🌐 Mailman 3 July 2001 - Tutor - python.org

closing files
by fleet＠teachout.org Aug. 8, 2001

Aug. 8, 2001

In the discussion about the non-operating CGI form, I noticed that noone mentioned that the file opened for reading was never closed. I recreated the situation in an interactive session (ie, opened a file for reading, then without closing it, opened it again for writing and wrote to it). Python didn't appear to care that the file opened for reading was not closed before the same file gets opened for writing. When opened files are not closed, what are the dangers? Are files closed by Python under some circumstances and, if so, which circumstances? If I write something like: open("writefile","w").write(open("readfile","r").read()) are both the read and write files closed on execution of the command - or should I spoil the beauty of a one-liner by adding: readfile.close() writefile.close() - fleet -

3 3

Re: [Tutor] python source code
by Michael P. Reilly Aug. 8, 2001

Aug. 8, 2001

support wrote > I am trying to obtain the source code for python ver. 1.5.2 for purposes > of rebuilding, as I need to change what functions are actually called in > C for pythons calls of open, write close, etc... (File I/O).. currently > python doesn't use the win32API funcs for this, which i need to use for > the program im working on to operate on NT, and over networks. IF > anyone knows where i can get the source or has done this or has any > advice for this (such as how to include the win32API in the C file, or > which functions from python are defined where in the source code) I > would be greatly appreciated. Yikes, that seems like a LOT of overkill. Firstly, as a bad thing to do, you could put other function in place of the built-in "open" file (in the __builtin__ module). Changing things in the __builtin__ module is not good because it makes for some dangerous side-effects when something you don't expect happens. The Python core does not use win32 APIs because it is not a win32 application; it is supported on multiple platforms. However there are full extensions for the win32 APIs, including CreateFile and CreatePipe. Before you start cracking the code, you probably want to download and look at those extensions: win32all at <URL: http://www.python/windows>. The module you are looking for in that package is called "win32file". Mark Hammond has also written a good book called "Python Programming on Win32", which describe the win32all package and other things. Good luck, -Arcege P.S.: Python uses C's standard library for files. Win32all can give you better control for windoze, but won't let you get portable results. -- +----------------------------------+-----------------------------------+ | Michael P. Reilly | arcege(a)speakeasy.net |

2 1

RE: [Tutor] using IDLE and changing variables
by alan.gauld＠bt.com Aug. 8, 2001

Aug. 8, 2001

> when you make changes to a module that you are importing, > the changes won't be recognized unless you force a reload Correct, at the interactive prompt. > Let's say at one point you had a variable x defined > and you used it in several places. Then you change > the variable from x to y but forget to change it > everywhere. That'd be a programmer error in every environment I've ever used. Some language compilers will tell you if you remove the original declaration/definition of the variable (x in your case) and then try to use it. Python will warn you in some situations. But it has nothing to do with IDLE per se, it's all to do with Python itself. > I guess what I'm looking for is something like BASIC's "new" > command that flushes out everything of the session and begins > fresh. The interactive prompt doesn't have that feature but working in IDLE the best thing IMHO is simply to create a new file using File|New... The >>> prompt is great for playing with ideas but if its more than a few lines of code I prefer to work in a file and edit/save/run - even if the file is just a throw-away like tmp.py of foo.py or whatever. Alan G

