quantlib-dev — Discussion of QuantLib development

Re: [Quantlib-dev] Adding new functions to blackcalculator

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

On Wed, 2010-03-03 at 09:47 +0100, Dima wrote:
> Thanks Luigi and sorry for not having answered earlier. I'm busy. I
> tried to build QL from the latest trunk
> but was not sucessfull. And the stuff you added is not in the official
> 1.0 version, so I couldn't really check. 
> Don't know what happened to the trunk files, they seem to be
> corrupted. For example, in smilesection.hpp 
> I find 
> 
> <<<<<<<<<<.mine
> 
> in the code. Its either a problem here or in the trunk.

The trunk is clean.  You modified the code on your machine, made a svn
update, and got a conflict between your local changes and the changes
pulled in by the update. Subversion doesn't know how to merge them and
keeps both sets of changes between markers such as the one above. You'll
have to resolve the conflict manually. You can see
<http://svnbook.red-bean.com/en/1.5/svn.tour.cycle.html#svn.tour.cycle.resolve> for more details.

Luigi


-- 

The box said "Use Windows 95 or better," so I got a Macintosh. 

Re: [Quantlib-dev] Adding new functions to blackcalculator

[Dima <dim...@go...>]

I see, thanks. Will get back to you

2010/3/3 Luigi Ballabio <lui...@gm...>

> On Wed, 2010-03-03 at 09:47 +0100, Dima wrote:
> > Thanks Luigi and sorry for not having answered earlier. I'm busy. I
> > tried to build QL from the latest trunk
> > but was not sucessfull. And the stuff you added is not in the official
> > 1.0 version, so I couldn't really check.
> > Don't know what happened to the trunk files, they seem to be
> > corrupted. For example, in smilesection.hpp
> > I find
> >
> > <<<<<<<<<<.mine
> >
> > in the code. Its either a problem here or in the trunk.
>
> The trunk is clean.  You modified the code on your machine, made a svn
> update, and got a conflict between your local changes and the changes
> pulled in by the update. Subversion doesn't know how to merge them and
> keeps both sets of changes between markers such as the one above. You'll
> have to resolve the conflict manually. You can see
> <
> http://svnbook.red-bean.com/en/1.5/svn.tour.cycle.html#svn.tour.cycle.resolve>
> for more details.
>
> Luigi
>
>
> --
>
> The box said "Use Windows 95 or better," so I got a Macintosh.
>
>
>

Re: [Quantlib-dev] [Quantlib-users] QuantlibXL with Office 2003 Service Pack 3

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

On Wed, Mar 3, 2010 at 9:09 AM, Yang Ye <lea...@ya...> wrote:
> We got a problem for functions returning double with Office 2003 SP3.
> E.g.  function qlSmileSectionVolatility. Sometimes, it return just 0;
> Sometimes, it return a very big number (negative/positive). We could
> trace to
>
> returnValue = ObjectIdLibObjPtr->volatility(
>               StrikeCpp);
> return &returnValue;
>
> returnValue is still correct before return. But Excel displayed the
> wrong number.

this is the second report I receive about this kind of problem. what
version of Excel are you using?

An user wrote me about Excel 2003 (11.8169.8172) SP3 having the bug
while Excel 2003 (11.8237.8221) SP3 is working correctly.
I use Excel 2003 (11.8220.8202) SP3 and don't have this problem.

I googled around with no luck, I hope somebody more knowledgeable will help

ciao -- Nando

Re: [Quantlib-dev] Adding new functions to blackcalculator

[Dima <dim...@go...>]

Thanks Luigi and sorry for not having answered earlier. I'm busy. I tried to
build QL from the latest trunk
but was not sucessfull. And the stuff you added is not in the official 1.0
version, so I couldn't really check.
Don't know what happened to the trunk files, they seem to be corrupted. For
example, in smilesection.hpp
I find

<<<<<<<<<<.mine

in the code. Its either a problem here or in the trunk.



