New winetricks 20080226: adds dotnet11 verb

[Dan Kegel]

In honor of .net 1.1 installing on today's git, here's
a version of winetricks that knows how to grab and
run the installer.   Amusingly, I noticed that there
are dozens of scripts out there that do this... it seems
to be not uncommon in the windows world.

As always, winetricks is at
http://kegel.com/wine/winetricks

and its source repository is at
http://winezeug.googlecode.com

BTW if anybody has any other small whiny things that
almost but not quite fit in winehq, I'd be happy to
consider hosting them in the winezeug repository.
- Dan

-- 
Wine for Windows ISVs: http://kegel.com/wine/isv

[bill lam]

Thank you for update. I also notice that the version of comctl32 and gdiplus 
that installed from winetricks are quite old. I have to manually copy those 
files from my win-xp to make my application work.  Could you also consider 
revise them?

Dan Kegel wrote:
> In honor of .net 1.1 installing on today's git, here's
> a version of winetricks that knows how to grab and

[Dan Kegel]

On Tue, Feb 26, 2008 at 5:25 PM, bill lam <cbill.lam(a)gmail.com> wrote:
> Thank you for update. I also notice that the version of comctl32 and gdiplus
>  that installed from winetricks are quite old. I have to manually copy those
>  files from my win-xp to make my application work.  Could you also consider
>  revise them?

Which apps?
What version of comctl32 and gdiplus do they require?

According to
http://msdn2.microsoft.com/en-us/library/bb776779(VS.85).aspx
comctl32 version 6 is nonredistributable, which means
it might be a bit of a challenge to find a free download from
microsoft.com that offers it.
- Dan

[James McKenzie]

Dan Kegel wrote:
> On Tue, Feb 26, 2008 at 5:25 PM, bill lam <cbill.lam(a)gmail.com> wrote:
> > Thank you for update. I also notice that the version of comctl32 and gdiplus
> >  that installed from winetricks are quite old. I have to manually copy those
> >  files from my win-xp to make my application work.  Could you also consider
> >  revise them?
>
> Which apps?
> What version of comctl32 and gdiplus do they require?
>
> According to
> http://msdn2.microsoft.com/en-us/library/bb776779(VS.85).aspx
> comctl32 version 6 is nonredistributable, which means
> it might be a bit of a challenge to find a free download from
> microsoft.com that offers it.
> - Dan

Dan:

What version of comctl32 is packaged with VB6SP5 (if I remember this 
correctly, the files are on my work machine which I cannot access from 
here)?  That package is freely redistributable from what I remember and 
is available from Microsoft.

James McKenzie

[bill lam]

Dan Kegel wrote:
> Which apps?
> What version of comctl32 and gdiplus do they require?

depends.exe show the following info for comctl32

COMCTL32.DLL from win-xp
File Ver     5.82.2900.2180
Product Ver  6.0.2900.2180

COMCTL32.DLL from cc580
File Ver     5.80.2614.3600
Product Ver  5.0.2614.3600

The last time that I used the old comctl32, the bmp inside toolbar cannot 
display, just black holes. I tested again today, it work correctly even with the 
old comctl32.  May be I forgot exactly where it failed or may be I had updated 
wine from .53 to .55. Anyway I will stick to the old comctl32 and report failure 
if it happen again.

regards,

[bill lam]

Dan Kegel wrote:
> Which apps?
> What version of comctl32 and gdiplus do they require?

I check version of gdiplus from pp viewer 2003 (winetricks) and that from 
win-xp, they are the same version. I must had use my old version from win-me 
that caused problem. Sorry for the noise.

In current wine version *.55, the built-in version of comctl32 does work for my 
app so that I do not need the native comctl32 anymore. Now only odbc and gdiplus 
need to be overrided.

Thank you to all wine developers.

regards,

[Dan Kegel]

On Wed, Feb 27, 2008 at 7:00 AM, bill lam <cbill.lam(a)gmail.com> wrote:
> In current wine version *.55, the built-in version of comctl32 does work for my
>  app so that I do not need the native comctl32 anymore. Now only odbc and gdiplus
>  need to be overrided.

What app is this, and what is it about Wine's gdiplus that
makes you need to override it?  Does the app crash,
or does it just look bad?  If it crashes, what's the crash log?

[bill lam]

Dan Kegel wrote:
> What app is this, and what is it about Wine's gdiplus that
> makes you need to override it?  Does the app crash,
> or does it just look bad?  If it crashes, what's the crash log?

The app is a package called 'platimg' of J language.

I had reported this and I believed that it is the negative stride that caused 
the problem.

