//volumecontrol.cpp

#include "usleep_win32.h"

And

In kMuteState VolumeControl::IterateMutedChannels(void)

#endif

to

#else
return MUTE_OFF;
#endif

And

#include <sys/ioctl.h>

To

#ifndef __MINGW32__
#include <sys/ioctl.h>
#endif

//mythplugin.cpp






#include <dlfcn.h>

To

#ifndef __MINGW32__
#include <dlfcn.h>
#else
#include "dlfcn_win32.h"
#endif

//httpcomms.h

#ifdef __MINGW32__
#include "usleep_win32.h"
#endif
--------------------------
Still wont compile all the way, but that is about as far as I'll be
able to get for a few days, I'll be a bit busy.
RE: Re: compiling mythtv frontend for cygwin [ In reply to ]
Thanks to Eric's help I've managed to compile 3.1 and 3.3 against the libs
he provided (see his email for the link). QT3.1 compiled without issue with
mingw but 3.3 had some trouble with configure.exe.

Do you have a patch put together for myth which I can apply to cvs? I'd like
to start playing with this if I could.

Thanks, Mark.

-----Original Message-----
From: Mario L [mailto:superm1@gmail.com]
Sent: Friday, October 08, 2004 1:21 PM
To: Eric A. Litman; MythTV-Developer List
Subject: [mythtv] Re: compiling mythtv frontend for cygwin

Well not exactly, mingw gcc seems to do things alot differently then the
cygwin gcc does, and this makes for lots of troubles linking libraries
between the two. So here is where the unfortunateness (if thats even a
word) stands:

Options:
1) Compile QT 3_3_branch with mingw (Compiles fine), link with SQL from
undefined source (Links fine)
A) Can't link myth compiled with cygwin to QT with mingw, errors about
how mingw causes a jump into the QT code
B) Can't compile myth with mingw runtime within cygwin (-mno-cygwin),
Compiles and Links (with changes up to my old patch):
libavcodec, libavfomat
Doesn't finish compilation of libmyth, originally i was
getting errors about missing sys headers, but recently i have errors in
mythcontext.cpp about a undefined sleep function, which i haven't looked
too much into yet.

2) Compile QT 3_3_branch with cygwin, can't compile - no support yet

3) Compile QT3_1_branch with cygwin (Compiles fine), link with SQL from
tarball
A) QT3_1 branch has link problems with current cygwin ld, and causes
apps to crash

4) Compile QT3_1 branch with mingw (Compiles fine), but no support to link
SQL

Of these options, 1B is the most promissing, and would run the quickest, but
is still turning out to be pain. I have a QT binary up if you want to give
this a partial shot. Just make sure you declare your QMAKESPEC to be
win32-g++ and your QTDIR to wherever you put this binary. The SQL file is a
dynamic library and if you actually get myth to compile will need to be
copied along with the qt-mt file.

On Fri, 8 Oct 2004 12:46:34 -0400, Eric A. Litman <elitman@yahoo.com> wrote:
> Have you managed to make any more progress w/ mingw+QT/Win32?
>
>
>
>
> ----- Original Message -----
> From: "Mario L" <superm1@gmail.com >
> To: "Buechler, Mark R" <mark_r_buechler@bausch.com >; "Development of
mythtv"
> <mythtv-dev@mythtv.org >
> Sent: Tuesday, October 05, 2004 1:27 PM
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> > Well Actually mysql is in there, but it is linked dynamically. If
> > you look at the package i posted up, there is a libmysql.dll that is
> > referenced when a QT app is run. I did have a bit of a hard time to
> > get everything in QT linked right, so actually QT is linked to a
> > static mysql library which references the dynamic mysql library I
> > provided.
> >
> >
> > On Tue, 5 Oct 2004 13:01:47 -0400, Buechler, Mark R
> > <mark_r_buechler@bausch.com > wrote:
> >> I'm compiling Win32-QT3 now with MinGW. Question - did you manage
> >> to get MySQL plugin compiled with it? I'd like to stay completely
> >> away from Cygwin if possible but I don't think that'll be possible
> >> with MySQL.
> >>
> >> - Mark.
> >>
> >> -----Original Message-----
> >> From: Mario L [mailto:superm1@gmail.com ]
> >> Sent: Tuesday, October 05, 2004 1:47 AM
> >> To: Ian Caulfield; Development of mythtv
> >> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
> >>
> >> I managed to finally get a build of QT3.3 compiled. I had to get
> >> it done with mingw to avoid all of the linking issues in cygwin,
> >> and then rebuild qmake for cygwin to allow myth to be built in
> >> cygwin. For some reason libmyth is having a really hard time
> >> linking against it however, and starts referring to all of this x11
> >> stuff for some reason. I did check, and this build works with all
> >> of the standard qt stuff, ex designer, linguist, and all of the
> >> tutorials. I posted a rar including all of the necessities to QT
> >> (libs, binaries, and
> >> includes) if you want to try to link libmyth against it and see the
> >> link errors for yourself.
> >>
> >> http://clue.eng.iastate.edu/~superm1/qt-3-win32-mingw.rar
> >>
> >> Also, everything I used to build QT3 is at the root of that directory.
> >> I built against the 3_3_branch of CVS, and had to make a couple of
> >> my own changes. I used mingw provided with devcpp.
> >>
> >> I do have to say though, WOW there is a difference in speed between
> >> cygwin compiled QTWin32 ( for what worked at least) and QTWin32
> >> compiled under mingw. If designer wasn't so huge (or if I had more
> >> file space provided by
> >> school) then I would post so you could see the difference. Let me
> >> know if any progress is made by anyone to get this link working
> >>
> >>
> >> Bausch & Lomb
> >> 150 Years of Perfecting Vision, Enhancing Life (TM)
> >>
> >> EMAIL DISCLAIMER
> >>
> >> Please Note: The information contained in this message may be
> >> privileged and confidential, protected from disclosure, and/or
> >> intended only for the use of the individual or entity named above.
> >> If the reader of this message is not the intended recipient, or an
> >> employee or agent responsible for delivering this message to the
> >> intended recipient, you are hereby notified that any disclosure,
> >> distribution, copying or other dissemination of this communication
> >> is strictly prohibited. If you received this communication in
> >> error, please immediately reply to the sender, delete the message
> >> and destroy all copies of it.
> >>
> >> Thank You
> >>
> >>
> >
>
>
> ----------------------------------------------------------------------
> ----------
>
> _______________________________________________
> mythtv-dev mailing list
> mythtv-dev@mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
>
>