2010/2/15 Luigi Ballabio <lui...@gm...>

> On Tue, 2009-04-21 at 09:49 +0200, Dima wrote:
> > I've uploaded
> > my current working version of the deltacalculator including a
> > testsuite(your recent change not incl.).
> >
> > You can get it here:
> >
> > longvega.com/DeltaEngine.zip
> >
>
> Dimitri,
>        I've added your stuff to the Subversion trunk under
> q;/experimental/fx.
> May you check it out and see that I haven't broken anything?
>
> Thanks,
>        Luigi
>
>
> --
>
> Academic: a term of opprobrium applied to those that do their job well
> by those who cannot.
> -- Sir Ernest Gowers
>
>
>

Re: [Quantlib-dev] I think I have a need for another pricing engine for the bond instrument

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

On Tue, Mar 2, 2010 at 9:21 PM, Nathan Abbott <nka...@gm...> wrote:
> the user wants the NPV of the bond to be the amount
> purchased * clean price  / 100.

and you know what... he's simply right

The current engine does calculate the dirty price
(results_.settlementValue) and some "npv" (results_.value) of
questionable usefulness. My reservation about the current npv are:
1) it does include cashflows before or on bond settlement date, which
in most situation do not belong to the npv
2) is valid for an holding amount of faceAmount (typically 100)

> I am thinking the best way to solve this
> problem is to create another pricing engine for the bond that calculates the
> NPV of the bond to equal the amount purchased * the clean price / 100.

the real issue in not the engine, but the current lack for the
"holding amount" piece of info
I would put it in the Bond class since it inherits from Instrument and
so it should be homogeneous to other Instruments in portfolio
evaluation.

Luigi?

ciao -- Nando

[Quantlib-dev] I think I have a need for another pricing engine for the bond instrument

[Nathan A. <nka...@gm...>]

I think I have a need for another pricing engine for the bond instrument. My
users want to place bond in a portfolio that contains different type of
instruments. Right now, I place all the instruments in a collection and then
loop through the collection and call the NPV method of each instrument. The
problem is that the user wants the NPV of the bond to be the amount
purchased * clean price  / 100.  I am thinking the best way to solve this
problem is to create another pricing engine for the bond that calculates the
NPV of the bond to equal the amount purchased * the clean price / 100.  Is
this a common proiblem? If I wrote the engine, would people want it include
in QuantLib?

[Quantlib-dev] [PATCH] Hybrid products in mcbasket

[Andrea <mar...@go...>]

This patch (which contains all outstanding patches for experimental/mcbasket)
enables products to price instruments depending on the yieldTermStructure (which might be stochastic).

For the time being it is still deterministic, but in the future it might change.

The key method of the payoff class is

        virtual void value(const Matrix       & path,
                           const std::vector<Handle<YieldTermStructure> > & forwardTermStructures,
                           Array              & payments,
                           Array              & exercises,
                           std::vector<Array> & states) const = 0;

the 2nd argument is the yield term structure on each fixing date.
An example to use it is

member variables

private:
    mutable RelinkableHandle<YieldTermStructure> m_termStructure;
    boost::shared_ptr<IborIndex> m_euribor3m;
	

then in value()

        const Handle<YieldTermStructure> & yieldTermStructure = forwardTermStructures[i];
        const Date & referenceDate = yieldTermStructure->referenceDate();

        // relink term structure to actual value
        m_termStructure.linkTo(yieldTermStructure.currentLink());

        const Date spotDate = m_euribor3m->fixingCalendar().adjust(referenceDate);
        const Rate spotEuribor = m_euribor3m->fixing(spotDate);

Attached the patch and a little example.