4 4

[Long!] Varying degrees of precision
by kromag＠nsacom.net July 31, 2001

July 31, 2001

Hi all! Sorry I've been a stranger lately, but I have been decending into C. I have noticed that there are some weirdnesses occouring when converting integers to floats in python. I wrote the following module as the beginnings of a "hot rod math" project. The first function, suck() gives the proper cfm flow for a carburetor on an engine of a given size at a given rpm. ------------begin rodney.py----------- '''Suck gives a good idea of the required carburetor flow for an engine of a given volume in cubic centimeters at a given RPM. Note: suck does not round up to nearest cfm, but merely strip everything to the right of the decimal. If you are really, really anal, replace 'int(cfm) with 'cfm' in the lines below and revel in your increased accuracy... then cry as you realize you can't buy a carb for your Buick that flows at the exactly optimal 347.83669982303041 CF/M. Now calm down.''' '''usage: suck(350,5500,'sae') or: suck(2000.00,6000.00,'cc') <---here it is! Hint: 'sae' is the only thing you have to remember. 'cc' could be replaced by 'booger' and it would still work.''' def suck(cc,rpm,increment): if increment == 'sae': cid=cc else: cid=cc/16.387 # This line converts from cc to cid- lazy. cfm=( cid/2 )*( rpm/1728 )*.85 #Assumes 85% volumetric effeciency. #return int( cfm ) return cfm ------------end rodney.py--------------- rodney.py returns: ------------begin rodney output--------- Python 2.0 (#8, Oct 16 2000, 17:27:58) [MSC 32 bit (Intel)] on win32 Type "copyright", "credits" or "license" for more information. IDLE 0.6 -- press F1 for help >>> >>> from rodney import suck >>> suck(2000,6000,'cc') 155.61115518398728 >>> suck(2000.00,6000.00,'cc') 180.10550368517048 >>> ------------end rodney output------------ note the differences in the results! I figured this out when I wrote the following lame C program: ----------lame c program----------------- #include <stdio.h> float rpm; float cc; float cid; float cfm; main() { rpm=6000; cc=2000; cid=cc/16.387; cfm=(cid/2)*(rpm/1728)*.85; printf("%f",cfm); return(0); } ---------end lame c program--------------- which returns 180.105499 Is this a case of weak typing biting me on the backside, or is there a pythonic way to specify a float? Thanks!

2 1

[Long!] Varying degrees of precision
by kromag＠nsacom.net July 31, 2001

July 31, 2001

Hi all! Sorry I've been a stranger lately, but I have been decending into C. I have noticed that there are some weirdnesses occouring when converting integers to floats in python. I wrote the following module as the beginnings of a "hot rod math" project. The first function, suck() gives the proper cfm flow for a carburetor on an engine of a given size at a given rpm. ------------begin rodney.py----------- '''Suck gives a good idea of the required carburetor flow for an engine of a given volume in cubic centimeters at a given RPM. Note: suck does not round up to nearest cfm, but merely strip everything to the right of the decimal. If you are really, really anal, replace 'int(cfm) with 'cfm' in the lines below and revel in your increased accuracy... then cry as you realize you can't buy a carb for your Buick that flows at the exactly optimal 347.83669982303041 CF/M. Now calm down.''' '''usage: suck(350,5500,'sae') or: suck(2000.00,6000.00,'cc') <---here it is! Hint: 'sae' is the only thing you have to remember. 'cc' could be replaced by 'booger' and it would still work.''' def suck(cc,rpm,increment): if increment == 'sae': cid=cc else: cid=cc/16.387 # This line converts from cc to cid- lazy. cfm=( cid/2 )*( rpm/1728 )*.85 #Assumes 85% volumetric effeciency. #return int( cfm ) return cfm ------------end rodney.py--------------- rodney.py returns: ------------begin rodney output--------- Python 2.0 (#8, Oct 16 2000, 17:27:58) [MSC 32 bit (Intel)] on win32 Type "copyright", "credits" or "license" for more information. IDLE 0.6 -- press F1 for help >>> >>> from rodney import suck >>> suck(2000,6000,'cc') 155.61115518398728 >>> suck(2000.00,6000.00,'cc') 180.10550368517048 >>> ------------end rodney output------------ note the differences in the results! I figured this out when I wrote the following lame C program: ----------lame c program----------------- #include <stdio.h> float rpm; float cc; float cid; float cfm; main() { rpm=6000; cc=2000; cid=cc/16.387; cfm=(cid/2)*(rpm/1728)*.85; printf("%f",cfm); return(0); } ---------end lame c program--------------- which returns 180.105499 Is this a case of weak typing biting me on the backside, or is there a pythonic way to specify a float? Thanks!