Bausch & Lomb
150 Years of Perfecting Vision, Enhancing Life (TM)






EMAIL DISCLAIMER

Please Note: The information contained in this message may be privileged and
confidential, protected from disclosure, and/or intended only for the use of
the individual or entity named above. If the reader of this message is not
the intended recipient, or an employee or agent responsible for delivering
this message to the intended recipient, you are hereby notified that any
disclosure, distribution, copying or other dissemination of this
communication is strictly prohibited. If you received this communication in
error, please immediately reply to the sender, delete the message and
destroy all copies of it.

Thank You
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
I have a patch (previously posted) which will allow it to compile with
cygwin. The problem is that a cygwin build of myth won't link
properly with a mingw build of QT. There are some fixes that i have
on my site to compile QT3.3 ( preferred over 3.1), meaningly the fix
for configure.exe and the library you need to link SQL with it. The
library for mysql is taken from a native windows build with symbols
exported. It is linked statically, but dependent on a dynamic dll
file along with the qt library.

Working from that basis I am still working on a patch to allow mingw
compilation, there are A LOT of things that mingw doesn't define, that
cygwin does which I am now realizing. I can release a patch that will
bring it up to allow libavcodec, libavforamt, and libmyth, but I am
still working with libmythtv.

If you could apply my old patch to CVS, that is a great start, and
then I will get the new CVS and redo the changes I have done and make
a second patch. If it is easier to just do one patch alltogether,
i'll do that instead though, just let me know what is most convenient.

Also, if you don't want to fuss around with compiling QT3.3 under
mingw, i do have a binary at my site
http://clue.eng.iastate.edu/~superm1/qt-3-win32-mingw.rar .

Again just give me the signal what is best to do at this point.
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Hey mark - scratch that last patch, here is a patch to get the
compilation complete. To compile properly it needs to be done using
the mingw runtime under a cygwin environment. This is all pretty much
automated thanks to qmake, as long as mingw runtime is installed in
the cygwin envionment.
Also, even though a mingw build of QT is used, a cygwin version of
qmake has to be used. My binary (which i just updated), has a version
of qmake that is intended for cygwin usage, however the qt libraries
are built for mingw usage. The other dependencies are pthreads-win32
(as previously stated), freetype2, and lamemp3 (built under mingw),
and directX.

All the dependencies are at my site, in usable format. Unfortuantely
there is a problem when launching which I am beginning to think might
have been a cause of using the cygwin environment. The problem is
looking very similar to the previous problems with QT3.1.

All together, attached it the patch, and all the dependencies are at
my site. If you are going to compile, be sure that all of these
dependencies are put into /usr/local/lib, as that is hardcoded into a
couple of files to hunt there.

http://clue.eng.iastate.edu/~superm1
I re-arranged everything there to be more logically put. Dependencies
are all in the same folder.
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Okay - So as for my patch, I just did a build with just mingw and msys
and ran into link errors with QT. I redid QT from CVS, and redid
myth, and did a successful build. I realized that there is one more
dependency missing, zlib, and added that to my site.

First of all launching is obviously quicker under mingw for what works
about it. It doesn't seem to launch at all from a command line,
unless launching from msys, and even then it doesn't seem to like
windows filenames. This is finally at the point where it will compile
all the way with that patch I posted, and it is just working out
symantecs of a non posix file structure where it is hard coded in
myth's code. If that patch could get applied to CVS by someone, then
some other people can finally join in on this since it is basically
down to a science.


For those who don't want to follow the whole thread:

My site: http://clue.eng.iastate.edu/~superm1

Dep's Needed to build Myth under MinGW
1) MinGW
2) mSYS
3) Freetype2
4) Zlib
5) QT 3.3 w/ mysql support
6) P-Threads Win32
7) LameMp3 built w/ mingw
8) DirectX includes

All dep's needed to build are at my site.

SideNotes:
1) The directX headers have to be mounted under msys under
/mingw/includes/directx
2) When launching one of the executables, you will need dll's from
freetype,zlib,qt-mt333, mysql, pthreadsGCE, and myth all in the same
directory. The mysql dll is in the QT binary I provided. (If you
want to be crazy, build QT yourself, everything i changed and needed
is there)
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
On Tue, 12 Oct 2004 03:47:43 -0500, Mario L wrote:
> .... If that patch could get applied to CVS by someone, then
> some other people can finally join in on this since it is basically
> down to a science.

While Isaac is probably the only one with a say in whether something
like this goes in, I wouldn't count on it. First off, there is no
reason to put it into the code yet, as it doesn't work. Anyone
competent enough to do hacking on it can install your patch and make
new ones. Second the code will be nearly impossible to maintain.
There are ifdefs all over the place. The ones controlling headers I
can understand, but there are many others. Since almost all of the
developers run Linux, they don't want to worry about whether they are
breaking the win32 build, and I would guess that Isaac doesn't want to
have to wait for the win32 maintainers to fix bugs before releasing
new versions. I would say (a) get it working first. I don't see nay
compelling reason to add it otherwise, and (b) after you do that,
clean up the patch to minimize differences such that the maintenance
will be easier.

