Extracting a short using struct - won't print.

[Graham Nicholls]

Thanks to Richard and Alex, I'm doing this:

seg_stru=">BHH"
seg_data=self.f hand.read(struc t.calcsize(seg_ stru))
data=struct.unp ack(seg_stru,se g_data)
x=seg_data[1]
y=seg_data[2]
if debug:
print ("DEBUG: Image [%s] is [%d] by [%d] pixels" % (self.fname,x,y ))

Expecting to extract a filler byte (ignored), then a big-endian short (a
python integer) , then another of the same. When I try to print, I get
this:

File "./pdf_merge", line 426, in getsize
print ("DEBUG: Image [%s] is [%d] by [%d] pixels" % (self.fname,x,y ))
TypeError: int argument required

What am I missing - its probably something simple - earlier I spent half an
hour wondering why I was getting a message that "img_file has no object
getsize()" from x,y=img.getsize (). It turned out to be because I had got
def getsize: indented one indent too far in the class definition. Aaargh!

Python is 2.3b1.

Thanks
--
Graham Nicholls
Rock Computer Consultancy

[Graham Nicholls]

Peter Hansen wrote:

Graham Nicholls wrote:

Thanks to Richard and Alex, I'm doing this:

seg_stru=">BHH"
seg_data=self.f hand.read(struc t.calcsize(seg_ stru))
data=struct.unp ack(seg_stru,se g_data)
x=seg_data[1]
y=seg_data[2]
Why are you unpacking the struct if you aren't going to use the
unpacked data?

Err, I'm trying to extract two shorts - the x and y size in pixels of an
image, which I'll use to scale it to fit another pair of parameters. What
makes you think I'm not using the data? Am I not - I thought I was!

I realise that ">BHH" above ouhjt really to be "xHH" but aside from that, I
don't understand your point - theres more code after the debug in the rel
app - returning x and y.

if debug:
print ("DEBUG: Image [%s] is [%d] by [%d] pixels" %
(self.fname,x,y ))
Best way to troubleshoot this kind of thing, in my experience,
is to insert "import pdb; pdb.set_trace() " into the code just
above the failing line, then use the debugger to inspect the
values directly just before they are used. (You need to execute
the "r" or "return" command immediately after the debugger is
entered... the rest is pretty straightforward to learn.)

I'll give that a go, thanks -Peter

--
Graham Nicholls
Rock Computer Consultancy

[Richard Brodie]

"Graham Nicholls" <gr****@rockcon s.co.uk> wrote in message
news:3f******** *************** @auth.uk.news.e asynet.net...

seg_stru=">BHH"
seg_data=self.f hand.read(struc t.calcsize(seg_ stru))
data=struct.unp ack(seg_stru,se g_data)
x=seg_data[1]
y=seg_data[2]

Why are you unpacking the struct if you aren't going to use the
unpacked data?

Err, I'm trying to extract two shorts - the x and y size in pixels of an
image, which I'll use to scale it to fit another pair of parameters. What
makes you think I'm not using the data? Am I not - I thought I was!

You aren't using the returned value from struct.unpack() anywhere.
You need something like x, y = struct.unpack(s eg_stru,seg_dat a)

[Graham Nicholls]

Graham Nicholls wrote:

Peter Hansen wrote:

Graham Nicholls wrote:

Thanks to Richard and Alex, I'm doing this:

seg_stru=">BHH"
seg_data=self.f hand.read(struc t.calcsize(seg_ stru))
data=struct.unp ack(seg_stru,se g_data)
x=seg_data[1]
y=seg_data[2]
Why are you unpacking the struct if you aren't going to use the
unpacked data?

Err, I'm trying to extract two shorts - the x and y size in pixels of an
image, which I'll use to scale it to fit another pair of parameters. What
makes you think I'm not using the data? Am I not - I thought I was!

I realise that ">BHH" above ouhjt really to be "xHH" but aside from that,
I don't understand your point - theres more code after the debug in the
rel app - returning x and y.

if debug:
print ("DEBUG: Image [%s] is [%d] by [%d] pixels" %
(self.fname,x,y ))

Best way to troubleshoot this kind of thing, in my experience,
is to insert "import pdb; pdb.set_trace() " into the code just
above the failing line, then use the debugger to inspect the
values directly just before they are used. (You need to execute
the "r" or "return" command immediately after the debugger is
entered... the rest is pretty straightforward to learn.)

I'll give that a go, thanks

-Peter

OK, the code now looks like this:

seg_stru=">xHH"
seg_data=self.f hand.read(struc t.calcsize(seg_ stru))
data=struct.unp ack(seg_stru,se g_data)

x=seg_data[1]
y=seg_data[2]
import pdb ; pdb.set_trace()
if debug:
print ("DEBUG: Image [%s] is [%d] by [%d] pixels" % (self.fname,x,y ))
return(x,y)

When I run it, I get this:

EBUG: Scaling image file hb.jpg to fit 100 100
DEBUG: Searching hb.jpg for dimensions
DEBUG: found header type e0
DEBUG: found header type db
DEBUG: found header type db
DEBUG: found header type c0
--Return-- /usr/local/lib/python2.3/pdb.py(972)set_ trace()->None -> Pdb().set_trace ()
(Pdb)
(Pdb) r /home/graham/work/hsb/pdflive/pdf_merge(427)g etsize()

-> if debug:
(Pdb) print x

(Pdb) print y
m
(Pdb)

Which _seems_ to indicate that unpack is not unpacking the shorts as shorts,
but as chars. Is this right?

Thanks
Graham
--
Graham Nicholls
Rock Computer Consultancy

[Graham Nicholls]

Graham Nicholls wrote:

Richard Brodie wrote:

....snipped

I told you it'd be something stupid!
I mean
x=data[1]
y=data[2]
don't I!
Thanks!!!!
Graham

Which, of course works - one of the annoying things about learning a
language is that you tend to assume its something more complex than it
really is when you get an error. I'd see that in 2 seconds flat in a c
program , of course, but then the compiler would help..

BTW How does one test python programs - because its interpreted, it often
won't run bits of code till unexpected circumstances, so a syntax error can
be hidden until runtime?

Graham
--
Graham Nicholls
Rock Computer Consultancy

[Graham Nicholls]

Peter Hansen wrote:

Graham Nicholls wrote:

don't I!
Yes, you do! :-) Sorry for not being more obvious about the answer,
and I thought you might figure it out yourself by using the "import pdb"
trick I showed. "Teach a man to fish" and all that. <grin>

Too tired to fish, but thanks, I suspect will be a useful technique.
Graham
-Peter

--
Graham Nicholls
Rock Computer Consultancy