🌐 Mailman 3 January 2000 - Python-list - python.org

tabs do WHAT?
by Warren Postma Feb. 1, 2000

Feb. 1, 2000

Let's see how many languages I have run into that make whitespace and/or tabs important: 1. Makefiles: I once had a problem where a rule had a space where one was not allowed, at the end of a line, causing me to spend three days finding the problem. 2. Multi-line C Macros. Double Blech! The '\' macro-continuation-character must be the last character on the line. Ever been hit with this one? Yuck. At least you only do it once and then you learn. 3. Anything programmed on punch cards. I seem to remember that the first 8 columns were used for line numbers (Fortran?) and the middle bunch were for code, and the last column was special somehow, and whitespace could sometimes make or break your code. Extra Yucky. 4. Python. The two polar opposites (otherwise beautiful syntax, and the tabbed scoping) are so fascinatingly horrible that you have to do a double- It's like the commercial of the beautiful girl who's otherwise perfect smile is marred by yellow nicotine stains. Sorry for the visual image, but hey, I gotta say, it just doesn't fit with the rest of the language, and I don't buy it one bit. A programmer/professor I once had for a software engineering course told us about the "principle of least amazement": Anything weird or hard to explain that you write into your software will later become the thing you regret. As an embedded language, Python needs to be usable beyond the ivory towers where functional programming or object oriented programming or metaclasses are fodder for discussion. Python looked eminently useable by non-programmers until I found out that whitespace was meaningful. Try to explain scope to someone who doesn't write programs for a living. Even BASIC, ugly as it is from a functionality point of view, is miles ahead from the point of view of useability by people who are skilled technical people, yet who are not "programmers". Ergo, Visual Basic's popularity. The language is a dog's breakfast, but at least the basic feature set you use every day is clear and clean enough to keep in your head, and still have room for your other domain-specific knowledge. Python seems one of the few languages that is elegant enough for a computer science type to appreciate, and yet not cluttered up with syntactic garbage as to become an impenetrable mess (TCL, Perl, and the Granddaddy of Unreadable Languages, )... So close, and yet so far away. Has anyone built a version of Python that totally ignores tabs and whitespace, or a filter that you can use on top of Python? Perhaps it could support both 'begin' and 'end', or '{' and '}' for scoping. Heck, at least you could write it in Python. An interpreter within an interpreter. That would probably be very slow, wouldn't it? If Guido reads this, please note my personal feature request for Python 2.0: Make the "whitespace" stuff optional, for people who like weird syntactical features only. I'll love you forever. Thanks. Warren Postma

30 53

Namespace pollution
by Dirk-Ulrich Heise Feb. 1, 2000

Feb. 1, 2000

In the recent whitespace war, some people said "Python is such an expressive language, you need only short functions.", "don't indent too much, that's bad style", even Linus Thorvalds was quoted (who likes 1 TAB = 8 spaces indentation to not lose control)... I've got one problem with all this stuff: I end up with dozens of helper functions that are called exactly one time. They pollute the namespaces of the module or class they're defined in. I hate polluted namespaces. How do others cope with this problem (that does not only apply to Python)? -- Dipl.Inform. Dirk-Ulrich Heise hei(a)adtranzsig.de dheise(a)debitel.net

6 6

list -> dict help
by Bill Scherer Feb. 1, 2000

Feb. 1, 2000

What's the best way for me to do this: say I have a list: >>> l = ['a', 'b', 'c', 1] and I want to turn it into a dictionary that looks like this: >>> print d {'a': {'b': {'c': 1}}} where the length of list l is >= 3. Basically, list[:-2] represents the nested keys, list[-2] is a terminal key, and list[-1] is the value. It seems I need somthing to recur over list[:-2], then add the terminal info but I can't figure it out... Thanks! -- William K. Scherer Sr. Member of Applications Staff Bell Atlantic Mobile

3 5

Re: Stackless Python 1.0 + Continuations 0.6
by Magnus Lie Hetland Feb. 1, 2000

Feb. 1, 2000

Cool! :) Without *really* having read all the documentation, I venture a question... By using S.P. and continuations etc. - would it be possible to stop and pickle a thread? This might be very interesting for Mobile Agent Technology etc. (I.e. a running program can migrate over a network and resume its execution etc...) -- Magnus Lie Hetland mailto:<anything>@hetland.org

5 6

DL_EXPORT (Extending on Windows question)
by Thomas Heller Feb. 1, 2000

Feb. 1, 2000

How do I compile my C-extensions so that the usual DL_EXPORT (void) initmodule (void) is expanded to __declspec (dllexport) void initmodule (void) ? Cannot figure out the right defines... Thomas Heller ION-TOF GmbH

2 1

Win NT Workstation (v4, sp #5) PWS Unable to use Python?
by junkster＠nospam.rochester.rr.com Feb. 1, 2000

Feb. 1, 2000

Folks, I just changed my "Python" development machine, and even though it's configured "identically" (as far as I can tell), PWS doesn't want to allow me to deal with python scripts. I've tried Aaron Watters instructions for installing python for IIS, via REGEDIT....But even after adding the reg keys it doesn't work. I finally uninstalled PWS, reinstalled it with the IIS manager, and added .py extension info via IIS Manager. That seems to work, but (!!) if I have just "c:\progra~1\python\python.exe" the task just "sits" there until it times out. If it's "c:\progra~1\python\python.exe -u %s %s" , the result is always "python.exe: unable to open <filename>". Anyone have any suggestions? - Benjamin ==================================================== (Remove "NoSpam" to Email me) ==================================================== Please feel free to copy any and or all of this sig. A little something for spam bots: root@localhost postmaster@localhost admin@localhost abuse@localhost postmaster(a)127.0.0.1 Chairman William Kennard: bkennard(a)fcc.gov Commissioner Susan Ness: sness(a)fcc.gov Commissioner Harold Furchtgott-Roth: hfurchtg(a)fcc.gov Commissioner Michael Powell: mpowell(a)fcc.gov Commissioner Gloria Tristani: gtristan(a)fcc.gov consumerline(a)ftc.gov fccinfo(a)fcc.gov ssegal(a)fcc.gov

