f



AIX + GCC 2.95.3 - Python (2.1 or 2.2) - building the shared library (.so) for python - HOW TO?!?

Hi all,
(If I use standard -shared linking, during the _import python is crashing.)
As seen in in AIX-NOTES there should be used ld_so_aix. But I suspect 
that it was prepared for standard xlC (CC) compiler. How to make it 
running for GCC compiler?

Now I receive following error:

gcc -Wl,-einitlib.a -Wl,-bE:lib.a.exp -Wl,-bI:./python.exp -Wl,-bhalt:4 
-Wl,-bM:
SRE -Wl,-T512 -Wl,-H512 -lm -o lib.a. -shared -Wl,-bbigtoc 
-L../../../../lib/pow
erpc_aix_4.3_gcc -L/users/tbech/ND/Python/usr/local/lib/python2.1/config 
python.
exp y.tab.o lex.yy.o idlerr.o idlutil.o idltype.o idlrepoId.o idlscope.o 
idlexpr
..o idlast.o idlvalidate.o idldump.o idlconfig.o idlfixed.o idlpython.o
ld: 0711-418 ERROR: Import or export file lib.a.exp at line 10:
         A symbol name may only be followed by an export attribute
         or an address. The line is being ignored.

When I remove this line 10 from lib.a.exp it is telling that several Py* 
functions are not exported and that entry point initlib.a not found.

Additional questions:
1. Why is such nonuderstandable-ultimate-tricky solution prepared for AIX?
2. Does it work only for xlC (CC) or also for GCC? If, how to do it?
3. What is the entry point function? How I can find it in the sources?
4. Seems that makexp_aix wrongly makes exports :(

Thanks in advance,
	Tom

0
hab (4)
7/4/2003 12:04:57 PM
comp.lang.python 77058 articles. 6 followers. Post Follow

2 Replies
1659 Views

Similar Articles

[PageSpeed] 21

hab <hab@polbox.com> writes:

> (If I use standard -shared linking, during the _import python is crashing.)
> As seen in in AIX-NOTES there should be used ld_so_aix. But I suspect
> that it was prepared for standard xlC (CC) compiler. How to make it
> running for GCC compiler?

Nobody knows anything about AIX in the Python world. You have to make
it work yourself.

> 1. Why is such nonuderstandable-ultimate-tricky solution prepared for AIX?

There are two possible reasons:
1. AIX is such a strange system that you need to play dirty tricks to
   make it load modules dynamically.
2. Whoever ported shared loading to AIX didn't know anything about the
   system, and tried random things until he got a working solution.

I don't know which one it is, but I'm leaning towards 1)

> 2. Does it work only for xlC (CC) or also for GCC? If, how to do it?

Nobody knows.

> 3. What is the entry point function? How I can find it in the sources?

You mean, of python itself? It's main(), and it is defined in
Modules/python.c.

Regards,
Martin
0
martin
7/5/2003 8:04:21 AM
In article <m3y8zdz9fe.fsf@mira.informatik.hu-berlin.de>,
 martin@v.loewis.de (Martin v. Lowis) wrote:

> hab <hab@polbox.com> writes:
> 
> > (If I use standard -shared linking, during the _import python is crashing.)
> > As seen in in AIX-NOTES there should be used ld_so_aix. But I suspect
> > that it was prepared for standard xlC (CC) compiler. How to make it
> > running for GCC compiler?
> 
> Nobody knows anything about AIX in the Python world. You have to make
> it work yourself.
> 
> > 1. Why is such nonuderstandable-ultimate-tricky solution prepared for AIX?
> 
> There are two possible reasons:
> 1. AIX is such a strange system that you need to play dirty tricks to
>    make it load modules dynamically.
> 2. Whoever ported shared loading to AIX didn't know anything about the
>    system, and tried random things until he got a working solution.
> 
> I don't know which one it is, but I'm leaning towards 1)