As an aside, if you want unix people to look at your patches .tar.bz2
(or .gz) is a much better format that rar. I hate having to use
proprietary programs to just look at a patch. And again anyone
familiar with cygwin will know how to use tar, so you aren't
alienating your users.

.Geoff
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
On Tuesday 12 October 2004 10:51 am, Geoffrey Hausheer wrote:
> On Tue, 12 Oct 2004 03:47:43 -0500, Mario L wrote:
> > .... If that patch could get applied to CVS by someone, then
> > some other people can finally join in on this since it is basically
> > down to a science.
>
> While Isaac is probably the only one with a say in whether something
> like this goes in, I wouldn't count on it. First off, there is no
> reason to put it into the code yet, as it doesn't work. Anyone
> competent enough to do hacking on it can install your patch and make
> new ones. Second the code will be nearly impossible to maintain.
> There are ifdefs all over the place. The ones controlling headers I
> can understand, but there are many others. Since almost all of the
> developers run Linux, they don't want to worry about whether they are
> breaking the win32 build, and I would guess that Isaac doesn't want to
> have to wait for the win32 maintainers to fix bugs before releasing
> new versions. I would say (a) get it working first. I don't see nay
> compelling reason to add it otherwise, and (b) after you do that,
> clean up the patch to minimize differences such that the maintenance
> will be easier.
>
> As an aside, if you want unix people to look at your patches .tar.bz2
> (or .gz) is a much better format that rar. I hate having to use
> proprietary programs to just look at a patch. And again anyone
> familiar with cygwin will know how to use tar, so you aren't
> alienating your users.

I agree with all this - there is too much cruft in the patch that should be
slimmed down by quite a bit before I'd consider applying it.

Isaac
_______________________________________________
mythtv-dev mailing list
mythtv-dev@mythtv.org
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
RE: Re: compiling mythtv frontend for cygwin [ In reply to ]
After some trouble (most having to do with not replacing qmake.exe) I've
managed to get it completely compiled with your patch. I'm attempting to do
an install now. Did you manage to successfully do a "mingw32-make install"
or did you simply copy the files in place by hand?

- Mark.

-----Original Message-----
From: Mario L [mailto:superm1@gmail.com]
Sent: Tuesday, October 12, 2004 4:48 AM
To: Buechler, Mark R; MythTV-Developer List
Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin

Okay - So as for my patch, I just did a build with just mingw and msys and
ran into link errors with QT. I redid QT from CVS, and redid myth, and did
a successful build. I realized that there is one more dependency missing,
zlib, and added that to my site.

First of all launching is obviously quicker under mingw for what works
about it. It doesn't seem to launch at all from a command line,
unless launching from msys, and even then it doesn't seem to like windows
filenames. This is finally at the point where it will compile all the way
with that patch I posted, and it is just working out symantecs of a non
posix file structure where it is hard coded in myth's code. If that patch
could get applied to CVS by someone, then some other people can finally join
in on this since it is basically down to a science.


For those who don't want to follow the whole thread:

My site: http://clue.eng.iastate.edu/~superm1

Dep's Needed to build Myth under MinGW
1) MinGW
2) mSYS
3) Freetype2
4) Zlib
5) QT 3.3 w/ mysql support
6) P-Threads Win32
7) LameMp3 built w/ mingw
8) DirectX includes

All dep's needed to build are at my site.

SideNotes:
1) The directX headers have to be mounted under msys under
/mingw/includes/directx
2) When launching one of the executables, you will need dll's from
freetype,zlib,qt-mt333, mysql, pthreadsGCE, and myth all in the same
directory. The mysql dll is in the QT binary I provided. (If you want to
be crazy, build QT yourself, everything i changed and needed is there)




Bausch & Lomb
150 Years of Perfecting Vision, Enhancing Life (TM)






EMAIL DISCLAIMER

Please Note: The information contained in this message may be privileged and
confidential, protected from disclosure, and/or intended only for the use of
the individual or entity named above. If the reader of this message is not
the intended recipient, or an employee or agent responsible for delivering
this message to the intended recipient, you are hereby notified that any
disclosure, distribution, copying or other dissemination of this
communication is strictly prohibited. If you received this communication in
error, please immediately reply to the sender, delete the message and
destroy all copies of it.

Thank You
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
They are all copied by hand, you'll need mythavcodec*.dll
mythavformat*.dll myth*.dll mythtv*.dll mythfrontend.exe mythtv.exe
mythepg.exe mythtvosd*.exe and mythprogfind.exe in the same directory.
Also each of the dependency dlls has to be there, ex
pthreadsGCE.dll, freetype-6.dll, zlib.dll, and qt-mt333.dll. You can
also add another directory to your PATH and put the dlls in that path.

Things may or may not run for you at that point, I played around alot
with my prefix and changed alot of lines to try to elimate posix
stuff, but now i am realizing i probably just had to change that
prefix since windows doesnt really care if you use a / or \ in
directories. Also, QT seems to have trouble eveywhere myth used
iterators in a QStringList, which seems to be pretty common. I have
been trying to use gdb to find out a little bit more about why, but
have been having a pretty rough time with it. I think I'm going to
revert back to some older code where I hadn't changed the posix stuff
and then rebuild QT with debug symbols and try again.


