quantlib-dev — Discussion of QuantLib development

Re: [Quantlib-dev] QuantLib-Python patch

[Luigi B. <lui...@fa...>]

On 2004.04.01 02:52, Mark Treiber wrote:
> I've attached a patch for QuantLib-Python to compile on darwin (the  
> BSD core to Mac OS X).

Applied, thanks.

	Luigi

[Quantlib-dev] QuantLib-Python patch

[Mark T. <mrt...@en...>]

I've attached a patch for QuantLib-Python to compile on darwin (the BSD 
core to Mac OS X).  Basically darwin distinguishes between shared 
libraries and loadable modules so the patch detects if darwin is the 
build environment and if that is the case is uses different compiler 
switches.  That's it.  Unfortunately I wasn't able to create this 
before 0.3.5 but compiling on OSX is the only think affected so the 
fink quantlib-python package uses the attached patch on the 0.3.5 tgz.

Mark.

Re: [Quantlib-dev] 0.3.5 preview

[Mark T. <mrt...@en...>]

On Mar 31, 2004, at 7:45 AM, Luigi Ballabio wrote:

> On 2004.03.31 14:30, Dirk Eddelbuettel wrote:
>> On Wed, Mar 31, 2004 at 09:11:28AM +0200, Luigi Ballabio wrote:
>> > How about:
>> >
>> > Debian users can find the "quantlib" package in the "unstable"
>> > distribution.
>> >
>> > to be changed to "testing" when the packages enter Sarge?
>> Better. But how about something like this so that people don't trip
>> over the
>> previous release in testing:
>>     Debian users can find version 0.3.5 of the different quantlib
>> packages in the "unstable" (aka "sid") distribution. The previous 
>> release, 0.3.4, is still in the "testing" (aka "sarge") distribution, 
>> but should be replace by 0.3.5 in a few days.
>
> Good. I'm changing the page in a few minutes.

The quantlib package is now available from fink.  I was actually 
thinking it my head that mentioning the debian and fink packages 
separate from the files is the way to go but I see that's what you 
thought as well.  Could I suggest changing the text for fink to instead 
read:

"For Mac OS X users, Fink packages are available in the 
"current-10.3/unstable" tree.  For a list of available QuantLib related 
files see the <a href 
=http://fink.sourceforge.net/pdb/search.php?s=quantlib>Package 
Database</a>.  To enable the "unstable" tree, please read the <a 
href=http://fink.sourceforge.net/faq/usage-fink.php#unstable>Fink 
faq</a>.  Be sure to run 'fink selfupdate' to obtain the latest package 
definitions, and then run 'fink install quantlib' to install the main 
library."

Thanks..

Mark.

[Quantlib-dev] QuantLib 0.3.5 released

[Ferdinando A. <na...@qu...>]

QuantLib [1] is a cross-platform, free/open-source quantitative finance C++ 
library for modeling, pricing, trading, and risk management in real-life. 
Version 0.3.5 has been released: see [2] for an overview of the library and 
[3] for a summary of the changes since version 0.3.4.

QuantLib is distributed in a number of formats suitable for most operating 
systems. Debian, Fink, and RPM packages are also available.

Furthermore, Python, Ruby, Guile, and MzScheme bindings are available for 
QuantLib 0.3.5 as well as an Excel add-in.

Feedback is welcome.

Ferdinando Ametrano

[1] http://quantlib.org
[2] http://quantlib.org/html/overview.html
[3] http://sf.net/project/shownotes.php?group_id=12740&release_id=223212
  

Re: [Quantlib-dev] 0.3.5 preview

[Luigi B. <lui...@fa...>]

On 2004.03.31 14:30, Dirk Eddelbuettel wrote:
> On Wed, Mar 31, 2004 at 09:11:28AM +0200, Luigi Ballabio wrote:
> > How about:
> >
> > Debian users can find the "quantlib" package in the "unstable"
> > distribution.
> >
> > to be changed to "testing" when the packages enter Sarge?
> 
> Better. But how about something like this so that people don't trip
> over the
> previous release in testing:
> 
>     Debian users can find version 0.3.5 of the different quantlib
> packages in the "unstable" (aka "sid") distribution. The previous  
> release, 0.3.4, is still in the "testing" (aka "sarge") distribution,  
> but should be replace by 0.3.5 in a few days.

