🌐 Mailman 3 June 2002 - Python-list - python.org

Re: Python daemon
by claird＠starbase.neosoft.com Oct. 18, 2002

Oct. 18, 2002

In article <mailman.1024999953.15765.python-list(a)python.org>, Roman Suzi <rnd(a)onego.ru> wrote: . . . >> 3. restarting of the program if it fails, or the watchdog doesn't trigg= >er > >Run the program from the /etc/inittab - if it terminates,=20 >init will restart it for you. . . . This is one of the great FMMs under Unix--the under-utilization of init(1). From everything I know, Roman's exactly right, that /etc/inittab is the right way to set up a process that you truly want to keep running. However, many, MANY Unix hosts have all kinds of ugly home-brewed hacks to duplicate this functionality. I don't have an explanation, beyond the rather tautologous ob- servation that init(1) simply isn't as widely understood as it deserves to be. -- Cameron Laird <Cameron(a)Lairds.com> Business: http://www.Phaseit.net Personal: http://starbase.neosoft.com/~claird/home.html

7 10

lists and sequences
by Collin Monahan Oct. 15, 2002

Oct. 15, 2002

What is a python list? Is it a linked list or a random access structure in contiguous memory? What type of situations should be avoided when using it? E.g. when does n^2 time happen using it? What sorting algorithm is used by its sort member function? Are the answers to these questions the same for Jython? If there is a good situation to use a Java collection from Jython because of above considerations, how bad are the constant factors which arise from doing so? That is, how preferable is it to stick with the python-language structures when using Jython? I'm not on the list so please copy me. Thank you Collin Monahan __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com

4 4

Removing ^M
by Michael Hall Oct. 9, 2002

Oct. 9, 2002

I am trying remove ^M characters (some kind of newline character) from an HTML file. I've tried all sorts of string.replace and sed possibilities but the things just won't go away. Does anyone have a way of removing such characters? TIA Mick

6 6

Unicode -> String problem
by Jay Parlar Sept. 8, 2002

Sept. 8, 2002

I'm having a problem converting unicode text to string type with str(). The code snippet causing the problem is if type(pageText) == UnicodeType: newText = str(pageText) and the error message I receive is Traceback (most recent call last): File "<interactive input>", line 1, in ? File "D:\MyData\HOME\PWA\Scripts\filter.py", line 60, in parser newText = str(pageText) UnicodeError: ASCII encoding error: ordinal not in range(128) Now, I know there is a lot of precedence for these "...ordinal not in range(128)" questions, but I've looked around, and I haven't found anything that will explicitly do what I want, namely, completely remove any uncovertable unicode characters. I have to be able to parse this text afterwards, using a lot of Python's string functions, so I need 'newText' to be a string, but I'd really prefer not to have the various unicode strings (ie \xa0) showing up. Is there a simple way to convert the unicode text to StringType, removing the resulting unicode strings for unrepresentable characters? Thanks in advance to anyone who's looking at this, Jay P.