AIX is a classic example of underemployed engineers working in a
vacuum.  It's strange, sometimes arguably better but rarely worth
the trouble.  Linux strikes me a bit the same way, actually, and
IBM's interest in Linux is no surprise.  I doubt there are any really
dirty tricks, but there is a point at which things like this are
going to depend on the compiler software.  ld_so_aix works fine with
xlc, in my experience, and there's no reason to look for malfeasance
on either side to explain why it stops there.  If you want a normal
UNIX platform with an open architecture, try one of the BSDs, or Linux
if you're not that picky about the normal UNIX part.  If you're obliged
to work on AIX, then whoever's paying for AIX should happy to pay for
the compiler too.

   Donn Cave, donn@u.washington.edu
0
Donn
7/7/2003 4:35:40 PM
Reply:

Similar Artilces:

how can i write this statement labels=[1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;1;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;2;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;3;]; in compress fo
How to write it in compact form to avoid repetition "shah " <shahkhn3@gmail.com> wrote in message <lodugh$8hr$1@newscl01ah.mathworks.com>... > How to write it in compact form to avoid repetition You'll find in this thread different solutions for similar question http://www.mathworks.fr/matlabcentral/newsreader/view_thread/335814 Bruno ...

Upgrade from VO 1 - 2 - 2..1 -2.2 - 2.3 -2.4 - 2.5 - 2.6 - 2.7
About upgrade 2.5 - 2.7 at $ 384.00 For most products I'm using - the updates (2.0 - 2.1 - 2.7) are free... 2.0 to 3.0 might be worth + 10-30% of the original price? VO is + full price again & again - Full program price for every minor upgrade/bug-fix? I dropped out of the money/bugs [VO] at 2.5 after starting at 1.0 more than 10 years ago. Certainly whoever is making VO have to make a living; earning moneys: - that is OK! A 10 years old VO site; - 10 years later, how much have changed? http://www.yi.com/prany/cavo/cavofront.htm Even MS is not ...

compiling gcc 3.2.3 on aix 4.3.2 using gcc 3.2.1
All, I am trying to compile gcc-3.2.3 on aix-4.3.2 using a prebuilt gcc-3.2.1. I am using GNU make, native as,ar,ld. There are the errors I get. Can anyone help? thx balaji $ gmake bootstrap gmake[1]: Entering directory `/mnt/gcc-3.2.3/aix/libiberty' gmake[2]: Entering directory `/mnt/gcc-3.2.3/aix/libiberty/testsuite' gmake[2]: Nothing to be done for `all'. gmake[2]: Leaving directory `/mnt/gcc-3.2.3/aix/libiberty/testsuite' gmake[1]: Leaving directory `/mnt/gcc-3.2.3/aix/libiberty' gmake[1]: Entering directory `/mnt/gcc-3.2.3/aix/zlib' : gmake ; exec true "AR_...

[1 1 1 1 1 ;2 2 2 2 2 ;3 3 3 3 3 3;....;n n n n n]
Hi, Anybody knows how to create this matrix but without using any loops? a=[1 1 1 1 1 ;2 2 2 2 2 ;3 3 3 3 3 ;.......;n n n n n ] Thank you. Hana. Hana wrote: > Hi, > Anybody knows how to create this matrix but without using any loops? > a=[1 1 1 1 1 ;2 2 2 2 2 ;3 3 3 3 3 ;.......;n n n n n ] HELP REPMAT - Randy Hana wrote: > > > Hi, > Anybody knows how to create this matrix but without using any > loops? > a=[1 1 1 1 1 ;2 2 2 2 2 ;3 3 3 3 3 ;.......;n n n n n ] > > Thank you. > Hana. Hope this isn't homework. >> repmat([1:n]'...