Andrea

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17110] trunk/QuantLib/ql/math/interpolations/ mixedinterpolation.hpp

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

On Thu, Feb 25, 2010 at 4:45 PM, Luigi Ballabio
<lui...@gm...> wrote:
> You can remove the warning by removing
>
> xBegin2_(this->xBegin_+n)
>
> from the initializer list and adding
>
> xBegin2_ = this->xBegin_+n;
>
> to the body of the constructor.

done, thank you

ciao -- Nando

Re: [Quantlib-dev] QuantLib 1.0 released

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

On Thu, 2010-02-25 at 14:01 +0900, dikshie wrote:
> successfully run quantlib-1.0 on FreeBSD.
> I filled FreeBSD's problem report for quantlib-1.0:
> http://www.freebsd.org/cgi/query-pr.cgi?pr=144276
> 
> quantlib-1.0 probably will be committed to FreeBSD's ports tree
> after FreeBSD-7.3 release.

I didn't even know you were maintaining a port for FreeBSD.
Thanks for the support.

Luigi


-- 

The first thing we do, let's kill all the lawyers. 
-- W. Shakespeare, "King Henry VI, Part II" 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17110] trunk/QuantLib/ql/math/interpolations/ mixedinterpolation.hpp

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

On Wed, 2010-02-17 at 18:46 +0100, Ferdinando Ametrano wrote:
> On Wed, Feb 17, 2010 at 10:36 AM,  <lba...@us...> wrote:
> > I have no idea what VC9 complains about, but the 'this->' is
> > required by the C++ standard.
> 
> see Visual Studio Help below
> 
> ================================================
> 
> Error Message
> 'this' : used in base member initializer list
> 
> 
> The this pointer is valid only within nonstatic member functions. It
> cannot be used in the initializer list for a base class.

I see.  Except we're using it for our own data members, not for a base
class.

Oh well.  You can remove the warning by removing 

xBegin2_(this->xBegin_+n)

from the initializer list and adding

xBegin2_ = this->xBegin_+n;

to the body of the constructor.

Luigi


-- 

All generalizations are false, including this one. 
-- Mark Twain 

Re: [Quantlib-dev] QuantLib 1.0 released

[dikshie <di...@gm...>]

On Thu, Feb 25, 2010 at 1:35 AM, Luigi Ballabio
<lui...@gm...> wrote:
>
> QuantLib is a cross-platform, free/open-source quantitative finance C++
> library for modeling, pricing, trading, and risk management in
> real-life.
>
> Version 1.0 has been released and is available for download at
> <http://quantlib.org/download.shtml>.

successfully run quantlib-1.0 on FreeBSD.
I filled FreeBSD's problem report for quantlib-1.0:
http://www.freebsd.org/cgi/query-pr.cgi?pr=144276

quantlib-1.0 probably will be committed to FreeBSD's ports tree
after FreeBSD-7.3 release.

thanks!

-dikshie-
"FreeBSD's quantlib ports maintainer"

Re: [Quantlib-dev] QuantLib 1.0 released

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

On Wed, 2010-02-24 at 17:35 +0100, Luigi Ballabio wrote:
> Version 1.0 has been released

...and let me thank all those who made it possible, especially Nando and
Eric.  I risked my sanity once or twice during the release process, and
man, it's good to see the release out there.  Now I just hope that any
unknown bugs stay hidden for a few more weeks...

Thanks again to everybody,
	Luigi


-- 

Every solution breeds new problems. 
-- unknown 

[Quantlib-dev] QuantLib 1.0 released

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

QuantLib is a cross-platform, free/open-source quantitative finance C++
library for modeling, pricing, trading, and risk management in
real-life.

Version 1.0 has been released and is available for download at
<http://quantlib.org/download.shtml>.

Please log any problems you have with this release in the SourceForge
bug tracker at
<http://sourceforge.net/tracker/?group_id=12740&atid=112740>
specifying that you're using QuantLib 1.0.