3 2

automating Outlook [repost]
by klassa＠ipass.net Aug. 30, 2002

Aug. 30, 2002

[ I managed to let this one slip by without a subject last time. Sorry for the repost. ] If you're interested in my motivation for this, see the footnote[1]. Bottom line: I wanted to be able to edit my drafts, in Microsoft Outlook, with the editor of my choice. In this case, Emacs. The following script does this for me... You tell Outlook to start a reply, and then you hit the "Edit" button that this script creates. It sucks the text out of Outlook, puts it into Emacs (you need to tweak the paths), then puts it back into Outlook when you're finished. The thing I'd like to add, still, is the ability to put an icon into the system tray, rather than have it be a free-floating application with a button. The sample code I found was a bit hard to grok (given that I'm not a Windows programmer, nor more than a novice python programmer). If you know how to do that, and would care to add it in, I'd love to hear from you. :-) The only snag I've found is that you can't run the script without having run makepy.py first, to make the Outlook library available. The full dynamic dispatch mechanism doesn't work, for some reason. That said, here's the script. It's undoubtedly ugly, but the important stuff is there. :-) -- snip -- import win32com.client import os import Tkinter from Tkconstants import * def launch(): # Default to an empty body. body = "" # Get a handle to Outlook. o = win32com.client.Dispatch("Outlook.Application") # Work our way down to the reply (a "MailItem"). insp = o.ActiveInspector() if insp == None: return item = insp.CurrentItem if item == None: return # Grab the body. body = item.Body # Should make this a guaranteed-unique file... fh = open("c:/temp/editor.txt", "w") # Write the body. Had to add a try/except because of ASCII # encoding problems when the reply is in one of Outlook's more # funky formats. try: fh.write(body) except: fh.write("") fh.close() # Launch emacs to edit the file. Should make this configurable. # Note that by default, Emacs seems to come up in Unix mode, and # so the ^M characters are visible. A persistent, bound-to-a-key # Emacs macro takes care of that nicely, however. os.spawnv(os.P_WAIT, "d:/Editors/emacs-20.7/bin/emacs", ["d:/Editors/emacs-20.7/bin/emacs", "c:/temp/editor.txt"]) # Read the result back into memory. fh = open("c:/temp/editor.txt", "r") body = fh.read() fh.close() # Store it as the body of the reply. item.Body = body # Create a single button that, when clicked, takes care of the rest. if __name__=='__main__': tk = Tkinter.Tk() frame = Tkinter.Frame(tk, relief=RIDGE, borderwidth=2, background="white ") frame.pack(fill=BOTH, expand=1) button = Tkinter.Button(frame, text="Edit", command=launch, background="white") button.pack(fill=BOTH, expand=1) tk.mainloop() -- snip -- Comments welcomed. John [1] Every so often, I make it a point to try some other language or tool, so as to widen my perspective a bit. For example, I'm an ardent emacs fan, but I've learned vi so that I'm able to edit anywhere (well, on any Unix machine), and so as to appreciate other ways to edit files. In this same way, I decided to try Microsoft Outlook. I'm an exmh hacker/fan (have been, for years), but thought I'd try Outlook just to see what my business/marketing/managements friends have to deal with. :-) My one biggest gripe is that you can't use the editor of your choice when you edit your drafts. The Outlook editor is okay... I mean, it's like every other Microsoft editor -- that is, reasonable, but lacking features (like the ability to reflow your text). So, I decided to give Outlook the ability to use the editor of my choice. I'm also a perl fan, but discovered that what I wanted to do was a lot harder (or at least seemed to be) in perl. So, I turned to python. Python seems to have a lot of nice Win32 support... This is maybe my fifth or sixth small python script, and I'm liking it (python) more and more.

3 3

Re: automatic "lotjes trekken"
by François Pinard Aug. 26, 2002

Aug. 26, 2002

"Tim Peters" <tim_one(a)email.msn.com> writes: > [...] with a belly full of far too much turkey. Strange. I was imagining you rather slim. Are you? :-) > where p[:] ranges over all permutations of range(N) [...] By the way, I do not remember any standard library helper to enumerate all permutations (or combinations, or arrangements) of a given list (or tuple). Ideally one at a time, of course. Would this be useful? -- François Pinard http://www.iro.umontreal.ca/~pinard