Good. I'm changing the page in a few minutes.

Thanks,
	Luigi

Re: [Quantlib-dev] 0.3.5 preview

[Dirk E. <ed...@de...>]

On Wed, Mar 31, 2004 at 09:11:28AM +0200, Luigi Ballabio wrote:
> On 2004.03.30 19:18, Dirk Eddelbuettel wrote:
> >The Debian text, as it currently stand on the page, is plain WRONG as
> >it refers to the testing distribution containing 0.3.4 -- as opposed  
> >to the unstable distribution containing 0.3.5.
> 
> Dirk,
> 	I removed the download link as it's easier to go through
> apt-get (or dpkg, or synaptic...)
> How about:
> 
> Debian users can find the "quantlib" package in the "unstable"  
> distribution.
> 
> to be changed to "testing" when the packages enter Sarge?

Better. But how about something like this so that people don't trip over the
previous release in testing:

    Debian users can find version 0.3.5 of the different quantlib packages
    in the "unstable" (aka "sid") distribution. The previous release, 0.3.4,
    is still in the "testing" (aka "sarge") distribution, but should be 
    replace by 0.3.5 in a few days.

Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

Re: [Quantlib-dev] 0.3.5 preview

[Luigi B. <lui...@fa...>]

On 2004.03.30 19:47, Dirk Eddelbuettel wrote:
> That said, I think my point still applies. Either we ship 'source
> only' (with a possible exception for win, and maybe os x), or we ship  
> source as well as prebuilt. As you seem to have opted for choice two,  
> I personally would find it natural to give Debian the same shelf  
> space as you allocated to RH/FC.

Dirk,
	from the download page, I gather we're actually going for  
choice one: the quantlib rpm is a packaging of the sources (don't ask  
me what use that is--I don't know)

As for shelf space, we're shipping rpms only because RH users (poor  
souls) don't have a central package repository. Personally, as a Debian  
user, I'd rather use apt-get than downloading and installing a .deb  
manually---especially in the future, as dependencies on the Boost  
packages will arise. The same applies to Mac OS X: as soon as the  
packages are ready, an entry will appear which says "Fink packages are  
available in the 10.3/unstable tree."

Anyway: I modified the download page. May you check it out?

Later,
	Luigi

Re: [Quantlib-dev] 0.3.5 preview

[Luigi B. <lui...@fa...>]

On 2004.03.30 19:18, Dirk Eddelbuettel wrote:
> The Debian text, as it currently stand on the page, is plain WRONG as
> it refers to the testing distribution containing 0.3.4 -- as opposed  
> to the unstable distribution containing 0.3.5.

Dirk,
	I removed the download link as it's easier to go through
apt-get (or dpkg, or synaptic...)
How about:

Debian users can find the "quantlib" package in the "unstable"  
distribution.

to be changed to "testing" when the packages enter Sarge?

Later,
	Luigi

Re: [Quantlib-dev] 0.3.5 preview

[Mark T. <mrt...@en...>]

>
> Mark: anything I should write and/or link to for the fink package?
>
I submitted my updated .info to the fink tracker so the package should 
be available a couple of days.  Once its in a good text to write would 
be along the lines of:

"Fink packages are available in the 10.3/unstable tree.  Just run 'fink 
install quantlib' to install.  To enable the unstable tree, please read 
the <a 
href=http://fink.sourceforge.net/faq/usage-fink.php#unstable>Fink 
faq</a>."

I'll let you know once the main library is available.  Python and Ruby 
packages will follow but I'm diagnosing some build issues with these.  
I'll keep you posted.

Mark.