On Thu, 14 Oct 2004 09:44:41 -0400, Buechler, Mark R
<mark_r_buechler@bausch.com> wrote:
> After some trouble (most having to do with not replacing qmake.exe) I've
> managed to get it completely compiled with your patch. I'm attempting to do
> an install now. Did you manage to successfully do a "mingw32-make install"
> or did you simply copy the files in place by hand?
>
> - Mark.
>
>
>
> -----Original Message-----
> From: Mario L [mailto:superm1@gmail.com ]
> Sent: Tuesday, October 12, 2004 4:48 AM
> To: Buechler, Mark R; MythTV-Developer List
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> Okay - So as for my patch, I just did a build with just mingw and msys and
> ran into link errors with QT. I redid QT from CVS, and redid myth, and did
> a successful build. I realized that there is one more dependency missing,
> zlib, and added that to my site.
>
> First of all launching is obviously quicker under mingw for what works
> about it. It doesn't seem to launch at all from a command line,
> unless launching from msys, and even then it doesn't seem to like windows
> filenames. This is finally at the point where it will compile all the way
> with that patch I posted, and it is just working out symantecs of a non
> posix file structure where it is hard coded in myth's code. If that patch
> could get applied to CVS by someone, then some other people can finally join
> in on this since it is basically down to a science.
>
> For those who don't want to follow the whole thread:
>
> My site: http://clue.eng.iastate.edu/~superm1
>
> Dep's Needed to build Myth under MinGW
> 1) MinGW
> 2) mSYS
> 3) Freetype2
> 4) Zlib
> 5) QT 3.3 w/ mysql support
> 6) P-Threads Win32
> 7) LameMp3 built w/ mingw
> 8) DirectX includes
>
> All dep's needed to build are at my site.
>
> SideNotes:
> 1) The directX headers have to be mounted under msys under
> /mingw/includes/directx
> 2) When launching one of the executables, you will need dll's from
> freetype,zlib,qt-mt333, mysql, pthreadsGCE, and myth all in the same
> directory. The mysql dll is in the QT binary I provided. (If you want to
> be crazy, build QT yourself, everything i changed and needed is there)
>
>
> Bausch & Lomb
> 150 Years of Perfecting Vision, Enhancing Life (TM)
>
> EMAIL DISCLAIMER
>
> Please Note: The information contained in this message may be privileged and
> confidential, protected from disclosure, and/or intended only for the use of
> the individual or entity named above. If the reader of this message is not
> the intended recipient, or an employee or agent responsible for delivering
> this message to the intended recipient, you are hereby notified that any
> disclosure, distribution, copying or other dissemination of this
> communication is strictly prohibited. If you received this communication in
> error, please immediately reply to the sender, delete the message and
> destroy all copies of it.
>
> Thank You
>
>
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
> On Tuesday 12 October 2004 10:51 am, Geoffrey Hausheer wrote:
> > On Tue, 12 Oct 2004 03:47:43 -0500, Mario L wrote:
> > > .... If that patch could get applied to CVS by someone, then
> > > some other people can finally join in on this since it is basically
> > > down to a science.
> >
> > While Isaac is probably the only one with a say in whether something
> > like this goes in, I wouldn't count on it. First off, there is no
> > reason to put it into the code yet, as it doesn't work. Anyone
> > competent enough to do hacking on it can install your patch and make
> > new ones. Second the code will be nearly impossible to maintain.
> > There are ifdefs all over the place. The ones controlling headers I
> > can understand, but there are many others. Since almost all of the
> > developers run Linux, they don't want to worry about whether they are
> > breaking the win32 build, and I would guess that Isaac doesn't want to
> > have to wait for the win32 maintainers to fix bugs before releasing
> > new versions. I would say (a) get it working first. I don't see nay
> > compelling reason to add it otherwise, and (b) after you do that,
> > clean up the patch to minimize differences such that the maintenance
> > will be easier.
> >
> > As an aside, if you want unix people to look at your patches .tar.bz2
> > (or .gz) is a much better format that rar. I hate having to use
> > proprietary programs to just look at a patch. And again anyone
> > familiar with cygwin will know how to use tar, so you aren't
> > alienating your users.

>I agree with all this - there is too much cruft in the patch that should be
>slimmed down by quite a bit before I'd consider applying it.

Okay sounds like a plan, once I actually get it to a point of working,
I will clean it up. Also I'll remember to release it in a more unix
friendly format next time. Thanks for the input.
RE: Re: compiling mythtv frontend for cygwin [ In reply to ]
Pretty much the same here. Mythfrontend.exe will startup from either msys or
cmd but immediately segfaults. Starting from cmd shows no output though,
which is standard with Win apps. It looks like it's finding all the files
correctly - or at least mysql.txt. I'm unable to put it in verbose mode to
get more output since that makes it segfault earlier. I need to put some
more debug messages in the code to figure out where it's bombing or attempt
the gdb route.

- Mark.

-----Original Message-----
From: Mario L [mailto:superm1@gmail.com]
Sent: Thursday, October 14, 2004 10:28 AM
To: Buechler, Mark R; MythTV-Developer List
Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin

They are all copied by hand, you'll need mythavcodec*.dll mythavformat*.dll
myth*.dll mythtv*.dll mythfrontend.exe mythtv.exe mythepg.exe mythtvosd*.exe
and mythprogfind.exe in the same directory.
Also each of the dependency dlls has to be there, ex pthreadsGCE.dll,
freetype-6.dll, zlib.dll, and qt-mt333.dll. You can also add another
directory to your PATH and put the dlls in that path.

Things may or may not run for you at that point, I played around alot with
my prefix and changed alot of lines to try to elimate posix stuff, but now i
am realizing i probably just had to change that prefix since windows doesnt
really care if you use a / or \ in directories. Also, QT seems to have
trouble eveywhere myth used iterators in a QStringList, which seems to be
pretty common. I have been trying to use gdb to find out a little bit more
about why, but have been having a pretty rough time with it. I think I'm
going to revert back to some older code where I hadn't changed the posix
stuff and then rebuild QT with debug symbols and try again.