        The QuantLib group

[Quantlib-dev] ugly unreasonable patch, help would be appreciated

[Andrea <mar...@go...>]

Hi,

I saw this comment and read the code for curiosity:

for (; cf>=leg.begin() && (*cf)->date()==paymentDate; --cf)

my question is

is it legal to decrement an iterator when it is at .begin()?
can this ever be false: "cf>=leg.begin()"?

it reminds me the trick to do a backward for loop with a size_t.

or maybe it is just nonsense...

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17128] trunk/QuantLib/ql/cashflows/cashflows.cpp

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

On Mon, 2010-02-22 at 20:05 +0100, Ferdinando Ametrano wrote:
> On Mon, Feb 22, 2010 at 9:36 AM, Luigi Ballabio
> <lui...@gm...> wrote:
> > Try using reverse iterators.
> 
> following your suggestion I found out a bug in the previousAmount and
> previousRate functions, which was easily fixed switching to reverse
> iterator (please find the patch below).
> 
> Unfortunately the fix is not backward compatible as it requires
> CashFlows::previousCashFlow and BondFunctions::previousCashFlow to
> return a Leg::const_reverse_iterator instead of Leg::const_iterator.

Well, it doesn't strictly require it---you could retrieve the iterator
from the reverse iterator.  But I see that the common use case would be
to turn it into a reverse itereator again, so it makes more sense to
change the signature.

<sigh> Ok, I'm applying it. Commit the thing on the trunk asap.

Luigi


-- 

Do the right thing. It will gratify some people and astonish the rest. 
-- Mark Twain 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17128] trunk/QuantLib/ql/cashflows/cashflows.cpp

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

Hi Luigi

On Mon, Feb 22, 2010 at 9:36 AM, Luigi Ballabio <lui...@gm...>
wrote:
> Try using reverse iterators.

following your suggestion I found out a bug in the previousAmount and
previousRate functions, which was easily fixed switching to reverse iterator
(please find the patch below).

Unfortunately the fix is not backward compatible as it
requires CashFlows::previousCashFlow and BondFunctions::previousCashFlow to
return a Leg::const_reverse_iterator instead of Leg::const_iterator. I know
you just tagged 1.0, but this fix might be worth porting to 1.0 before
release. What is your opinion ?

ciao -- Nando

PS If you go for fixing the 1.0, less relevant backward-compatible bug-fix
Rev17144 and Rev17143 can be ported too

============ patch ===============================================


Index: ql/cashflows/cashflows.cpp
===================================================================
--- ql/cashflows/cashflows.cpp (revision 17128)
+++ ql/cashflows/cashflows.cpp (working copy)
@@ -81,29 +81,22 @@
        return true;
    }

-    Leg::const_iterator
+    Leg::const_reverse_iterator
    CashFlows::previousCashFlow(const Leg& leg,
                                bool includeSettlementDateFlows,
                                Date settlementDate) {
        if (leg.empty())
-            return leg.end();
+            return leg.rend();

        if (settlementDate == Date())
            settlementDate = Settings::instance().evaluationDate();

-        if ( ! (*leg.begin())->hasOccurred(settlementDate,
-                                           includeSettlementDateFlows) )
-            return leg.end();
-
-        Leg::const_iterator i = nextCashFlow(leg,
-                                             includeSettlementDateFlows,
-                                             settlementDate);
-        // --i is not what we're looking for since there
-        // might be more than one CashFlow at (*--i)->date()
-        Date beforeLastPaymentDate = (*--i)->date()-1;
-        return nextCashFlow(leg,
-                            includeSettlementDateFlows,
-                            beforeLastPaymentDate);
+        Leg::const_reverse_iterator i;
+        for (i = leg.rbegin(); i<leg.rend(); ++i) {
+            if ( (*i)->hasOccurred(settlementDate,
includeSettlementDateFlows) )
+                return i;
+        }
+        return leg.rend();
    }