Printing 2^1, 2^2, 2^3, 2^4 and 2^5
Hi, When I use a while loop, the output is: 2 4 8 16 32, as desired: <?php $c=1; while($c<=5) { $d = newpow(2, $c); print "$d "; $c++; } function newpow($base, $power) { // statements... } ?> The But when I use a for loop, the output is: 64: <?php for($c=1; $c<=5; $c++); { $d = newpow(2, $c); print "$d "; } function newpow($base, $power) { // statements... } ?> Why is this?? Many thanks, Michael mejpark wrote: > When I use a while loop, the output is: 2 4 8 16 32, as desired: [...] > The But when I use a for loop, the output is: 64...

how to "(1-3*x^2)/(1-3*x+x^2+2*x^3) => 1/(1-2*x) + x/(1-x-x^2)"?
hi normal(1/(1-2*x) + x/ (1-x-x^2),expanded); 1/(1-2*x) + x/(1-x-x^2) => (1-3*x^2)/(1-3*x+x^2+2*x^3) how to do the reverse? dillogimp@gmail.com writes: > normal(1/(1-2*x) + x/ (1-x-x^2),expanded); > 1/(1-2*x) + x/(1-x-x^2) => (1-3*x^2)/(1-3*x+x^2+2*x^3) > how to do the reverse? convert(%,parfrac,x); -- Joe Riel ...

Aix 5.2 and gcc 3.3
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Any bleading edgers out there get this to work? Building with BerkeleyDb 4.1.25 and TLS Had some problems with propolice (stack smashing protection) causing gcc to fail. Removed that option and now it compiles but won't link... yeah I know AIX what's new? The configure script for newdb had some of the same errors while testing for mutex's but eventually found configure:17733: result: UI/threads/library Fails with ld: 0711-317 ERROR: Undefined symbol: .mutex_init ld: 0711-317 ERROR: Undefined symbol: .cond_init ld: 0711-317 ERRO...

Tornado 2.2.1/GCC 2.96 #2
vxWorks 5.5.1/Tornado 2.2.1 is shipped with GCC 2.96. GCC 2.96 is _very_ old. The question. Short of purchasing/ugrading to the latest version of Tornado ( which I believe is shipped with GCC 3.3.4 ), has anyone tried upgrading the GCC binarys. If so what was the prognosis? forums_mp@hotmail.com wrote: > vxWorks 5.5.1/Tornado 2.2.1 is shipped with GCC 2.96. GCC 2.96 is > _very_ old. The question. Short of purchasing/ugrading to the latest > version of Tornado ( which I believe is shipped with GCC 3.3.4 ), has > anyone tried upgrading the GCC binarys. If so what was the ...

2 * 2 * 2 * 2 or pow(2, 4)?
Hi , I hope this isn't too trivial to ask, but I'm evaluating a polynomial that involve lots of terms like x^2y^2z^3, and although I can use the power function pow(a, b), I wonder if just writing x*x*y*y*z*z*z for example wouldn't be more efficient than pow(x, 2)*pow(y, 2)*pow(z,3). Or is it all the same? Thanks, - Olumide Olumide wrote: > I hope this isn't too trivial to ask, but I'm evaluating a polynomial > that involve lots of terms like x^2y^2z^3, and although I can use the > power function pow(a, b), I wonder if just writing x*x*y*y*z*z*z for > exam...

linspace(2,2,2) returns [2 2] ??
May be the help for linspace should change? It says: "LINSPACE(X1, X2, N) generates N points between X1 and X2." Notice, it says BETWEEN. But when I ask for 2 points, between 2 and 2: ---------------- EDU>> linspace(2,2,2) ans = 2 2 --------------- May be mathematically this is ok? i.e. there are infinite points between the point and itself? (on the real line), but from programming point of view, it is confusing. no? --Nasser Maybe it should say "generates N points between X1 and X2, inclusive." What did you expect to...

2**2**2**2**2 wrong? Bug?
In Python 2.5 on intel, the statement 2**2**2**2**2 evaluates to >>> 2**2**2**2**2 20035299304068464649790723515602557504478254755697514192650169737108940595563114 53089506130880933348101038234342907263181822949382118812668869506364761547029165 04187191635158796634721944293092798208430910485599057015931895963952486337236720 30029169695921561087649488892540908059114570376752085002066715637023661263597471 44807111774815880914135742720967190151836282560618091458852699826141425030123391 10827360384376787644904320596037912449090570756031403507616256247603186379312648 4703743782954975613770...

