Loh, Joe | 6 Nov 17:23 2006

Missing DEV_SD1_MAJOR in dtable.cc

Thank you so much Eric for setting us straight on submitting patches,
and Corinna for making change before we had a chance to followup with
Eric's recommendation.

We will be sorting out the necessary documentation in order to
submitting the non-trivial patch for mapping up to 64 volumes, as
pointed out by Eric.

One other behavior we noticed is that /proc/partition currently shows
multiple entries of /dev/sdz if number of volumes exceeds 26.  It's a
pretty benign behavior and we don't have any recommended patches.

Thank you all for your support of Cygwin.

Joe Loh

Christopher Faylor | 6 Nov 17:31 2006

Re: Missing DEV_SD1_MAJOR in dtable.cc

On Mon, Nov 06, 2006 at 10:23:15AM -0600, Loh, Joe wrote:
>Thank you so much Eric for setting us straight on submitting patches,
>and Corinna for making change before we had a chance to followup with
>Eric's recommendation.
>
>We will be sorting out the necessary documentation in order to
>submitting the non-trivial patch for mapping up to 64 volumes, as
>pointed out by Eric.
>
>One other behavior we noticed is that /proc/partition currently shows
>multiple entries of /dev/sdz if number of volumes exceeds 26.  It's a
>pretty benign behavior and we don't have any recommended patches.
>
>Thank you all for your support of Cygwin.

You're welcome, but since this isn't actually a patch, please don't
clutter the mailing list with this kind of thing.

cgf

Loh, Joe | 13 Nov 23:42 2006

Patch to mapping up to 128 SCSI Disk Devices


This is a modified patch for up to 128 SCSI Disk Devices as discussed in
http://cygwin.com/ml/cygwin/2006-11/msg00060.html.

As suggested by Eric Blake, we have snail mailed the copyright
assignment to Rose Naftaly.

ChangeLog for winsup/cygwin:

2006-11-13  Joe Loh  <joel at pivot3 dot com>

        * devices.h: Add additional SCSI disk block device numbers per
        http://www.kernel.org/pub/linux/docs/device-list/devices.txt 
        up to 128 devices.
        (DEV_SD{2..7}_MAJOR): Define.
        (FH_SD{2..7}): Define.
        (FH_SDA{A..Z}): Define.
        (FH_SDB{A..Z}): Define.
        (FH_SDC{A..Z}): Define.
        (FH_SDD{A..X}): Define.
        * devices.in: Add additional SCSI disk block device numbers per
        http://www.kernel.org/pub/linux/docs/device-list/devices.txt 
        up to 128 devices.
        (/dev/sda{a..z}): Define.
        (/dev/sdb{a..z}): Define.
        (/dev/sdc{a..z}): Define.
        (/dev/sdd{a..x}): Define.
        (device::parsedisk): Add additonal else-if cases for decoding
base 
        and drive indices.
(Continue reading)

Corinna Vinschen | 14 Nov 11:02 2006

Re: Patch to mapping up to 128 SCSI Disk Devices

On Nov 13 16:42, Loh, Joe wrote:
> 
> This is a modified patch for up to 128 SCSI Disk Devices as discussed in
> http://cygwin.com/ml/cygwin/2006-11/msg00060.html.
> 
> As suggested by Eric Blake, we have snail mailed the copyright
> assignment to Rose Naftaly.

This will take a couple of days, probably.  In the meantime...

> Index: devices.h
> ===================================================================

I must admit that I don't quite understand why that happens, but
when I save your patch into a file, all '=' characters are converted
into a '=3D' sequence.  This is a bit weird given that you're using
us-ascii encoding.  Does anybody know why this happens?

> +++ devices.in  13 Nov 2006 22:30:44 -0000
>  <at>  <at>  -85,7 +85,15  <at>  <at> 
>  "/dev/scd%(0-15)d", BRACK(FHDEV(DEV_CDROM_MAJOR, {$1})),
> "\\Device\\CdRom{$1}"
>  "/dev/sr%(0-15)d", BRACK(FHDEV(DEV_CDROM_MAJOR, {$1})),
> "\\Device\\CdRom{$1}"
>  "/dev/sd%{a-z}s", BRACK(FH_SD{uc $1}), "\\Device\\Harddisk{ord($1) -
> ord('a')}\\Partition0"
> +"/dev/sda%{a-z}s", BRACK(FH_SDA{uc $1}), "\\Device\\Harddisk{26 +
> ord($1) - ord('a')}\\Partition0"
> +"/dev/sdb%{a-z}s", BRACK(FH_SDB{uc $1}), "\\Device\\Harddisk{52 +
> ord($1) - ord('a')}\\Partition0"
(Continue reading)

Brian Dessent | 14 Nov 11:25 2006
Picon

Re: Patch to mapping up to 128 SCSI Disk Devices

Corinna Vinschen wrote:

> I must admit that I don't quite understand why that happens, but
> when I save your patch into a file, all '=' characters are converted
> into a '=3D' sequence.  This is a bit weird given that you're using
> us-ascii encoding.  Does anybody know why this happens?

That's because of:

> Content-Transfer-Encoding: quoted-printable

..but your email client should undo the encoding if you tell it to save
the message as a file.  Otherwise:

perl -MMIME::QuotedPrint -ne 'print decode_qp($_)' <in >out

> The patch is also broken due to unexpected line breaks, see above.

That's always a pain... attachments are really the way to go.