Re: [Quantlib-dev] 0.3.5 preview

[Dirk E. <ed...@de...>]

On Tue, Mar 30, 2004 at 07:33:37PM +0200, Ferdinando Ametrano wrote:
> Hi Dirk
> 
> >If we list / ship the binary rpms, should we ship debs too?

s/binaries/pre-built packages/

I didn't mean binaries in the sense of /usr/bin, I meant .rpm 'binary
packages' as opposed to source tarballs. Sorry for mixing lingo. that really
was not clear.

That said, I think my point still applies. Either we ship 'source only'
(with a possible exception for win, and maybe os x), or we ship source as
well as prebuilt. As you seem to have opted for choice two, I personally
would find it natural to give Debian the same shelf space as you allocated
to RH/FC.  

Dirk


> I wouldn't ship QuantLib binaries, QuantLib developers are supposed to be 
> able to build QuantLib  (and hopefully able to build and run the associated 
> test-suite)
> 
> I would ship QuantLib-Python, QuantLib-Ruby, etc binaries, since users of 
> those packages are not required to build them.
> 
> This is what should be available on quantlib.org/nextrelease.html
> 
> Is it OK?
> 
> ciao -- Nando 
> 
> 

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

Re: [Quantlib-dev] 0.3.5 preview

[Ferdinando A. <na...@am...>]

Hi Dirk

>The Debian text, as it currently stand on the page, is plain WRONG as it
>refers to the testing distribution containing 0.3.4 -- as opposed to the
>unstable distribution containing 0.3.5.

this afternoon Luigi committed a change to the nextrelease page, I don't 
know if it was in accord with you.
Anyway if you write those few lines about Debian I would use them to 
replace the current description before announcing 0.3.5 tomorrow.

Thank you for your help

ciao -- Nando 

Re: [Quantlib-dev] barrier options greeks lead to seg.fault.

[Ferdinando A. <na...@am...>]

Hi Dirk

>Nando, Luigi: Any word on my problem with the barrier option?  I had the
>same problem last night with plain American options as soon as I tried any
>of the Greeks. I must be doing something wrong, but still, it shouldn't
>seg.fault, should it?

I hope to be able to take a look at it by the end of the week. Sorry Dirk, 
but I'm running short on free-time


ciao -- Nando 

Re: [Quantlib-dev] 0.3.5 preview

[Ferdinando A. <na...@am...>]

Hi Dirk

>If we list / ship the binary rpms, should we ship debs too?

I wouldn't ship QuantLib binaries, QuantLib developers are supposed to be 
able to build QuantLib  (and hopefully able to build and run the associated 
test-suite)

I would ship QuantLib-Python, QuantLib-Ruby, etc binaries, since users of 
those packages are not required to build them.

This is what should be available on quantlib.org/nextrelease.html

Is it OK?

ciao -- Nando 

Re: [Quantlib-dev] 0.3.5 preview

[Dirk E. <ed...@de...>]

On Tue, Mar 30, 2004 at 10:45:00AM -0600, Dirk Eddelbuettel wrote:
> On Tue, Mar 30, 2004 at 02:39:42PM +0200, Ferdinando Ametrano wrote:
> > Hi all
> > 
> > the 0.3.5 release is available from the unofficial "hidden" page 
> > http://quantlib.org/nextrelease.html
> > 
> > I would appreciate if anyone could check its own (Win32, Unix, RPM, Debian, 
> > ect) stuff.

The Debian text, as it currently stand on the page, is plain WRONG as it
refers to the testing distribution containing 0.3.4 -- as opposed to the
unstable distribution containing 0.3.5.

Dirk

--
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

Re: [Quantlib-dev] 0.3.5 preview

[Dirk E. <ed...@de...>]

On Tue, Mar 30, 2004 at 02:39:42PM +0200, Ferdinando Ametrano wrote:
> Hi all
> 
> the 0.3.5 release is available from the unofficial "hidden" page 
> http://quantlib.org/nextrelease.html
> 
> I would appreciate if anyone could check its own (Win32, Unix, RPM, Debian, 
> ect) stuff.