Does 4/2/2= (4/2)/2 =1 OR = 4/(2/2)=4
I don't have matlab and I am trying to translate some simple matlab code. Does 4/2/2 = (4/2)/2 = 1 or 4/2/2 = 4/(2/2) = 4 On 9/11/2012 9:43 AM, sarah.englander@gmail.com wrote: > I don't have matlab and I am trying to translate some simple matlab code. > > Does 4/2/2 = (4/2)/2 = 1 yes. At a given precedence level (and obviously two of the same operators are the same level) evaluation proceeds from LtoR. ....snip... <http://www.mathworks.com/help/techdoc/matlab_prog/f0-40063.html#f0-38155> -- dpb <none@non.net> wrote in message <...

#1 #n+1 #2n+1 #2 #n+2 #2n+2 #3 #n+3 #2n+3 #n #2n #3n #1 #2 #3 #n #n+1 #n+2 #n+3 #2n #2n+1 100 Gbps aggregate stream of 64/66b words virtual lane markers virtual lane 1 virtual lane 2 virtual lan
#1 #n+1 #2n+1 #2 #n+2 #2n+2 #3 #n+3 #2n+3 #n #2n #3n #1 #2 #3 #n #n+1 #n+2 #n+3 #2n #2n+1 100 Gbps aggregate stream of 64/66b words virtual lane markers virtual lane 1 virtual lane 2 virtual lane 3 virtual lane n Simple 66-bit word level round robin distribution Transmit PM Transmit PM Receive PM Receive PM Transmit PC Receive PC 1 0.0 1.0 0.1 0.2 0.3 1.2 1.2 1.1 0.0 1.0 0.1 1.1 2.0 3.0 4.0 6.0 7.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 17.0 18.0 19.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 15.0 16.0 17.0 1...

2.4.2 -> 2.5.2 / Mingw/GCC 3.2 / W2K
hi, confirming my app-code to 2.5.2, two problems are left: wxSlider wouldn't show the selected value field anymore: wxSlider( m_p_editmask, 196, 0, 0, 120, wxPoint(20, 240), wxSize(300, 24), wxSL_HORIZONTAL | wxSL_AUTOTICKS | wxSL_LABELS, wxDefaultValidator, "wait" ); as it did before with 2.4.2 Got a lot of Can't verify memory struct - all bets are off! (common/memory.cpp[424]) messages, removed most by taking away the deletes for the class objects created by new, but my (pro)file-handling is managed by: WX_DE...

LaTex equation numbering like 2.1, 2.2, also 2.0.1, 2.0,2, 2.1.1, etc
David Roach wrote: > Is there a simple command to have latex number > the equations by the section numbers automatically. > > David From the LaTeX Companion (what a wonderful book!) %%%%%%%%%%%%%%%%%%%%%% % number equations within sections % In preamble \makeatletter \@addtoreset{equation}{section} \makeatother \renewcommand{\theequation}{\thesection.\arabic{equation}} %%%%%%%%%%%%%%%%%%%%%% Numbers within sections...I use this all the time in articles. Even works with ******************* ******************* Brilliant. And then I got numbering within subsections by repe...

OSK 2.0, 2.2 and 2.3
Does anyone here have any of the versions named in the subject? I am unravelling 68K Basic I-Code, and I need to compare 2.0, 2.2 and 2.3 modules with 2.1 and 2.4 modules. There are differences between 2.1 and 2.4, and I want to know how many other differences there are between versions. Thanks in advance. Wayne Campbell ...