    Leg::const_iterator
@@ -127,61 +120,52 @@
    Date CashFlows::previousCashFlowDate(const Leg& leg,
                                         bool includeSettlementDateFlows,
                                         Date settlementDate) {
-        Leg::const_iterator cf = previousCashFlow(leg,
-
 includeSettlementDateFlows,
-                                                  settlementDate);
-        if (cf==leg.end()) return Date();
+        Leg::const_reverse_iterator cf;
+        cf = previousCashFlow(leg, includeSettlementDateFlows,
settlementDate);
+
+        if (cf==leg.rend())
+            return Date();
+
        return (*cf)->date();
    }

    Date CashFlows::nextCashFlowDate(const Leg& leg,
                                     bool includeSettlementDateFlows,
                                     Date settlementDate) {
-        Leg::const_iterator cf = nextCashFlow(leg,
-                                              includeSettlementDateFlows,
-                                              settlementDate);
-        if (cf==leg.end()) return Date();
+        Leg::const_iterator cf;
+        cf = nextCashFlow(leg, includeSettlementDateFlows, settlementDate);
+
+        if (cf==leg.end())
+            return Date();
+
        return (*cf)->date();
    }

    Real CashFlows::previousCashFlowAmount(const Leg& leg,
                                           bool includeSettlementDateFlows,
                                           Date settlementDate) {
-        Leg::const_iterator cf = previousCashFlow(leg,
-
 includeSettlementDateFlows,
-                                                  settlementDate);
-        if (cf==leg.end()) return Real();
+        Leg::const_reverse_iterator cf;
+        cf = previousCashFlow(leg, includeSettlementDateFlows,
settlementDate);

+        if (cf==leg.rend())
+            return Real();
+
        Date paymentDate = (*cf)->date();
        Real result = 0.0;
-
-        // when cf==leg.begin() the following code crashes at --cf
-        // on VC8/9 Debug (boundary check)
-        // but it also crashes in Release mode as if it would evaluate
-        // (*cf) even when cf<leg.begin()
-        //
-        // help or suggestion would be appreciated
-        //
-        //for (; cf>=leg.begin() && (*cf)->date()==paymentDate; --cf)
-        //    result += (*cf)->amount();
-
-        // ugly patch...
-        for (; cf>leg.begin() && (*cf)->date()==paymentDate; --cf)
+        for (; cf<leg.rend() && (*cf)->date()==paymentDate; ++cf)
            result += (*cf)->amount();
-        if (cf==leg.begin() && (*cf)->date()==paymentDate)
-            result += (*cf)->amount();
-
        return result;
    }

    Real CashFlows::nextCashFlowAmount(const Leg& leg,
                                       bool includeSettlementDateFlows,
                                       Date settlementDate) {
-        Leg::const_iterator cf = nextCashFlow(leg,
-                                              includeSettlementDateFlows,
-                                              settlementDate);
-        if (cf==leg.end()) return Real();
+        Leg::const_iterator cf;
+        cf = nextCashFlow(leg, includeSettlementDateFlows, settlementDate);

+        if (cf==leg.end())
+            return Real();
+
        Date paymentDate = (*cf)->date();
        Real result = 0.0;
        for (; cf<leg.end() && (*cf)->date()==paymentDate; ++cf)
@@ -192,18 +176,20 @@
    // Coupon utility functions
    namespace {

+        template<typename Iter>
        Rate aggregateRate(const Leg& leg,
-                           Leg::const_iterator cf) {
-            if (cf==leg.end()) return 0.0;
+                           Iter first,
+                           Iter last) {
+            if (first==last) return 0.0;