NB. load and decode a jpeg file into 2D integer matrix with each integer
NB. for one pixel element.
readimg1 'picture.jpg'

where definition of readimg1 is as follows

readimg1=: 3 : 0
   GdiplusStartup (TOK=.,_1);GdiplusStartupInput;0
   GdipCreateBitmapFromFile y;BMP=.,_1
   z=. bmpARGB {.BMP
   GdiplusShutdown {.TOK
   z
)

bmpARGB=: 3 : 0
   DATA=. i.4%~#BitmapData
   GdipBitmapLockBits y;0;ImageLockModeRead;PixelFormat32bppARGB;DATA
   'w h s f p r'=. DATA
   z=. (h,w)$memr p,0,(w*h),JINT
   GdipBitmapUnLockBits y;DATA
   GdipDisposeImage y
   z
)

Here GdipBitmapLockBits has 2 issues,

1. GdipBitmapLockBits accept a null for rect (the second argument) to stands for 
the entire region. I can workaround this by calling GdipGetImageWidth and 
GdipGetImageHeight to construct the rect. However

2. the data returned GdipBitmapLockBits does not point a valid bitmap data, so 
the when trying memory read in the following
z=. (h,w)$memr p,0,(w*h),JINT
J gives raise a domain error. (but no crash, so that no crash log)

Scripts of J can be difficult to understand, but since this example calls 
gdiplus api so that you can follow it without much difficulty and write an 
equivalent c program to test it.

current version of J can be downloaded here.
http://www.jsoftware.com/beta.htm
after installation, run package manager that comes with J to install the platimg 
addon.

regards,

[Dan Kegel]

On Wed, Feb 27, 2008 at 8:48 AM, bill lam <cbill.lam(a)gmail.com> wrote:
> The app is a package called 'platimg' of J language.
>
> I had reported this and I believed that it is the negative stride that caused
>  the problem.

Where did you report it?  I can't find a bug report.

Would you like to file a Wine bug, or should I?
> NB. load and decode a jpeg file into 2D integer matrix with each integer
>  NB. for one pixel element.
>  readimg1 'picture.jpg'
>
> where definition of readimg1 is as follows
>
> readimg1=: 3 : 0
>    GdiplusStartup (TOK=.,_1);GdiplusStartupInput;0
>    GdipCreateBitmapFromFile y;BMP=.,_1
>    z=. bmpARGB {.BMP
>    GdiplusShutdown {.TOK
>    z
>  )
>
> bmpARGB=: 3 : 0
>    DATA=. i.4%~#BitmapData
>    GdipBitmapLockBits y;0;ImageLockModeRead;PixelFormat32bppARGB;DATA
>    'w h s f p r'=. DATA
>    z=. (h,w)$memr p,0,(w*h),JINT
>    GdipBitmapUnLockBits y;DATA
>    GdipDisposeImage y
>    z
>  )
>
> Here GdipBitmapLockBits has 2 issues,
>
> 1. GdipBitmapLockBits accept a null for rect (the second argument) to stands for
>  the entire region. I can workaround this by calling GdipGetImageWidth and
>  GdipGetImageHeight to construct the rect. However
>
> 2. the data returned GdipBitmapLockBits does not point a valid bitmap data, so
>  the when trying memory read in the following
>  z=. (h,w)$memr p,0,(w*h),JINT
>  J gives raise a domain error. (but no crash, so that no crash log)
>
> Scripts of J can be difficult to understand, but since this example calls
>  gdiplus api so that you can follow it without much difficulty and write an
>  equivalent c program to test it.
>
> current version of J can be downloaded here.
>  http://www.jsoftware.com/beta.htm
>  after installation, run package manager that comes with J to install the platimg
>  addon.

Cool, thanks.  I also see it's documented here:
http://olegykj.sourceforge.net/scripts/platimg.htm
- Dan

[bill lam]

Dan Kegel wrote:
> Where did you report it?  I can't find a bug report.
> not formal report, just post to this forum days ago.
> Would you like to file a Wine bug, or should I?
> Could you please file a Wine bug. It is easier for you to write an equivalent C 
> script to reproduce the bug than I. And I'm not sure if any wine developers will 
> be able to testing using J if I report it in the current format.

regards,

[bill lam]

I patched platimg to work with negative stride and now it works with wine 
built-in gdiplus.

FYI, native gdiplus gives a top-down bitmap but wine's gdiplus gives a 
bottom-up. It is not a bug in wine's but might causes compatibility problem.

[James Hawkins]

On Thu, Feb 28, 2008 at 10:07 PM, bill lam <cbill.lam(a)gmail.com> wrote:
> I patched platimg to work with negative stride and now it works with wine
>  built-in gdiplus.
>
> FYI, native gdiplus gives a top-down bitmap but wine's gdiplus gives a
>  bottom-up. It is not a bug in wine's but might causes compatibility problem.

How is it not a bug?

-- 
James Hawkins

[bill lam]

James Hawkins wrote:
> On Thu, Feb 28, 2008 at 10:07 PM, bill lam <cbill.lam(a)gmail.com> wrote:
> > I patched platimg to work with negative stride and now it works with wine
> >  built-in gdiplus.
> >
> > FYI, native gdiplus gives a top-down bitmap but wine's gdiplus gives a
> >  bottom-up. It is not a bug in wine's but might causes compatibility problem.
>
> How is it not a bug?