Brian

Corinna Vinschen | 14 Nov 11:54 2006

Re: Patch to mapping up to 128 SCSI Disk Devices

On Nov 14 02:25, Brian Dessent wrote:
> Corinna Vinschen wrote:
> 
> > I must admit that I don't quite understand why that happens, but
> > when I save your patch into a file, all '=' characters are converted
> > into a '=3D' sequence.  This is a bit weird given that you're using
> > us-ascii encoding.  Does anybody know why this happens?
> 
> That's because of:
> 
> > Content-Transfer-Encoding: quoted-printable
> 
> ..but your email client should undo the encoding if you tell it to save
> the message as a file.  Otherwise:
> 
> perl -MMIME::QuotedPrint -ne 'print decode_qp($_)' <in >out

Uh, thanks for the explanation.  I don't know why mutt doesn't convert
this...

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Loh, Joe | 14 Nov 18:52 2006

Re: Patch to mapping up to 128 SCSI Disk Devices


> > +  else if (drive < 112)          // /dev/sdcs -to- /dev/sddh
> > +    {
> > +      base = DEV_SD6_MAJOR;
> > +      drive -= 96;
> > +    }
> > +  // NOTE: This will cause multiple /dev/sddx entries in
> > +  //       /proc/partitions if there are more than 128 devices
>
> Any problem to fix that and to get rid of this comment?  If not,
> can you please convert the comment to C-style /**/?

We did not attempt to change the implementation due to lack of fully
understanding the implications of how this function is called.  We did
noticed the behavior in /proc/partition when we first attempted to
connect using the stock 1.5.19 cygwin.  Hence, we felt it may help
others to document the behavior.  Our intent is to increase the device
mapping with minimal change to existing implementation.

We have converted all comments to use C-style /**/ and attached the
patch as a file instead.

Thanks,
Joe
Attachment (patch.jl-061114): application/octet-stream, 12 KiB
Thomas Wolff | 22 Nov 18:20 2006
Picon

[Patch] bug # 514 deja vu (cygwin console color handling)

I noticed that the reverse color bug
 http://sourceware.org/bugzilla/show_bug.cgi?id=514
shows up again in recent cygwin1.dll updates. My previous patch 
is still in the source but additional code apparently has the same 
effect of rendering output unreadable; the effect is the following:
* foreground is set bright
* screen mode is set to reverse
* cygwin wrongly assumes that the reverse foreground colour (which 
  actually used to be the non-bright background color) should be 
  set to bright, which is obviously a wrong idea and often results 
  in a contrast that renders the output almost unreadable

The attached shell script test514 demonstrates the bug.
The attached patch is an attempt to fix the bug again.
Unfortunately, I could not compile it due to the following mysterious 
make error:

make[4]: Entering directory `/usr/src/cygwin-1.5.22-patch/i686-pc-cygwin/newlib'
rm -f libm.a
ln libm/libm.a libm.a >/dev/null 2>/dev/null || cp libm/libm.a libm.a
rm -rf libc.a libg.a tmp
mkdir tmp
cd tmp; \
         ar x ../libm.a lib_a-s_isinf.o lib_a-sf_isinf.o lib_a-s_isnan.o lib_a-sf_isnan.o lib_a-s_isinfd.o
lib_a-sf_isinff.o lib_a-s_isnand.o lib_a-sf_isnanf.o lib_a-s_nan.o lib_a-sf_nan.o
lib_a-s_ldexp.o lib_a-sf_ldexp.o lib_a-s_frexp.o lib_a-sf_frexp.o lib_a-s_modf.o
lib_a-sf_modf.o lib_a-s_scalbn.o lib_a-sf_scalbn.o lib_a-s_finite.o lib_a-sf_finite.o
lib_a-s_copysign.o lib_a-sf_copysign.o lib_a-s_infconst.o ; \
         ar x ../libc/libc.a ; \
         ar rc ../libc.a *.o
(Continue reading)

Corinna Vinschen | 23 Nov 11:00 2006

Re: Patch to mapping up to 128 SCSI Disk Devices

On Nov 14 11:52, Loh, Joe wrote:
> [...]
> We have converted all comments to use C-style /**/ and attached the
> patch as a file instead.

Your copyright assignment has arrived and got signed, so I took the
opportunity to apply your patch.  Thanks.

Looking forward to see more patches,
Corinna

P.S.: Just a friendly reminder, you're all the time using "we" in the
context of this patch.  I've positively assumed now that this patch has
been created by yourself.  If other developers in your company want to
provide patches, please make sure they all have a copyright assignment
in place.  I do not like these legal hassles, but I have to follow them,
sorry.

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

Corinna Vinschen | 23 Nov 11:08 2006

Re: [Patch] bug # 514 deja vu (cygwin console color handling)

On Nov 22 18:20, Thomas Wolff wrote:
> I noticed that the reverse color bug
>  http://sourceware.org/bugzilla/show_bug.cgi?id=514
> shows up again in recent cygwin1.dll updates. My previous patch 
> is still in the source but additional code apparently has the same 
> effect of rendering output unreadable; the effect is the following:
> * foreground is set bright
> * screen mode is set to reverse
> * cygwin wrongly assumes that the reverse foreground colour (which 
>   actually used to be the non-bright background color) should be 
>   set to bright, which is obviously a wrong idea and often results 
>   in a contrast that renders the output almost unreadable
> 
> The attached shell script test514 demonstrates the bug.
> The attached patch is an attempt to fix the bug again.
> [...]
> * fhandler_console.cc (set_color): Avoid (again) inappropriate intensity 
>      interchanging that used to render reverse output unreadable 
>      when (non-reversed) text is bright.
>      See http://sourceware.org/bugzilla/show_bug.cgi?id=514

Yup, that seems to do the trick.  Thanks, applied.

Corinna

--

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

(Continue reading)


Gmane