If we list / ship the binary rpms, should we ship debs too? You'll find them at

http://ftp.debian.org/debian/pool/main/q/quantlib/
http://ftp.debian.org/debian/pool/main/q/quantlib-python/
http://ftp.debian.org/debian/pool/main/q/quantlib-refman/
http://ftp.debian.org/debian/pool/main/q/quantlib-refman-html/
http://ftp.debian.org/debian/pool/main/q/quantlib-ruby/

and you could link directly to the i386 (or all) packages, as well as these
directories for the other binaries.  URLs also work as

http://ftp.de.debian.org/debian/pool/main/q/quantlib/
http://ftp.fr.debian.org/debian/pool/main/q/quantlib/
http://ftp.es.debian.org/debian/pool/main/q/quantlib/
http://ftp.it.debian.org/debian/pool/main/q/quantlib/
http://ftp.uk.debian.org/debian/pool/main/q/quantlib/
[...]


Nando, Luigi: Any word on my problem with the barrier option?  I had the
same problem last night with plain American options as soon as I tried any
of the Greeks. I must be doing something wrong, but still, it shouldn't
seg.fault, should it?

Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

[Quantlib-dev] 0.3.5 preview

[Ferdinando A. <na...@am...>]

Hi all

the 0.3.5 release is available from the unofficial "hidden" page 
http://quantlib.org/nextrelease.html

I would appreciate if anyone could check its own (Win32, Unix, RPM, Debian, 
ect) stuff.

Tomorrow I will announce the release, which will be then available from the 
usual http://quantlib.org/download.html page

Mark: anything I should write and/or link to for the fink package?

thank you everybody

ciao -- Nando

Re: [Quantlib-dev] NewYork Holiday calendar

[Dirk E. <ed...@de...>]

On Mon, Mar 29, 2004 at 05:53:05PM +0200, Andre Louw wrote:
> Anybody!!!!

Please chill. We're all volunteers here. 
 
> I am urgently trying to establish if Friday 31/12/2004 is a holiday, 
> according to our forward brokers it's not (and I'm inclined to go with 
> them), but QuantLib has it as one!

For what it;s worth, Bloomberg's CDR function does not show it as a 2004
holiday for the exchanges either, but the US Govt Bond market is listed as
having an early close at 2:00pm (i.e. no holiday either).

Hth, Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

[Quantlib-dev] NewYork Holiday calendar

[Andre L. <an...@de...>]

Anybody!!!!

I am urgently trying to establish if Friday 31/12/2004 is a holiday, 
according to our forward brokers it's not (and I'm inclined to go with 
them), but QuantLib has it as one!

Thanx

Andre

[Quantlib-dev] NewYork calendar - 31st December?

[Andre L. <an...@de...>]

Hi,

QuantLib has a rule in the Nework calendar stating that if Jan 1 2005 == 
Saturday, then Friday 31st 2004 == Holiday. Is this correct?

Andre

[Quantlib-dev] barrier options greeks lead to seg.fault.

[Dirk E. <ed...@de...>]

Hi all,

I spent some time rolling my RQuantLib package forward to new 0.3.5
standards. I generally followed the code I saw in test-suite. All in all,
things seems to be working now, and I am starting to get the hang of the
design: bloody elegant if I dare say so!  

One problem, though, seems to be that I no longer get greeks for barriers:

edd@basebud:/tmp> g++ -Wall -o barrier_blows_up barrier_blows_up.cc -lQuantLib
edd@basebud:/tmp> ./barrier_blows_up
Value: 3.73825
Aborted (core dumped)
edd@basebud:/tmp>

which happens as soon as I call   barrierOption.delta()  I tried both the
default engine and the MC engine, but no luck.