On Thu, 14 Oct 2004 09:44:41 -0400, Buechler, Mark R
<mark_r_buechler@bausch.com> wrote:
> After some trouble (most having to do with not replacing qmake.exe)
> I've managed to get it completely compiled with your patch. I'm
> attempting to do an install now. Did you manage to successfully do a
"mingw32-make install"
> or did you simply copy the files in place by hand?
>
> - Mark.
>
>
>
> -----Original Message-----
> From: Mario L [mailto:superm1@gmail.com ]
> Sent: Tuesday, October 12, 2004 4:48 AM
> To: Buechler, Mark R; MythTV-Developer List
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> Okay - So as for my patch, I just did a build with just mingw and msys
> and ran into link errors with QT. I redid QT from CVS, and redid
> myth, and did a successful build. I realized that there is one more
> dependency missing, zlib, and added that to my site.
>
> First of all launching is obviously quicker under mingw for what works
> about it. It doesn't seem to launch at all from a command line,
> unless launching from msys, and even then it doesn't seem to like
> windows filenames. This is finally at the point where it will compile
> all the way with that patch I posted, and it is just working out
> symantecs of a non posix file structure where it is hard coded in
> myth's code. If that patch could get applied to CVS by someone, then
> some other people can finally join in on this since it is basically down
to a science.
>
> For those who don't want to follow the whole thread:
>
> My site: http://clue.eng.iastate.edu/~superm1
>
> Dep's Needed to build Myth under MinGW
> 1) MinGW
> 2) mSYS
> 3) Freetype2
> 4) Zlib
> 5) QT 3.3 w/ mysql support
> 6) P-Threads Win32
> 7) LameMp3 built w/ mingw
> 8) DirectX includes
>
> All dep's needed to build are at my site.
>
> SideNotes:
> 1) The directX headers have to be mounted under msys under
> /mingw/includes/directx
> 2) When launching one of the executables, you will need dll's from
> freetype,zlib,qt-mt333, mysql, pthreadsGCE, and myth all in the same
> directory. The mysql dll is in the QT binary I provided. (If you
> want to be crazy, build QT yourself, everything i changed and needed
> is there)
>
>
> Bausch & Lomb
> 150 Years of Perfecting Vision, Enhancing Life (TM)
>
> EMAIL DISCLAIMER
>
> Please Note: The information contained in this message may be
> privileged and confidential, protected from disclosure, and/or
> intended only for the use of the individual or entity named above. If
> the reader of this message is not the intended recipient, or an
> employee or agent responsible for delivering this message to the
> intended recipient, you are hereby notified that any disclosure,
> distribution, copying or other dissemination of this communication is
> strictly prohibited. If you received this communication in error,
> please immediately reply to the sender, delete the message and destroy all
copies of it.
>
> Thank You
>
>




Bausch & Lomb
150 Years of Perfecting Vision, Enhancing Life (TM)






EMAIL DISCLAIMER

Please Note: The information contained in this message may be privileged and
confidential, protected from disclosure, and/or intended only for the use of
the individual or entity named above. If the reader of this message is not
the intended recipient, or an employee or agent responsible for delivering
this message to the intended recipient, you are hereby notified that any
disclosure, distribution, copying or other dissemination of this
communication is strictly prohibited. If you received this communication in
error, please immediately reply to the sender, delete the message and
destroy all copies of it.

Thank You
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Well i have broken it down to for sure that it is iterator problems.
There are only so many things that you can comment out in there that
use iterators before you have a non functioning program however :).
Launching from a command shell with the log "-l logfile.txt" will
maintain all paths, ex share/mythtv/mysql.txt etc... and log right to
a file. I have pinpointed the first place it segfaults to main.cpp in
mythfrontend, meaningly

for (QStringList::Iterator it = verboseOpts.begin();
it != verboseOpts.end(); ++it )

Of course that line is only reached if you try to use a switch for "-v
______", but this is almost the 15th line of the app, there are only
so many things that could cause it! That verboseOpts.begin() line
seems to seg fault it. The funny thing is that QT uses that same sort
of syntax all over with .begin() and .end() in alot of files, with no
problem. If you try to make a short app using iterators like:

#include <qapplication.h>
#include <iostream>
using namespace std;

int main()
{
QStringList theLIST;
theLIST << "Hello";
theLIST << "World";
for (QStringList::Iterator it = verboseOpts.begin();it !=
verboseOpts.end(); ++it )
{
cout<<*it<<endl;
}
return 0;
}


There are no issues whatsoever with the above code, so perhaps its not
even the fact of the using the iterators, but perhaps the text being
put into the QStringList.


The next instance for myth is in globalsettings.cpp

addSelection(QObject::tr("Desktop Style"), "");
QStyleFactory factory;
QStringList list = factory.keys();
QStringList::iterator iter = list.begin();
for (; iter != list.end(); iter++ )
addSelection(*iter);

Again it will segfault at the list.begin() line which makes me think
that maybe its not actually the text being put in wrong.