2 1

Note on style
by Charlie Clark July 31, 2001

July 31, 2001

>Someone posted a link to Guido's notes >on style and convention for me the other week. "InFile" would look like a >class to many a pythoneer. Sorry. I did read it but didn't bookmark it. But >it was over on Sourceforge In order to answer myself. Danny Yoo posted this link: http://python.sourceforge.net/peps/pep-0008.html Well worth reading for all newbies. Coming up with a good naming system (one which is easy to stick to and which others understand) is one of the most * underated* skills in programming in my opinion. Charlie -- Charlie Clark Helmholtzstr. 20 Düsseldorf D- 40215 Tel: +49-211-938-5360 GSM: +49-178-463-6199 http://www.begeistert.org

1 0

Re: [Tutor] Range limited?
by fleet＠teachout.org July 31, 2001

July 31, 2001

Recap: The discussion of the limits of an integer would not have occured (perhaps) if I had checked more than one source! The "docs/ref/integers.html" says: "Plain integer decimal literals must be at most 2147483647 (i.e., the largest positive integer, using 32-bit arithmetic). Plain octal and hexadecimal literals may be as large as 4294967295, but values larger than 2147483647 are converted to a negative value by subtracting 4294967296. There is no limit for long integer literals apart from what can be stored in available memory." Unfortunately, I had checked "Learning Python" only - and that reference (the one I've most used to date!) doesn't mention this limitation (or I can't find it.) Two other sources - "Python Essential Reference" and "Core Python Programming" - both provide the bounds of the python integer. (These are recent purchases - and were lying on the desk unopened at the time I asked the original question.) Also found (while researching the integer question) - both range and xrange operate with "integers" as arguments. My uncomplicated (?simplistic?) mind had seen this as "integer == number == (integer, float, long, complex)." Range and Xrange both convert the numbers to integers and are therefore limited to a range of 2147483647 (or -2147483648). >>> range(long(5)) [0, 1, 2, 3, 4] >>> type(range(long(5))[2]) <type 'int'> - fleet -

1 0

python source code
by support July 31, 2001

July 31, 2001

I am trying to obtain the source code for python ver. 1.5.2 for purposes of rebuilding, as I need to change what functions are actually called in C for pythons calls of open, write close, etc... (File I/O).. currently python doesn't use the win32API funcs for this, which i need to use for the program im working on to operate on NT, and over networks. IF anyone knows where i can get the source or has done this or has any advice for this (such as how to include the win32API in the C file, or which functions from python are defined where in the source code) I would be greatly appreciated. adam

2 2

backquotes and str() are not exactly equivalent
by Danny Yoo July 31, 2001

July 31, 2001

Hiya everyone, I'm doing a small project with Postgresql, and as I was inserting data into the database, I noticed that the backquote operator doesn't do exactly what str() does. Here's an example that shows what I mean: ### >>> myname = 'tutor-curator' >>> str(myname) 'tutor-curator' >>> `myname` "'tutor-curator'" >>> repr(myname) "'tutor-curator'" ### That is, backquotes (``) do a repr() of its input, and not an str() as I had thought before. Just wanted to share my goofup with everyone else... *grin* I'm having quite a bit of fun with the 'pygresql' and 'mailbox' modules, so if anyone has questions about these two modules, feel free to ask questions.

1 0

Range limited?
by fleet＠teachout.org July 30, 2001

July 30, 2001

Just for fun, I imported the smallest of the RSA Challenge numbers into a file, then (looking for factors): from math import sqrt import sys a=long(open("rsa.num","r").read()) b=long(sqrt(a)) for i in range(3,b,2): if 5 % i == 0: pass if a % i == 0: print i, b/i sys.exit() Apparently python (5.2.1) doesn't much care for 84 digit ranges. I couldn't get this to work until I shortened the range number to 8 digits. (I skipped from 10 digits to 8, so maybe 9 digits will also work). Did I mess something up or is 'range' limited? I suspect my math skills are not up to solving this. Obviously, iterating through an 84-digit number in this manner would take some real time - even on a much faster PC than my 733! :) - fleet -

4 6