4 5

Re: Passing values from C++ to embedded python.
by Justin Dubs Aug. 19, 2002

Aug. 19, 2002

"Mark Rowe" <bdash(a)gmx.net> wrote in message news:mailman.1022734969.31865.python-list@python.org... > Justin Dubs wrote: > > >Hey everyone, > > > >I have a C++ app in which I have embedded the python interpreter. I have a > >thread which does the simple Py_Initialize(), PyRun_InteractiveLoop(), > >Py_Finalize() business using stdin and stdout. Meanwhile the main thread > >does all the regular functionality of the app in another window. I have > >several classes, written in C++, which I have wrapped with SWIG and compiled > >into a shared library. C++ can instantiate the original classes, obviously, > >as they are written in C++. Python can instantiate the wrapped classes from > >the shared library and that also works perfectly. > > > >So, my problem is thus: > > > >If I have a pointer to an instance of one of these classes in C++, how can I > >create a python variable that points to the same instance? > > > >In other words, in pseudocode: > > > >In C++: > > MyClass *foo = new MyClass(); > > foo->SetAnswer(42); > > // insert unknown code to communicate foo to python > > > >In Python, from my interactive interpreter running on stdin and stdout > >within the C++ app: > > # insert unknown code to receive foo from c++ > > >>> foo.getAnswer(); > > 42 > > > >I hope you can understand what I am trying to accomplish. I need to make > >python aware of an instance of a variable that exists in C++. I tried doing > >this some simple ways from C++ using PyCObject_* and PyInstance_* and > >PyNew_* and other things. But, alas, I am but a beginner with the Python/C > >API and I could really use some help. > > > >Thanks a lot everyone, > > > >Justin Dubs > > > > > > > > > Hello, > > I had a similar problem recently. I'm unsure if the method that I used > to solve it is the simplest way to do so, but so far it has worked for me. > > The SWIG function SWIG_NewPointerObj can be used to return a PyObject * > for your wrapped class. If you are using SWIG shadow classes, you then > have to create a wrapper class instance that shadows the newly created > PyObject *. You can then insert the PyObject * into the appropriate > Python namespace. > > eg (Edited and untested): > > PyObject *ToConsole(Console *c); > > // Insert console object to into Python namespace > void InsertObject(Console *m_console) > { > PyObject * console_s = NULL, * a3d_m = NULL, * a3d_md = NULL, * > console_class = NULL; > PyObject * console = NULL, * main = NULL, * main_d = NULL, * > arg_tuple = NULL; > console_s = ToConsole(m_console); > > // Create Python shadow class from Python Object > // A3D.py is the shadow class file generated by SWIG > a3d_m = PyImport_ImportModule("A3D"); > a3d_md = PyModule_GetDict(a3d_m); > console_class = PyDict_GetItemString(a3d_md, "ConsolePtr"); > arg_tuple = PyTuple_New(1); > PyTuple_SetItem(arg_tuple, 0, console_s) ; > console = PyObject_CallObject(console_class, arg_tuple); > Py_XDECREF(arg_tuple); > arg_tuple = NULL; > > // Insert Python shadow class into global namespace > main = PyImport_AddModule("__main__"); > main_d = PyModule_GetDict(main); > > PyDict_SetItemString(main_d, "console", console); > > Py_XDECREF(console); > Py_XDECREF(a3d_m); > console = NULL; > a3d_m = NULL; > } > > and in your SWIG file, create a small section like the following: > > %{ > extern swig_type_info _swigt__p_Console[]; > PyObject *ToConsole(Console *c) > { > // _swigt__p_Console is a generated structure(?) that represents the > C++ class > extern swig_type_info _swigt__p_Console[]; > return SWIG_NewPointerObj(c, _swigt__p_Console, 0); > } > %} > > The C++ instance that you pass to InsertObject is then available in > Python as console in the __main__ namespace. I hope that this makes > sense, and helps you :D > > Mark > Thanks a lot Mark, I have a new problem now though. :-). our idea makes sense, I wrote the code, I put the stuff in the SWIG interface file, and it compiled perfectly. Woohoo. However, I can't figure out how to link it now. Here's what I have, simplified: classtowrap.cpp # the class i'm trying to wrap classtowrap.h # it's header file classtowrap.i # swig interface file classtowrap_wrap.cxx # generated c++ wrapper class from SWIG classtowrap.py # python shadow class from SWIG classtowrapc.so # the compiled, shared library from the classtowrap.cpp and the classtowrap_wrap.cxx that python can import main.cpp # the main driver program So, the main.cpp is what gets the python interpreter running and needs to send the instance into the python namespace. It, therefore, needs to use that method we defined in our classtowrap.i file, which was put in the classtowrap_wrap.cxx file, which is now in the shared library classtowrapc.so. I can't, of course, just do a -lclasstowrapc because libraries need to start with "lib", so I made a link from libclasstowrapc.so -> classtowrapc.so. Now I did the compilation of main.cpp with a -lclasstowrapc at the end of it. Here's the error I got: ld: libclasstowrapc.dylib is input for the dynamic link editor, is not relocatable by the static link editor again Mac OS X was nice enough to copy my symbolic link, this time with a .dylib extension, as is the default for dynamically linked libraries on this OS. However, being a dynamic library, you can't statically link it. So, how DO you dynamically link a library? I've never dealt with this kind of thing before. Thanks again for your help, Justin Dubs