Whichever the case, I made a post to the kde-cygwin list in hopes that
some of the QT experts out there can lend an ear.
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
whoops that code should for the small app should have been


int main()
{
QStringList theLIST;
theLIST << "Hello";
theLIST << "World";
for (QStringList::Iterator it = theLIST.begin();it !=
theLIST.end(); ++it )
{
cout<<*it<<endl;
}
return 0;
}


On Thu, 14 Oct 2004 13:46:36 -0500, Mario L <superm1@gmail.com> wrote:
> Well i have broken it down to for sure that it is iterator problems.
> There are only so many things that you can comment out in there that
> use iterators before you have a non functioning program however :).
> Launching from a command shell with the log "-l logfile.txt" will
> maintain all paths, ex share/mythtv/mysql.txt etc... and log right to
> a file. I have pinpointed the first place it segfaults to main.cpp in
> mythfrontend, meaningly
>
> for (QStringList::Iterator it = verboseOpts.begin();
> it != verboseOpts.end(); ++it )
>
> Of course that line is only reached if you try to use a switch for "-v
> ______", but this is almost the 15th line of the app, there are only
> so many things that could cause it! That verboseOpts.begin() line
> seems to seg fault it. The funny thing is that QT uses that same sort
> of syntax all over with .begin() and .end() in alot of files, with no
> problem. If you try to make a short app using iterators like:
>
> #include <qapplication.h>
> #include <iostream>
> using namespace std;
>
> int main()
> {
> QStringList theLIST;
> theLIST << "Hello";
> theLIST << "World";
> for (QStringList::Iterator it = verboseOpts.begin();it !=
> verboseOpts.end(); ++it )
> {
> cout<<*it<<endl;
> }
> return 0;
> }
>
> There are no issues whatsoever with the above code, so perhaps its not
> even the fact of the using the iterators, but perhaps the text being
> put into the QStringList.
>
> The next instance for myth is in globalsettings.cpp
>
> addSelection(QObject::tr("Desktop Style"), "");
> QStyleFactory factory;
> QStringList list = factory.keys();
> QStringList::iterator iter = list.begin();
> for (; iter != list.end(); iter++ )
> addSelection(*iter);
>
> Again it will segfault at the list.begin() line which makes me think
> that maybe its not actually the text being put in wrong.
>
> Whichever the case, I made a post to the kde-cygwin list in hopes that
> some of the QT experts out there can lend an ear.
>
RE: Re: compiling mythtv frontend for cygwin [ In reply to ]
Agreed. Looks to me like the problem is with QStringList::split(). As you
said, if you create a QStringList and fill it by hand everything works. If
you attempt to split a string it fails even if you put a static value in to
be split. The only thing I can think is that QStringList::split() is
returning a null value which obviously will break the iterator.

- Mark.

-----Original Message-----
From: Mario L [mailto:superm1@gmail.com]
Sent: Thursday, October 14, 2004 2:48 PM
To: Buechler, Mark R; MythTV-Developer List
Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin

whoops that code should for the small app should have been


int main()
{
QStringList theLIST;
theLIST << "Hello";
theLIST << "World";
for (QStringList::Iterator it = theLIST.begin();it != theLIST.end();
++it )
{
cout<<*it<<endl;
}
return 0;
}


On Thu, 14 Oct 2004 13:46:36 -0500, Mario L <superm1@gmail.com> wrote:
> Well i have broken it down to for sure that it is iterator problems.
> There are only so many things that you can comment out in there that
> use iterators before you have a non functioning program however :).
> Launching from a command shell with the log "-l logfile.txt" will
> maintain all paths, ex share/mythtv/mysql.txt etc... and log right to
> a file. I have pinpointed the first place it segfaults to main.cpp in
> mythfrontend, meaningly
>
> for (QStringList::Iterator it = verboseOpts.begin();
> it != verboseOpts.end(); ++it )
>
> Of course that line is only reached if you try to use a switch for "-v
> ______", but this is almost the 15th line of the app, there are only
> so many things that could cause it! That verboseOpts.begin() line
> seems to seg fault it. The funny thing is that QT uses that same sort
> of syntax all over with .begin() and .end() in alot of files, with no
> problem. If you try to make a short app using iterators like:
>
> #include <qapplication.h>
> #include <iostream>
> using namespace std;
>
> int main()
> {
> QStringList theLIST;
> theLIST << "Hello";
> theLIST << "World";
> for (QStringList::Iterator it = verboseOpts.begin();it !=
> verboseOpts.end(); ++it )
> {
> cout<<*it<<endl;
> }
> return 0;
> }
>
> There are no issues whatsoever with the above code, so perhaps its not
> even the fact of the using the iterators, but perhaps the text being
> put into the QStringList.
>
> The next instance for myth is in globalsettings.cpp
>
> addSelection(QObject::tr("Desktop Style"), "");
> QStyleFactory factory;
> QStringList list = factory.keys();
> QStringList::iterator iter = list.begin();
> for (; iter != list.end(); iter++ )
> addSelection(*iter);
>
> Again it will segfault at the list.begin() line which makes me think
> that maybe its not actually the text being put in wrong.
>
> Whichever the case, I made a post to the kde-cygwin list in hopes that
> some of the QT experts out there can lend an ear.
>




Bausch & Lomb
150 Years of Perfecting Vision, Enhancing Life (TM)






EMAIL DISCLAIMER

Please Note: The information contained in this message may be privileged and
confidential, protected from disclosure, and/or intended only for the use of
the individual or entity named above. If the reader of this message is not
the intended recipient, or an employee or agent responsible for delivering
this message to the intended recipient, you are hereby notified that any
disclosure, distribution, copying or other dissemination of this
communication is strictly prohibited. If you received this communication in
error, please immediately reply to the sender, delete the message and
destroy all copies of it.