A full example, trimmed down from the R interface code I had, is included.
I probably am doing something wromg. Just let me know what :)

Dirk


#include <ql/quantlib.hpp>	// make QuantLib known

using namespace QuantLib;

extern "C" {

  Handle<TermStructure> makeFlatCurve(const Handle<Quote>& forward,
				      DayCounter dc) {
    Date today = Date::todaysDate();
    return Handle<TermStructure>(
	 new FlatForward(today, today, 
			 RelinkableHandle<Quote>(forward), dc));
  }

  Handle<BlackVolTermStructure> makeFlatVolatility(const Handle<Quote>& vol,
                                                     DayCounter dc) {
    Date today = Date::todaysDate();
    return Handle<BlackVolTermStructure>(
         new BlackConstantVol(today, 
			      RelinkableHandle<Quote>(vol), dc));
  }

  int main(void) { 

    char *barrType = "downin";
    char *type = "call";
    double underlying = 100;
    double strike = 100;
    Spread dividendYield = 0.02;
    Rate riskFreeRate = 0.03;
    Time maturity = 0.5;
    int length = int(maturity * 360); // FIXME: this could be better
    double volatility = 0.4;
    double barrier = 90;
    double rebate = 0.0;

    Barrier::Type barrierType;
    if (!strcmp(barrType, "downin")) {
      barrierType = Barrier::DownIn;
    } else if (!strcmp(barrType, "upin")) {
      barrierType = Barrier::UpIn;
    } else if (!strcmp(barrType, "downout")) {
      barrierType = Barrier::DownOut;
    } else if (!strcmp(barrType, "upout")) {
      barrierType = Barrier::UpOut;
    } else {
      std::cerr << "Unexpected barrier type " << barrType << " aborting\n";
    }
    Option::Type optionType;
    if (!strcmp(type, "call")) {
      optionType = Option::Call;
    } else if (!strcmp(type, "put")) {
      optionType = Option::Put;
    } else if (!strcmp(type, "straddle")) {
      optionType = Option::Straddle;
    } else {
      std::cerr << "Unexpected option type " << type << " aborting\n";
    }

    // new QuantLib 0.3.5 framework
    DayCounter dc = Actual360();
    Handle<SimpleQuote> spot(new SimpleQuote(0.0));
    Handle<SimpleQuote> qRate(new SimpleQuote(0.0));
    Handle<TermStructure> qTS = makeFlatCurve(qRate, dc);
    Handle<SimpleQuote> rRate(new SimpleQuote(0.0));
    Handle<TermStructure> rTS = makeFlatCurve(rRate, dc);

    Handle<SimpleQuote> vol(new SimpleQuote(0.0));
    Handle<BlackVolTermStructure> volTS = makeFlatVolatility(vol, dc);

    Date today = Date::todaysDate();

    Date exDate = today.plusDays(length);
    Handle<Exercise> exercise(new EuropeanExercise(exDate));

    spot ->setValue(underlying);
    qRate->setValue(dividendYield);
    rRate->setValue(riskFreeRate);
    vol  ->setValue(volatility);
  
    Handle<StrikedTypePayoff> payoff(new PlainVanillaPayoff(optionType, strike));

    Handle<BlackScholesStochasticProcess> 
      stochProcess(new BlackScholesStochasticProcess(
                 RelinkableHandle<Quote>(spot),
                 RelinkableHandle<TermStructure>(qTS),
                 RelinkableHandle<TermStructure>(rTS),
                 RelinkableHandle<BlackVolTermStructure>(volTS)));

    Size timeSteps = 1;
    bool antitheticVariate = false;
    bool controlVariate = false;
    Size requiredSamples = 10000;
    double requiredTolerance = 0.02;
    Size maxSamples = 1000000;
    bool isBiased = false;

    Handle<PricingEngine> engine(new AnalyticBarrierEngine);
    Handle<PricingEngine> mcEngine(
       new MCBarrierEngine<PseudoRandom>(timeSteps, antitheticVariate,
	  			         controlVariate, requiredSamples,
				         requiredTolerance, maxSamples,
				         isBiased, 5));

    BarrierOption barrierOption(barrierType,
				barrier,
				rebate,
				stochProcess,
				payoff,
				exercise,
				//mcEngine);
				engine);

    std::cout << "Value: " << barrierOption.NPV() << std::endl;
    std::cout << "Delta: " << barrierOption.delta() << std::endl;
    std::cout << "Gamma: " << barrierOption.gamma() << std::endl;
    std::cout << "Vega:  " << barrierOption.vega() << std::endl;
    std::cout << "Theta: " << barrierOption.theta() << std::endl;
    std::cout << "Rho:   " << barrierOption.rho() << std::endl;
    std::cout << "divRho:" << barrierOption.dividendRho() << std::endl;