If the user beware this possibility and write a more secure code, 
GdipBitmapLockBits from wine's gdiplus will work.

However this is another bug in wine gdiplus in that
value return from GdipGetImageWidth ( and GdipGetImageHeight ?) will often be 1 
pixel less than actual value, eg a jpeg of size 800x600 will be reported as 
799x600. The larger the picture size, the more often it goes wrong.

regards,

[Lei Zhang]

On Fri, Feb 29, 2008 at 1:51 AM, bill lam <cbill.lam(a)gmail.com> wrote:
> However this is another bug in wine gdiplus in that
>  value return from GdipGetImageWidth ( and GdipGetImageHeight ?) will often be 1
>  pixel less than actual value, eg a jpeg of size 800x600 will be reported as
>  799x600. The larger the picture size, the more often it goes wrong.

Can you file a bug on bugzilla and provide a sample image / test code
to demonstrate this problem?

[bill lam]

Lei Zhang wrote:
> Can you file a bug on bugzilla and provide a sample image / test code
> to demonstrate this problem?

Done.

Test code is as follows, beware of line wrap.

NB. tested on J602 beta / wine 0.9.56
NB. script to demostrate bug in GdipGetImageWidth GdipGetImageHeight
NB. =========================================================
require 'dll'

s=. 'Ok GenericError InvalidParameter OutOfMemory ObjectBusy InsufficientBuffer '
s=. s,'NotImplemented Win32Error WrongState Aborted FileNotFound ValueOverflow '
s=. s,'AccessDenied UnknownImageFormat FontFamilyNotFound FontStyleNotFound '
s=. s,'NotTrueTypeFont UnsupportedGdiplusVersion GdiplusNotInitialized '
s=. ;:s,'PropertyNotFound PropertyNotSupported ProfileNotFound'

assi=: {::&s assert 0=]

cdi=: (&cd)(assi@)

PixelFormat32bppARGB=: 16b26200a

GdiplusStartup           =: 'gdiplus GdiplusStartup            > i   *i *i i 
      ' &cd
GdiplusShutdown          =: 'gdiplus GdiplusShutdown           > n   i 
      ' &cd

GdipCreateBitmapFromFile =: 'gdiplus GdipCreateBitmapFromFile  > i   *w *i 
      ' cdi
GdipDisposeImage         =: 'gdiplus GdipDisposeImage          > i   i 
      ' cdi

GdipBitmapLockBits       =: 'gdiplus GdipBitmapLockBits        > i   i *i i i *i 
      ' &cd
GdipBitmapUnLockBits     =: 'gdiplus GdipBitmapUnlockBits      > i   i *i 
      ' cdi

GdipGetImageWidth        =: 'gdiplus GdipGetImageWidth         >  i    i  *i 
       ' &cd
GdipGetImageHeight       =: 'gdiplus GdipGetImageHeight        >  i    i  *i 
       ' &cd

ImageLockModeRead=: 1
GUID=: 'WWWWXXYYZZZZZZZZ'
BitmapData=: 'WdthHghtStrdFrmtScanRsrv'
GdiplusStartupInput=: 1 0 0 0

readimg1=: 3 : 0
DATA=. i.4%~#BitmapData
assi GdiplusStartup (TOK=. ,_1);GdiplusStartupInput;0
assi GdipCreateBitmapFromFile y;(BMP=. ,_1)
assi GdipGetImageWidth ({.BMP);(Width=. ,_1)
assi GdipGetImageHeight ({.BMP);(Height=. ,_1)
1!:2&2 'GdipGetImageWidth GdipGetImageHeight ', ": Width,Height
rt=. 0 0, Width, Height
assi GdipBitmapLockBits ({.BMP);rt;ImageLockModeRead;PixelFormat32bppARGB;DATA
'w h s f p r'=. DATA
1!:2&2 'width height stride format scan0 res ', ":DATA
z=. |.^:(s<0) w&{."1^:(w~:|s%4) (h,|s%4)$memr p,((s<0)*s*<:h),(h*<.|s%4),JINT
GdipBitmapUnLockBits ({.BMP);DATA
GdipDisposeImage {.BMP
GdiplusShutdown {.TOK
''
)
NB. =========================================================


NB. run the script
readimg1 'z:/tmp/2083.jpg'

[James McKenzie]

bill lam wrote:
> If the user beware this possibility and write a more secure code, 
> GdipBitmapLockBits from wine's gdiplus will work.
>
> However this is another bug in wine gdiplus in that
> value return from GdipGetImageWidth ( and GdipGetImageHeight ?) will often be 1 
> pixel less than actual value, eg a jpeg of size 800x600 will be reported as 
> 799x600. The larger the picture size, the more often it goes wrong.

This is a bug and should be reported.  If this happens for any jpeg file 
then an attached file is unnecessary, but if you provide one that is 
'more better' and will help.

James McKenzie