Thank You
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Well but that theory seems to be broken because the same thing happens
in globalsettings.cpp, unless that is one problem and this is another
completely unrelated problem:

class StyleSetting: public ComboBoxSetting, public GlobalSetting {
public:
StyleSetting():
GlobalSetting("Style") {
setLabel(QObject::tr("Qt Style"));
fillSelections();
setHelpText(QObject::tr("At startup, MythTV will change the Qt "
"widget style to this setting. If \"Desktop Style\" "
"is selected, MythTV will use the existing desktop "
"setting."));
};

void fillSelections(void) {
clearSelections();
addSelection(QObject::tr("Desktop Style"), "");
QStyleFactory factory;
QStringList list = factory.keys();
QStringList::iterator iter = list.begin();
VERBOSE(VB_ALL,"iteration started");
for (; iter != list.end(); iter++ )
addSelection(*iter);
};

void load(QSqlDatabase *db) {
fillSelections();
GlobalSetting::load(db);
};
};

If you look further into that keys function, it must yield something
out, each of the entries are manually added in:

QStringList QStyleFactory::keys()
{
QStringList list;
#ifndef QT_NO_COMPONENT
if ( !instance )
instance = new QStyleFactoryPrivate;

list = QStyleFactoryPrivate::manager->featureList();
#endif //QT_NO_COMPONENT

#ifndef QT_NO_STYLE_WINDOWS
if ( !list.contains( "Windows" ) )
list << "Windows";
#endif
#ifndef QT_NO_STYLE_WINDOWSXP
if ( !list.contains( "WindowsXP" ) && QWindowsXPStyle::resolveSymbols() )
list << "WindowsXP";
#endif
#ifndef QT_NO_STYLE_MOTIF
if ( !list.contains( "Motif" ) )
list << "Motif";
#endif
#ifndef QT_NO_STYLE_CDE
if ( !list.contains( "CDE" ) )
list << "CDE";
#endif
#ifndef QT_NO_STYLE_MOTIFPLUS
if ( !list.contains( "MotifPlus" ) )
list << "MotifPlus";
#endif
#ifndef QT_NO_STYLE_PLATINUM
if ( !list.contains( "Platinum" ) )
list << "Platinum";
#endif
#ifndef QT_NO_STYLE_SGI
if ( !list.contains( "SGI" ) )
list << "SGI";
#endif
#ifndef QT_NO_STYLE_COMPACT
if ( !list.contains( "Compact" ) )
list << "Compact";
#endif
#ifndef QT_NO_STYLE_AQUA
if ( !list.contains( "Aqua" ) )
list << "Aqua";
#endif
#if !defined( QT_NO_STYLE_MAC ) && defined( Q_WS_MAC )
QString mstyle = "Macintosh";
Collection c = NewCollection();
if (c) {
GetTheme(c);
Str255 str;
long int s = 256;
if(!GetCollectionItem(c, kThemeNameTag, 0, &s, &str))
mstyle += " (" + p2qstring(str) + ")";
}
if (!list.contains(mstyle))
list << mstyle;
DisposeCollection(c);
#endif

return list;
}
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
I went a little further into this and it isn't a null string being
returned, the object does exist. I am really starting to think that
we might have hit some sort of QT bug here:

QStringList verboseOpts;
verboseOpts =
QStringList::split(',',a.argv()[argpos+1],true);
++argpos;
if (&verboseOpts == NULL)
VERBOSE(VB_ALL, "null stringlist");
else
{
VERBOSE(VB_ALL,&verboseOpts);
VERBOSE(VB_ALL,verboseOpts.size());
VERBOSE(VB_ALL,"empty? " + verboseOpts.empty());
if (verboseOpts.front() == NULL)
VERBOSE(VB_ALL, "null iterator");
else
VERBOSE(VB_ALL, "crazy iterator");
}
QStringList::Iterator it = verboseOpts.begin();
VERBOSE(VB_ALL,"it has begun");
for (QStringList::Iterator it = verboseOpts.begin();
it != verboseOpts.end(); ++it )

Adding those extra output lines gets me this:
$ /c/myth/mythfrontend.exe -v all
2004-10-14 15:57:08 0x23fdb8
2004-10-14 15:57:08 213979536
2004-10-14 15:57:08 empty?

Which does prove that it isn't a NULL QStringList, and also the fact
that the iterator list is fairly large, a little too large if i would
say. Debugging iterator stuff in QT is not going to be a very
enjoyable task I'll recon. I'm going to rebuild my QT later tonight
with debug support, and see if later on after work I can try'n use gdb
to learn a bit about the problem.

On Thu, 14 Oct 2004 15:43:17 -0500, Mario L <superm1@gmail.com> wrote:
> Well but that theory seems to be broken because the same thing happens
> in globalsettings.cpp, unless that is one problem and this is another
> completely unrelated problem:
>
> class StyleSetting: public ComboBoxSetting, public GlobalSetting {
> public:
> StyleSetting():
> GlobalSetting("Style") {
> setLabel(QObject::tr("Qt Style"));
> fillSelections();
> setHelpText(QObject::tr("At startup, MythTV will change the Qt "
> "widget style to this setting. If \"Desktop Style\" "
> "is selected, MythTV will use the existing desktop "
> "setting."));
> };
>
> void fillSelections(void) {
> clearSelections();
> addSelection(QObject::tr("Desktop Style"), "");
> QStyleFactory factory;
> QStringList list = factory.keys();
> QStringList::iterator iter = list.begin();
> VERBOSE(VB_ALL,"iteration started");
> for (; iter != list.end(); iter++ )
> addSelection(*iter);
> };
>
> void load(QSqlDatabase *db) {
> fillSelections();
> GlobalSetting::load(db);
> };
> };
>
> If you look further into that keys function, it must yield something
> out, each of the entries are manually added in:
>
> QStringList QStyleFactory::keys()
> {
> QStringList list;
> #ifndef QT_NO_COMPONENT
> if ( !instance )
> instance = new QStyleFactoryPrivate;
>
> list = QStyleFactoryPrivate::manager->featureList();
> #endif //QT_NO_COMPONENT
>
> #ifndef QT_NO_STYLE_WINDOWS
> if ( !list.contains( "Windows" ) )
> list << "Windows";
> #endif
> #ifndef QT_NO_STYLE_WINDOWSXP
> if ( !list.contains( "WindowsXP" ) && QWindowsXPStyle::resolveSymbols() )
> list << "WindowsXP";
> #endif
> #ifndef QT_NO_STYLE_MOTIF
> if ( !list.contains( "Motif" ) )
> list << "Motif";
> #endif
> #ifndef QT_NO_STYLE_CDE
> if ( !list.contains( "CDE" ) )
> list << "CDE";
> #endif
> #ifndef QT_NO_STYLE_MOTIFPLUS
> if ( !list.contains( "MotifPlus" ) )
> list << "MotifPlus";
> #endif
> #ifndef QT_NO_STYLE_PLATINUM
> if ( !list.contains( "Platinum" ) )
> list << "Platinum";
> #endif
> #ifndef QT_NO_STYLE_SGI
> if ( !list.contains( "SGI" ) )
> list << "SGI";
> #endif
> #ifndef QT_NO_STYLE_COMPACT
> if ( !list.contains( "Compact" ) )
> list << "Compact";
> #endif
> #ifndef QT_NO_STYLE_AQUA
> if ( !list.contains( "Aqua" ) )
> list << "Aqua";
> #endif
> #if !defined( QT_NO_STYLE_MAC ) && defined( Q_WS_MAC )
> QString mstyle = "Macintosh";
> Collection c = NewCollection();
> if (c) {
> GetTheme(c);
> Str255 str;
> long int s = 256;
> if(!GetCollectionItem(c, kThemeNameTag, 0, &s, &str))
> mstyle += " (" + p2qstring(str) + ")";
> }
> if (!list.contains(mstyle))
> list << mstyle;
> DisposeCollection(c);
> #endif
>
> return list;
> }
>
RE: Re: compiling mythtv frontend for cygwin [ In reply to ]
Hmm, I'm afraid I don't. Furthermore, I don't think it would even be
possible without altering g++ a bit. Admittedly I'm just guessing. I don't
have VS installed but I'll have to get it and play around. How difficult do
you think it would be to get mythtv compiled under VC++? That might fix many
problems.