  }

}


-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

[Quantlib-dev] Problems with the man pages tarball

[Dirk E. <ed...@de...>]

I tried to package the manpages tarball, but there are issues concerning
(file-)name space pollution:

edd@basebud:/var/local/cache/pbuilder/result> wajig install quantlib-manpages_0.3.5-1_all.deb
(Reading database ... 108087 files and directories currently installed.)
Unpacking quantlib-manpages (from quantlib-manpages_0.3.5-1_all.deb) ...
dpkg: error processing quantlib-manpages_0.3.5-1_all.deb (--install):
 trying to overwrite /usr/share/man/man3/index.3.gz', which is also in package manpages-dev
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
dpkg-deb: subprocess paste killed by signal (Broken pipe)
Errors were encountered while processing:
 quantlib-manpages_0.3.5-1_all.deb
edd@basebud:/var/local/cache/pbuilder/result>

man/man3/index.3 is a little too generic. The documentation for one of our
build-tools, dh_installman, mentions 

     dh_installman is a debhelper program that handles installing man pages
     into the correct locations in package build directories. You tell it
     what man pages go in your packages, and it figures out where to install
     them based on the section field in their .TH line. If you have a prop-
     erly formatted .TH line, your man page will be installed into the right
     directory, with the right name (this includes proper handling of pages
     with a subsection, like "3perl", which are placed in man3, and given an
     extension of ".3perl"). If your .TH line is incorrect or missing, the
     program may guess wrong based on the file extension.

so I was hopeful, but it turns out that we already have a .TH line:

basebud:/home/edd/chibud/src/debian/QuantLib/QuantLib-manpages-0.3.5# head -1 man/man3/index.3
.TH "index" 3 "23 Mar 2004" "QuantLib" \" -*- nroff -*-

and I don't quite understand why that doesn't cut it. 

What would people think if we a) either told doxygen to create all pages as
.3ql (if there is such a switch), or barring that, rename all file .3 into
.3ql ourselves?

Another problem came up overnight:


Delivery-date: Sun, 28 Mar 2004 07:40:38 -0600
From: ro...@ed... (Anacron)
To: ro...@ed...
Subject: Anacron job 'cron.daily'

/etc/cron.daily/man-db:
mandb: warning: whatis for history.3.gz exceeds 2048 bytes, truncating.


So it looks as if at least that page isn't created quite right.

Shall we tackly these issues, or should we be content with the other
documentation (already in Debian as .pdf and .html) ?

Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

[Quantlib-dev] Re: [Quantlib-users] Re: boost/version.hpp is missing in the CVS snapshot

[Ferdinando A. <na...@am...>]

Hi Dirk,

> > In QuantLib 0.3.5, which will be released by the end of this month, the
> > boost library is optional. In 0.3.6 the boost library will be mandatory.
>
>Grrr, had I known that I would have build 0.3.5 with boost support for Debian.


sorry Dirk, we probably write too little about development on our mailing 
lists... anyway no features are missing from the non-boost version.

The boost issue is the first item in NEWS.TXT, and the announcement will 
point that out.

Since 0.3.5 will be released on March 31st, can we just rebuild for Debian? 
To enforce boost dependency in the QuantLib code it's as easy as 
uncommenting a line of code.

BTW what about not releasing binaries for QuantLib? I've decided not to 
provide binaries for Win32, since that would have been 30Mb (80MB if you 
consider Debug libraries too). The strong point should be that Quantlib 
users are C++ developers: they should be able to compile QuantLib on their 
own. Of course I'm not talking about QL_Python, etc.

ciao -- Nando 

[Quantlib-dev] RPMs for Final 0.3.5 tarballs

[Liguo S. <lig...@va...>]

Hi, Nando,

The RPM files are available now. Here is the link
http://nlog.phy.vanderbilt.edu/SoftwareProjects/QuantLib/

Let me know if there is any problem.