[wx 2.4.2, GTK+ 1.2, gcc 2.95] wxTextCtrl & Background Colour
I'm having some trouble with setting the background colour of a wxTextCtrl. The issue only occurs when I set the wxTE_MULTILINE style. Basically when the dialog containing the control is shown the background colour of the wxTextCtrl is always white regardless of what I set it to in the dialog's contructor. The 'SetBackgroundColour()' method of the wxTextCtrl returns false, however, 'GetBackgroundColour()' does return the correct colour. Even so the control is still rendered with a white background. I can change the background colour programmatically, after it ...

problems with in 2.3.2 developed modules used with 2.2.2
Sorry for the long header of this mail ;-) Has anyone experienced modules developed in 2.3.2 (using IDLE and = PythonWin as IDE) don't work under 2.2.2.... I mean basic stuff like it = can not see classes from your .py file, so an import fails? It almost likes that in my case 2.2.2 doesn't like the file format = generated by IDLE or PythonWin in 2.3.2. Does someone know if this is a known problem? Vincent Try deleting the .pyc version of the file... maybe 2.2.2 doesn't realize that the magic number is wrong and that it should recompile from the .py source. Kevin. "Raa...

Problem compiling Octave 2.1.50 on Solaris 2.8 con gcc 3.4.2 #2
Dear everybody I have a problem compiling octave on Solaris 8 I have got this compiling error: DiagArray2.h: In constructor `DiagArray2<T>::DiagArray2(T*, int, int)': DiagArray2.h:120: error: there are no arguments to `set_max_indices' that depend on a template parameter, so a declaration of `set_max_indices' must be available DiagArray2.h: In constructor `DiagArray2<T>::DiagArray2()': DiagArray2.h:129: error: there are no arguments to `set_max_indices' that depend on a template parameter, so a declaration of `set_max_indices' must be available DiagArray...

lilo-22.5.8, gcc-3.2.1, glibc-2.2.1
have anyone ever encountered the following error when compiling lilo-22.5.8? cc -c -O2 -Wall -g -DLILO=0x16c2d41b `( if [ -r $ROOT/etc/lilo.defines ]; then cat $ROOT/etc/lilo.defines; else echo -DBDATA -DDSECS=3 -DEVMS -DIGNORECASE -DLVM -DONE_SHOT -DPASS160 -DREISERFS -DREWRITE_TABLE -DSOLO_CHAIN -DVERSION -DVIRTUAL; fi ) | sed 's/-D/-DLCF_/g'` probe.c probe.c:250:1: directives may not be used inside a macro argument probe.c:249:15: unterminated argument list invoking macro "printf" probe.c: In function `notice': probe.c:256: parse error before string constant make: ***...

Why does '[1,2,3] EQ [2]' yield zero but '[1,2,3] EQ 2' yield [0,1,0] ?
Hi all, The post title says it all. I thought "everything" in IDL was an array, but using [2] and 2 with relational operators is not equivalent... I searched the doc and newsgroup for quite a while, but obviously I did not come up with search terms that are specific enough. It took me a while to track this down in some code I was writing. My bug is now fixed *but* I don't understand why exactly. And we all know voodoo and IDL don't mix :-) can somebody clarify please? Many thanks, Gianguido PS: I have a feeling this issue came up not too long ago, just couldn't find...

Freetype 2.2.1 / FontConfig 2.3.2 for OpenVMS 8.2 Alpha
This is Freetype 2.2.1 and FontConfig 2.3.2 packaged for OpenVMS Alpha. These are used in the Cairo Graphics package which is a component of GTK+. Freetype 2.2.1 is current with the UNIX release, and contains a work around for a bug in some versions of the CRTL that will cause an access violation when attempting to read a 0 length file using the mmap functions. This version was built using the UNIX build configure/make procedures. The Freetype 2.2.1 distribution has 3 bugs: apinames.c: Uses non-standard "wt" in fopen(). I worked around that with a /f...