- Mark.

-----Original Message-----
From: Mario L [mailto:superm1@gmail.com]
Sent: Friday, October 15, 2004 1:16 PM
To: Buechler, Mark R; MythTV-Developer List
Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin

Mark, I did a bit of expermenting with builds of QT from other compilers.
This only happens with a mingw build of QT. If you know at all how to
properly link a Visual Studio build of QT wtih a mingw build of myth then we
will be in business, as I can say VS builds QT properly.

Attached is a log of the problems that crop up trying to link a VS build of
QT to that little app that I wrote to test for the begin() bug.




Bausch & Lomb
150 Years of Perfecting Vision, Enhancing Life (TM)






EMAIL DISCLAIMER

Please Note: The information contained in this message may be privileged and
confidential, protected from disclosure, and/or intended only for the use of
the individual or entity named above. If the reader of this message is not
the intended recipient, or an employee or agent responsible for delivering
this message to the intended recipient, you are hereby notified that any
disclosure, distribution, copying or other dissemination of this
communication is strictly prohibited. If you received this communication in
error, please immediately reply to the sender, delete the message and
destroy all copies of it.

Thank You
Re: Re: compiling mythtv frontend for cygwin [ In reply to ]
Switching to all VC++ would fix problems with QT, as well as create
many others with myth. I started to see the difficulty involved, when
I tried to just use it to build from the makefile for libavcodec.
Just running nmake (the VS equivalent of make) yielded lots of
definitions for include files that just aren't there. I think leaving
it as a last resort to turn to is a better idea at this point. I
found some things about linking VC++ dlls with mingw online, but
nothing has really seemed to work yet. I think for now we're just at
a wall deciding what compiler to go with and route to go with then.

I'm waiting from the kde-cygwin list about some more ideas about
linking the QT dll built in VC++ with cygwin. As of right now that is
ideal. It is quite clear that there is a bug involving iterators,
however; when building QT with mingw. If you or anyone else on the
list has some input what we could do then feel free to voice.

BTW, I believe that the compiler (not the IDE) for VC++ is freely
avail, if you just want to build QT with it.


On Fri, 15 Oct 2004 19:47:16 -0400, Buechler, Mark R
<mark_r_buechler@bausch.com> wrote:
> Hmm, I'm afraid I don't. Furthermore, I don't think it would even be
> possible without altering g++ a bit. Admittedly I'm just guessing. I don't
> have VS installed but I'll have to get it and play around. How difficult do
> you think it would be to get mythtv compiled under VC++? That might fix many
> problems.
>
>
>
> - Mark.
>
> -----Original Message-----
> From: Mario L [mailto:superm1@gmail.com ]
> Sent: Friday, October 15, 2004 1:16 PM
> To: Buechler, Mark R; MythTV-Developer List
> Subject: Re: [mythtv] Re: compiling mythtv frontend for cygwin
>
> Mark, I did a bit of expermenting with builds of QT from other compilers.
> This only happens with a mingw build of QT. If you know at all how to
> properly link a Visual Studio build of QT wtih a mingw build of myth then we
> will be in business, as I can say VS builds QT properly.
>
> Attached is a log of the problems that crop up trying to link a VS build of
> QT to that little app that I wrote to test for the begin() bug.
>
> Bausch & Lomb
> 150 Years of Perfecting Vision, Enhancing Life (TM)
>
> EMAIL DISCLAIMER
>
> Please Note: The information contained in this message may be privileged and
> confidential, protected from disclosure, and/or intended only for the use of
> the individual or entity named above. If the reader of this message is not
> the intended recipient, or an employee or agent responsible for delivering
> this message to the intended recipient, you are hereby notified that any
> disclosure, distribution, copying or other dissemination of this
> communication is strictly prohibited. If you received this communication in
> error, please immediately reply to the sender, delete the message and
> destroy all copies of it.
>
> Thank You
>
>

1 2 3  View All