2 1

Date Time problem in the Win32 extensions
by George Thom＠s Feb. 1, 2000

Feb. 1, 2000

Hi everyone, I've managed to put together a rudimentary script that extracts all the summary information from a storage object using the amazing support for COM offered by Python. Part of the code was adapted from an example in the beta chapter of 'Python Programming on Win32' by Mark Hammond & Andy Robinson. I seem to have encountered a problem in PIDSI_EDITTIME(0x0000000a). Consider the following code fragment. import pythoncom from win32com import storagecon # constants related to storage functions. # These come from ObjIdl.h FMTID_UserDefinedProperties = "{F29F85E0-4FF9-1068-AB91-08002B27B3D9}" #------------- Lots of other COM stuff that can be obtained from the online beta chapter ------------------- ps = pss.Open(FMTID_UserDefinedProperties) # pss is a handle to the IPropertySetStorage interface props = PIDSI_TITLE, PIDSI_SUBJECT, PIDSI_AUTHOR, PIDSI_KEYWORDS, PIDSI_COMMENTS, PIDSI_TEMPLATE, PIDSI_LASTAUTHOR, PIDSI_REVNUMBER, PIDSI_EDITTIME, PIDSI_LASTPRINTED, PIDSI_CREATE_DTM, PIDSI_LASTSAVE_DTM, PIDSI_PAGECOUNT, PIDSI_WORDCOUNT, PIDSI_CHARCOUNT, PIDSI_THUMBNAIL, PIDSI_APPNAME, PIDSI_DOC_SECURITY data = ps.ReadMultiple( props ) On retrieval of the PIDSI_EDITTIME property, a PyTime object is obtained. It, however, gives me 1601 as the year and consequently the Format() method returns meaningless information. Although I initially thought this was a Y2K issue, I later noticed that the other PyTime fields (PIDSI_LASTPRINTED for instance) continued to return compliant date/time information. I would really appreciate it if someone could help and let me know if I'm missing something obvious here. regards, george -------------------------------------------------------------------------- George Thomas Resident,Planet Earth,Third Rock from the Sun georgethomas(a)pspl.co.in Flat No. FI-5, Fifth Floor, Himagirinath Co-op. Hsg. Society, 486A Elphinstone Road, Kirkee, Pune - 411003, Maharashtra State, India. -------------------------------------------------------------------------- Faced with the choice between changing one's mind and proving there is no need to do so, almost everybody gets busy on the proof. - John Kenneth Galbraith --------------------------------------------------------------------------

2 1

PIL Problem (Python Imaging Library)
by Jerome ALET Feb. 1, 2000

Feb. 1, 2000

Hi, Does someone know if the standard PIL library can allow me to extract many graphic files (JPEG in my case) from a special undocumented format archive ? the only thing I know is the offset at which begins each jpeg file, but I don't know their sizes, and I don't want to analyze the jpeg format to finally discover the size (in fact I want PIL to do it for me). I've made a very small patch to PIL which works and allows me to do the following (say the jpeg files are all on 256 bytes boundaries): -------------------- file = open(filename, "rb") # the 'b' is only for portability to winXX # ... szbuf = 256 # all included documents seem to be on a 256 bytes boundary count = 0 eof = 0 while not eof : jpegpos = file.tell() read = file.read(szbuf) if len(read) : if isjpeg(read) : # test if the read chunk ressembles a jpeg file count = count + 1 outfilename = ("%02i" % count) + '.jpeg' newpos = file.tell() file.seek(jpegpos, 0) # I know I'm at the beginning of a graphic file # but I don't know the graphic file's size image = Image.open(file) # fortunately the save function loads the image in # memory before saving, so I don't have to know its size image.save(outfilename, quality = 100) file.seek(newpos, 0) print "\t" + outfilename, "\tSize:", image.size else : eof = 1 file.close() -------------------- but I wanted to know if there's a standard way with PIL to do exactly this without any patch (and without knowing the included jpeg filesizes) in order for me to be able to redistribute my program without worrying people about having the patched PIL. I suppose if another way exists then it should be without calling Image.open because this function assumes the graphic file begins at offset 0 which is false in my case. (the same for the save function). Please could you CC any answer to my email address sine I don't read newsgroups very often ? -- Jerome ALET - alet(a)unice.fr - http://cortex.unice.fr/~jerome

2 1

Re: Python Rocks!
by François Pinard Feb. 1, 2000

Feb. 1, 2000

"Tim Peters" <tim_one(a)email.msn.com> writes: > Programs-writing-programs is a twist of mind conspicuously absent in > the Python world. It has probably been debated to death already, and I have a folder full of suggestions about scanners and parsers for Python, waiting for me to read and study, which I gleaned from this very mailing list. Isn't there a need for compiling grammars into fast scanners and parsers, a bit like Flex and Bison do for C people? To be happy, I presume one would have to write small amounts of Python code for each "action". Or maybe just provide the name of a function to call. Yet, the result has to be Python. Or are you suggesting that such tools should directly generate byte-code? -- François Pinard http://www.iro.umontreal.ca/~pinard

5 5

C, C++ to python
by Antoni Aloy Feb. 1, 2000

Feb. 1, 2000

Hi! Anybody knows if exits a tool to convert C or C++ code to Python. I like to make a port for TVision and such a tool would be a great help. Best regards Antoni Aloy

2 1