Liguo




Ferdinando Ametrano wrote:

> Hi Liguo
>
>> I will compile the RPM packages as soon as the tarballs are 
>> finalized. (Is it already?)
>
> the packages at www.quantlib.org/gm/ are final. I will upload them to 
> SF along with the Win32 packages by the end of the week. Liguo: as 
> soon as you have the RPMs just send me a link and I will upload them too.
>
> I will announce the 0.3.5 on March 31. By that date I guess Debian and 
> Fink will be ready too.
>
> ciao -- Nando
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Quantlib-dev mailing list
> Qua...@li...
> https://lists.sourceforge.net/lists/listinfo/quantlib-dev

[Quantlib-dev] Re: [Quantlib-users] Re: boost/version.hpp is missing in the CVS snapshot

[Dirk E. <ed...@de...>]

Hi Nando

On Thu, Mar 25, 2004 at 04:56:27PM +0100, Ferdinando Ametrano wrote:
> Hi Dirk,
> 
> >> In QuantLib 0.3.5, which will be released by the end of this month, the
> >> boost library is optional. In 0.3.6 the boost library will be mandatory.
> >
> >Grrr, had I known that I would have build 0.3.5 with boost support for 
> >Debian.
> 
> 
> sorry Dirk, we probably write too little about development on our mailing 
> lists... anyway no features are missing from the non-boost version.
> 
> The boost issue is the first item in NEWS.TXT, and the announcement will 
> point that out.

Yeah, my bad. Too many balls in the air ... I guess I missed the News.txt
file.

> Since 0.3.5 will be released on March 31st, can we just rebuild for Debian? 
> To enforce boost dependency in the QuantLib code it's as easy as 
> uncommenting a line of code.

Yes, that is probably a good idea. Are you guys still hacking on Debian
machines. Is what I need among these? Libboost itself, do I need any add-ons?

edd@homebud:~/debian/QuantLib/QuantLib-0.3.5> apt-cache search libboost
libboost-date-time-dev - Set of date-time libraries based on generic programming concepts
libboost-date-time1.30.2 - Set of date-time libraries based on generic programming concepts
libboost-dev - Boost.org libraries development files
libboost-doc - Boost.org libraries documentation
libboost-graph-dev - Generic graph components and algorithms in C++
libboost-graph1.30.2 - Generic graph components and algorithms in C++
libboost-python-dev - The Boost Python Library development files
libboost-python1.30.2 - Boost Python Library
libboost-regex-dev - Regular expression library for C++
libboost-regex1.30.2 - Regular expression library for C++
libboost-signals-dev - Managed signals and slots library for C++
libboost-signals1.30.2 - Managed signals and slots library for C++
libboost-test-dev - Components for writing and executing test suites
libboost-test1.30.2 - Components for writing and executing test suites
libboost-thread-dev - Portable C++ multi-threading
libboost-thread1.30.2 - Portable C++ multi-threading
edd@homebud:~/debian/QuantLib/QuantLib-0.3.5> 
 
> BTW what about not releasing binaries for QuantLib? I've decided not to 
> provide binaries for Win32, since that would have been 30Mb (80MB if you 
> consider Debug libraries too). The strong point should be that Quantlib 
> users are C++ developers: they should be able to compile QuantLib on their 
> own. Of course I'm not talking about QL_Python, etc.

The quantlib-examples package has become semi-famous in mail exchanges I am
having with the openMosix folks as I use BermudanSwaption as a migration
test for openMosix :)

The package is only 2mb under Debian, and fully optional. That strikes me as
the right thing to do.


Dirk

-- 
The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page

Re: [Quantlib-dev] Final tarballs

[Ferdinando A. <na...@am...>]

Hi Liguo

>I will compile the RPM packages as soon as the tarballs are finalized. (Is 
>it already?)
the packages at www.quantlib.org/gm/ are final. I will upload them to SF 
along with the Win32 packages by the end of the week. Liguo: as soon as you 
have the RPMs just send me a link and I will upload them too.

I will announce the 0.3.5 on March 31. By that date I guess Debian and Fink 
will be ready too.

ciao -- Nando