HoldForm[1*2*3] should give 1 x 2 x 3 not 2 x 3
HoldForm[] is loosing the 1* when it apparently should not: In[1]:= HoldForm[1*2*3] Out[1]= 2 x 3 In[2]:= HoldForm[1*1*1] Out[2]= 1 x 1 x 1 In[3]:= HoldForm[3*2*1] Out[3]= 3 x 2 In[4]:= HoldForm[2*2*2] Out[4]= 2 x 2 x 2 In[5]:= HoldForm[2*1*3] Out[5]= 2 x 3 In[6]:= HoldForm[1*2] Out[6]= 1 x 2 In[7]:= HoldForm[1*2*1] Out[7]= 1 x 2 x 1 Q.E.D. Indeed, the same happens with Hold and HoldComplete. I'd say this is a bug. Cheers -- Sjoerd On Feb 25, 11:07 am, "Q.E.D." <a...@netzero.net> wrote: > HoldForm[] is loosing the 1* when it...

Changing (1,2,3,4,5...) to (1,1,2,2,3,3,4,4,5,5...)
Hi All, I have a column vector of numbers that I would like to extrapolate by "doubling up": e.g. turn 2 4 6 8 into: 2 2 4 4 6 6 8 8 any ideas on an easy vectorised way to do this? TIA, Rob v=repmat([2; 4; 6; 8].',2,1); v=v(:) "Robert Barrie" <askme@ifuwant.com> wrote in message <hlt7kp$rut$1@fred.mathworks.com>... > Hi All, > > I have a column vector of numbers that I would like to extrapolate by "doubling up": > > e.g. turn > 2 > 4 > 6 > 8 > > into: > 2 > 2 > 4 > 4 > 6 > 6 > 8 >...

Integral: Exp[-(x-m)^2/(2 s^2)] x^3 (1+x^2)^-1
Hi, any suggestion to make the integral of: Exp[-(x-m)^2/(2 s^2)] x (1+x^2)^-1 Exp[-(x-m)^2/(2 s^2)] x^2 (1+x^2)^-1 Exp[-(x-m)^2/(2 s^2)] x^3 (1+x^2)^-1 between -inf and +inf (or indefinite)? Look like it is not possible, but it is too long time I do not make integrals with more advanced techinques (as going to the complex plane)... so if you have suggestions (wonderful a solution :) ).... THANKS Ale ...

Web resources about - AIX + GCC 2.95.3 - Python (2.1 or 2.2) - building the shared library (.so) for python - HOW TO?!? - comp.lang.python

Berlaymont building - Wikipedia, the free encyclopedia
...  / 50.84361; 4.38278 Coordinates : 50°50′37″N 4°22′58″E  /  50.84361°N 4.38278°E  / 50.84361; 4.38278 The Berlaymont is an office building ...


Now you can live in a dorm-like apartment building where neighbors use a work app called Slack to communicate ...
Slack is a hot San Francisco startup that makes a chat app for the workplace. Now, it's coming to a dorm-like apartment building in the trendy ...

New Study Finds Global Green Building Expected To Double By 2018
A new study has concluded that green building around the globe is expected to double by 2018. According to the results of the study, the percentage ...

Housing starts, building permits fall
U.S. housing starts unexpectedly fell in January likely as bad weather disrupted building projects.

Aston Martin is building an electric car with Faraday Future's financier
The British automaker and LeEco will eventually work alongside Faraday Future, although details are scant at the moment.

Pope Francis on Trump: Building walls 'is not Christian' - CNNPolitics.com
Pope Francis said GOP frontrunner Donald Trump "is not Christian" if he pledges to build a wall between the United States and Mexico.

Pope Francis: Trump building wall is 'not Christian' - CNN Video
Pope Francis said that GOP frontrunner Donald Trump "is not Christian" if he calls for the deportation of undocumented immigrants and pledges ...

South African Students Loot University Buildings, Torch Portraits Of White People
South African Students Loot University Buildings, Torch Portraits Of White People

Pope Francis: Wall-building Donald Trump 'is not Christian'
x Embedded Content Welcome to Thursday. Today we have the Actual Pope Himself weighing in on Donald Trump. Po...

Resources last updated: 2/19/2016 10:51:25 PM