[ python-Feature Requests-1351692 ] Switch to make pprint.pprint display ints and longs in hex

SourceForge.net noreply at sourceforge.net
Thu Apr 6 20:15:55 CEST 2006


Feature Requests item #1351692, was opened at 2005-11-08 16:29
Message generated for change (Settings changed) made by gvanrossum
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=355470&aid=1351692&group_id=5470

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: Python Library
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Mark Hirota (markhirota)
>Assigned to: Anthony Baxter (anthonybaxter)
Summary: Switch to make pprint.pprint display ints and longs in hex

Initial Comment:
It would be nice to have some sort of switch or hook to 
allow 'pretty-printing' of integers and long integers in 
hexidecimal. So, for example:

>>> import pprint
>>> pprint.pprint(range(10)) # instead of this:
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
>>> pprint.hexint = True
>>> pprint.pprint(range(10)) # you would get this:
[0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9]
>>> pprint.pprint(range(0x100000000,0x100000010)) # 
and this:
[0x100000000L,
 0x100000001L,
 0x100000002L,
 0x100000003L,
 0x100000004L,
 0x100000005L,
 0x100000006L,
 0x100000007L,
 0x100000008L,
 0x100000009L,
 0x10000000AL,
 0x10000000BL,
 0x10000000CL,
 0x10000000DL,
 0x10000000EL,
 0x10000000FL]
>>>

Thanks,
--MH

----------------------------------------------------------------------

Comment By: Guido van Rossum (gvanrossum)
Date: 2006-04-06 14:13

Message:
Logged In: YES 
user_id=6380

IMO yes. Assigning to Anthony for pronouncement (it's a
pretty small feature as features go). Feature freeze isn't
until beta 1 is shipped (although for bigger features we'd
like to see them in, in rough form, sooner, of course).

Anthony, feel free to assign back to Walter to check it in
if you're ok with that.

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2006-04-06 09:02

Message:
Logged In: YES 
user_id=89016

Can this still be checked in after 2.5a1 is out the door? It
is a new feature not just a bug fix.

----------------------------------------------------------------------

Comment By: Guido van Rossum (gvanrossum)
Date: 2006-04-05 19:11

Message:
Logged In: YES 
user_id=6380

What's this waiting for?

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2005-11-22 15:56

Message:
Logged In: YES 
user_id=89016

I think this patch can go in. Any objections from Raymond or
Fred against this?

----------------------------------------------------------------------

Comment By: Mark Hirota (markhirota)
Date: 2005-11-16 20:04

Message:
Logged In: YES 
user_id=1375527

Walter, I agree that there still seems to be an issue with 
_format() calling _repr() multiple times, however this 
patch seems to enable my feature request using sub-
classing.  Can the other issue (which exists presently 
also) be ignored for now?  Could you submit the patch, or 
is that something I should do?  Thanks.

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2005-11-16 13:29

Message:
Logged In: YES 
user_id=89016

Here's your modified version as a patch against svn HEAD.
(Most of the patch is indentation changes. "svn diff
--diff-cmd diff -x -b Lib/pprint.py" gives this diff:

134d133
<         if sepLines:
154a154
>                         if sepLines:
155a156,157
>                         else:
>                             write(', %s: ' % rep)
180a183
>                         if sepLines:
181a185,186
>                         else:
>                             write(', ')
190d194
< 
192a197
>

The patch looks good, but it still has a problem that the
original pprint has: For each call to format() the object is
formatted twice, once to determine if it fits on one line
and once for the final output.

----------------------------------------------------------------------

Comment By: Mark Hirota (markhirota)
Date: 2005-11-14 13:33

Message:
Logged In: YES 
user_id=1375527

Did some digging into the code and found that the "if 
sepLines:" condition in the PrettyPrinter._format() method 
was the source of the limitation.

I've attached a modified version of pprint.py where the "if 
sepLines" is more embedded.  It gives the following results:

>>> import pprintmod
>>> class MyPrettyPrinter(pprintmod.PrettyPrinter):
	def format(self, object, context, maxlevels, level):
		if isinstance(object, int):
			return hex(object), True, 
False
		else:
			return 
pprintmod.PrettyPrinter.format(
				self, object, 
context, maxlevels, level)

		
>>> mpp = MyPrettyPrinter()
>>> mpp.pprint(range(10))
[0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9]
>>> mpp.pprint(range(0x10000000,0x10000010))
[0x10000000,
 0x10000001,
 0x10000002,
 0x10000003,
 0x10000004,
 0x10000005,
 0x10000006,
 0x10000007,
 0x10000008,
 0x10000009,
 0x1000000a,
 0x1000000b,
 0x1000000c,
 0x1000000d,
 0x1000000e,
 0x1000000f]
>>> 