-            Date paymentDate = (*cf)->date();
+            Date paymentDate = (*first)->date();
            bool firstCouponFound = false;
            Real nominal = 0.0;
            Time accrualPeriod = 0.0;
            DayCounter dc;
            Rate result = 0.0;
-            for (; cf<leg.end() && (*cf)->date()==paymentDate; ++cf) {
-                shared_ptr<Coupon> cp = dynamic_pointer_cast<Coupon>(*cf);
+            for (; first<last && (*first)->date()==paymentDate; ++first) {
+                shared_ptr<Coupon> cp =
dynamic_pointer_cast<Coupon>(*first);
                if (cp) {
                    if (firstCouponFound) {
                        QL_REQUIRE(nominal       == cp->nominal() &&
@@ -221,7 +207,7 @@
                }
            }
            QL_ENSURE(firstCouponFound,
-                      "next cashflow (" << paymentDate << ") is not a
coupon");
+                      "no coupon paid at cashflow date " << paymentDate);
            return result;
        }

@@ -230,19 +216,18 @@
    Rate CashFlows::previousCouponRate(const Leg& leg,
                                       bool includeSettlementDateFlows,
                                       Date settlementDate) {
-        Leg::const_iterator cf = previousCashFlow(leg,
-
 includeSettlementDateFlows,
-                                                  settlementDate);
-        return aggregateRate(leg, cf);
+        Leg::const_reverse_iterator cf;
+        cf = previousCashFlow(leg, includeSettlementDateFlows,
settlementDate);
+
+        return aggregateRate<Leg::const_reverse_iterator>(leg, cf,
leg.rend());
    }

    Rate CashFlows::nextCouponRate(const Leg& leg,
                                   bool includeSettlementDateFlows,
                                   Date settlementDate) {
-        Leg::const_iterator cf = nextCashFlow(leg,
-                                              includeSettlementDateFlows,
-                                              settlementDate);
-        return aggregateRate(leg, cf);
+        Leg::const_iterator cf;
+        cf = nextCashFlow(leg, includeSettlementDateFlows, settlementDate);
+        return aggregateRate<Leg::const_iterator>(leg, cf, leg.end());
    }

    Date CashFlows::accrualStartDate(const Leg& leg,
Index: ql/cashflows/cashflows.hpp
===================================================================
--- ql/cashflows/cashflows.hpp (revision 17127)
+++ ql/cashflows/cashflows.hpp (working copy)
@@ -55,7 +55,7 @@
        //! \name CashFlow functions
        //@{
        //! the last cashflow paying before or at the given date
-        static Leg::const_iterator
+        static Leg::const_reverse_iterator
        previousCashFlow(const Leg& leg,
                         bool includeSettlementDateFlows,
                         Date settlementDate = Date());
Index: ql/pricingengines/bond/bondfunctions.cpp
===================================================================
--- ql/pricingengines/bond/bondfunctions.cpp (revision 17144)
+++ ql/pricingengines/bond/bondfunctions.cpp (working copy)
@@ -47,8 +47,9 @@
        return bond.notional(settlement)!=0.0;
    }

-    Leg::const_iterator BondFunctions::previousCashFlow(const Bond& bond,
-                                                        Date settlement) {
+    Leg::const_reverse_iterator
+    BondFunctions::previousCashFlow(const Bond& bond,
+                                    Date settlement) {
        if (settlement == Date())
            settlement = bond.settlementDate();

Index: ql/pricingengines/bond/bondfunctions.hpp
===================================================================
--- ql/pricingengines/bond/bondfunctions.hpp (revision 17127)
+++ ql/pricingengines/bond/bondfunctions.hpp (working copy)
@@ -60,8 +60,9 @@

        //! \name CashFlow inspectors
        //@{
-        static Leg::const_iterator previousCashFlow(const Bond& bond,
-                                                    Date refDate = Date());
+        static Leg::const_reverse_iterator
+        previousCashFlow(const Bond& bond,
+                         Date refDate = Date());
        static Leg::const_iterator nextCashFlow(const Bond& bond,
                                                Date refDate = Date());
        static Date previousCashFlowDate(const Bond& bond,

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17134] trunk/QuantLib/ql/time/schedule.cpp