2 2

Re: tomorrow in yyyymmdd format
by Conrad Aug. 14, 2002

Aug. 14, 2002

Years ago, Nostradamus predicted that on Tue, 18 Jun 2002 20:52:25 -0500, Gerhard Häring would write, saying: > * Sean 'Shaleh' Perry <shalehperry(a)attbi.com> [2002-06-18 18:38 -0700]: >> >> >> >> Sure there is :-) time.mktime() can take a "malformed" time tuple >> >> and do something sensible with it. You can therefore get a time >> >> tuple from gmtime, add one to the day position and call >> >> time.mktime() on the result: >> >> >> >> >>> x = (2002, 2, 29, 0, 31, 42, 2, 170, 0) >> >> >>> time.gmtime(time.mktime(x)) >> >> (2002, 3, 1, 7, 31, 42, 4, 60, 0) >> >> >>> >> >> >>> >> > Pure luck that this works. Don't complain if it stops workin in >> > Python 2.7 and your code breaks ;-) >> > >> > >> actually that is a function of the mktime() function in the C library >> which the python module wraps. Based on my reading of a man page or >> two this behaviour seems guaranteed by POSIX. Now, the assumption made >> is the mktime() function on YOUR system follows POSIX. > > Fine. I've just read the Python documentation which doesn't make this > guarantee. I'll leave it to others to check if this works on win32, > MacOS, as I'm using mxDateTime for this kind of tasks, anyway. > > Gerhard Hmm - here's what I've done - a bit klutzy but it seems to work: from time import * one_day = 60*60*24 # number of seconds in day # make a julian for tomorrow - # note: depending on what you want be careful around midnight! tomorrow = mktime(localtime()) + one_day print 'tomorrow is',localtime(tomorrow) print 'today is',localtime() tomorrow is (2002, 6, 27, 10, 42, 11, 3, 178, 1) today is (2002, 6, 26, 10, 42, 11, 3, 178, 1) Conrad

3 2

Writing to Excel
by chouster＠uclink4.berkeley.edu Aug. 4, 2002

Aug. 4, 2002

Hi I'm using a Python cgi script to write the contents of a table to an Excel sheet. I retrieve all the information fine but it seems like Excel doesnt want to print more than one line for me. It prints 'headerLine' but nothing more. Why??? ... headerLine = "some crap" sys.stdout.write( Html.header( 'application/x-msexcel' ) ) print string.strip(headerLine) allText = "" ... # set allText = "some other crap" ... print allText Im dismayed that this only will print the first print line I specified to the Excel page. The Html module is a proven-to-work module I've obtained so I'm 99.99% sure that's not the problem. I'm using the latest version of python, winNT, and using IE Explorer 5.5. Chris Chris

5 4

Re: stackless python
by Paul Rubin Aug. 4, 2002

Aug. 4, 2002

Christian Tismer <tismer(a)tismer.com> writes: > I'm pretty sure Scheme can. It has raw continuations, so you can > implement everything with them, why not microthreads? > Continuations can do even more, they are able to create new > control structures for your language. And exactly that it the point > which makes them insuitable for Python: They are too powerful. > Python has its own control structures, and we don't need a construct > with the power to build some. This is oversized, and very oversized > construct turns out to be a drawback at some future. > What Python needs is a secure mechanism to switch frame changes > at certain times. This is not continuations, but microthreads with > explicit or implicit switching. The thing I don't fully understand is that several Scheme implementations are both smaller and faster than Python. I've been wondering for a while whether it's time to graft a Python parser onto a Scheme compiler/evaluator. The limitations of "simple generators" also seem kind of artificial--maybe Stackless for 2.2 can support calling a generator from multiple places.

9 17