Would this be an acceptable change?  Thanks,

----------------------------------------------------------------------

Comment By: Mark Hirota (markhirota)
Date: 2005-11-14 12:04

Message:
Logged In: YES 
user_id=1375527

Sorry, but maybe I could just interject here. Walter's sub-
classing example works fine, except for the fact that if the 
string is short enough, it gets handled slightly differently and 
the format() override gets bypassed.  I worked around this by 
shortening the width, but I've found that this screws with a lot 
of the nice formatting of other structures, and so basically 
creates a new set of problems.  If the "format() bypass" 
limitation can be removed, I think it would allow for more 
reliable extensibility by subclassing PrettyPrinter.

----------------------------------------------------------------------

Comment By: Raymond Hettinger (rhettinger)
Date: 2005-11-14 06:19

Message:
Logged In: YES 
user_id=80475

Fred, any thoughts on the vision for your module?

IMHO,  pprinting containers with ints represented in hex is
not an especially compelling motivation for a total rewrite.
 OTOH, it would be nice if the module were easily extensible..

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2005-11-14 04:54

Message:
Logged In: YES 
user_id=89016

I find pprint.py hard to understand as it is. I've been
staring at the code for several days now and the difference
between PrettyPrinter._format(), PrettyPrinter.format(),
PrettyPrinter._repr() and _safe_repr() is still not entirely
clear to me.

Using a subclass of int only makes sense, if it's your own
data structure that you're outputting. If you get it from
somewhere else, traversing it and replacing every int with
an Int just for output really isn't an option.

----------------------------------------------------------------------

Comment By: Raymond Hettinger (rhettinger)
Date: 2005-11-12 15:02

Message:
Logged In: YES 
user_id=80475

IMO, such a rewrite would expose too many of pprint's
internals and make the module harder to
use/understand/maintain.  Wouldn't it be better to stick
with the usual idiom for controlling the repr() formatting
of specific types by using a class wrapper:

>>> from pprint import pprint
>>> class Int(int):
	def __repr__(self):
		return hex(self)

>>> pprint([Int(x) for x in range(0x10000000,0x10000010)])
[0x10000000,
 0x10000001,
 0x10000002,
 0x10000003,
 0x10000004,
 0x10000005,
 0x10000006,
 0x10000007,
 0x10000008,
 0x10000009,
 0x1000000a,
 0x1000000b,
 0x1000000c,
 0x1000000d,
 0x1000000e,
 0x1000000f]

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2005-11-12 14:29

Message:
Logged In: YES 
user_id=89016

I think it's more of a limitation. I seems to me the main
focus in implementing pprint was speed not extensibility.
The code uses every trick in the book (e.g. turning globals
and builtins into locals, using bound methods etc.). I think
it was never ment to do anything other than what repr()
does, but with better formatting. However IMHO making pprint
extensible would be a worthwile project.

----------------------------------------------------------------------

Comment By: Mark Hirota (markhirota)
Date: 2005-11-11 12:47

Message:
Logged In: YES 
user_id=1375527

Is this bypassing considered a limitation or a bug?  I am, 
however, able to workaround the issue by setting the 
width=1: "mpp = MyPrettyPrinter(1,1)" -- it just means that 
instead of:

>>> mpp.pprint(range(10))
[0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9]

I get instead:

>>> mpp.pprint(range(10))
[0x0,
 0x1,
 0x2,
 0x3,
 0x4,
 0x5,
 0x6,
 0x7,
 0x8,
 0x9]

...which is OK for my uses.  Thanks!

----------------------------------------------------------------------

Comment By: Walter Dörwald (doerwalter)
Date: 2005-11-10 17:56

Message:
Logged In: YES 
user_id=89016

In theory this should be possible by subclassing
pprint.PrettyPrinter and overwritting the format method:

import pprint

class MyPrettyPrinter(pprint.PrettyPrinter):
  def format(self, object, context, maxlevels, level):
    if isinstance(object, int):
      return hex(object), True, False
    else:
      return pprint.PrettyPrinter.format(self, object,
context, maxlevels, level)

mpp = MyPrettyPrinter()
mpp.pprint(range(50))

This doesn't work reliable though: When the string is short
enough, format() seems to be bypassed and repr() is called
directly.


----------------------------------------------------------------------

Comment By: Georg Brandl (birkenfeld)
Date: 2005-11-09 16:45

Message:
Logged In: YES 
user_id=1188172

Moving to Feature Requests.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=355470&aid=1351692&group_id=5470


More information about the Python-bugs-list mailing list