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

On Fri, 2010-02-19 at 15:57 +0000, na...@us... wrote:
> Revision: 17134
>           http://quantlib.svn.sourceforge.net/quantlib/?rev=17134&view=rev
> Author:   nando
> Date:     2010-02-19 15:57:38 +0000 (Fri, 19 Feb 2010)
> 
> Log Message:
> -----------
> added tolerance for firstDate==effectiveDate and nextToLastDate==terminationDate
> 
> Modified Paths:
> --------------
>     trunk/QuantLib/ql/time/schedule.cpp

It would be nice to have test cases for these changes, wouldn't it? :)

Luigi


-- 

Any software problem can be solved by adding another layer of
indirection. 
-- Steven M. Bellovin 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17128] trunk/QuantLib/ql/cashflows/cashflows.cpp

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

On Fri, 2010-02-19 at 11:36 +0000, na...@us... wrote:
> Revision: 17128
>           http://quantlib.svn.sourceforge.net/quantlib/?rev=17128&view=rev
> Author:   nando
> Date:     2010-02-19 11:36:03 +0000 (Fri, 19 Feb 2010)
> 
> Log Message:
> -----------
> ugly unreasonable patch, help would be appreciated

Try using reverse iterators.

Luigi


-- 

Don't let school get in the way of your education. 
-- Mark Twain 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17110] trunk/QuantLib/ql/math/interpolations/ mixedinterpolation.hpp

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

On Wed, Feb 17, 2010 at 10:36 AM,  <lba...@us...> wrote:
> I have no idea what VC9 complains about, but the 'this->' is
> required by the C++ standard.

see Visual Studio Help below

ciao -- Nando

================================================

Error Message
'this' : used in base member initializer list


The this pointer is valid only within nonstatic member functions. It
cannot be used in the initializer list for a base class.

The base-class constructors and class member constructors are called
before this constructor. In effect, you've passed a pointer to an
unconstructed object to another constructor. If those other
constructors access any members or call member functions on this, the
result will be undefined. You should not use the this pointer until
all construction has completed.

This is a level-1 warning under Microsoft extensions (/Ze) and a
level-4 warning otherwise.

The following sample generates C4355:


// C4355.cpp
// compile with: /W1 /c
#include <tchar.h>

class CDerived;
class CBase {
public:
   CBase(CDerived *derived): m_pDerived(derived) {};
   ~CBase();
   virtual void function() = 0;

   CDerived * m_pDerived;
};

class CDerived : public CBase {
public:
   CDerived() : CBase(this) {};   // C4355 "this" used in derived c'tor
   virtual void function() {};
};

CBase::~CBase() {
   m_pDerived -> function();
}

int main() {
   CDerived myDerived;
}

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17105] trunk/QuantLib/ql/cashflows

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

On Wed, 2010-02-17 at 16:11 +0100, Ferdinando Ametrano wrote:
> On Wed, Feb 17, 2010 at 10:21 AM, Luigi Ballabio
> <lui...@gm...> wrote:
> > On Tue, 2010-02-16 at 19:09 +0000, na...@us... wrote:
> >> added more Coupon inspectors
> >
> > We're duplicating code all over the place---they're all basically the
> > same loop.  Why don't you write a nextCoupon() method instead and call
> > whatever method you want on the result?
> 
> you're right and this is true for both nextCashFlow and nextCoupon,
> their "previous" variants, and their CashFlows and Bond flavors
> 
> Anyway if bond_ is a shared_ptr<Bond> you would then have to write
> 
> (*BondFunctions::nextCashFlow(*bond_))->date();
> 
> instead of
> 
> bond_->nextCashFlowDate();

Worse---You'd have to check for a null iterator first.  Ok, I can see
it's good to have the whole bunch of functions for convenience.  Plus
there's the aggregation thing.  It might still be worth to write the
common loop in a nextCoupon() method (private, if need be) so that at
least the functions returning a date can call it and remove a good part
of the duplication.  For the aggregation, I don't know.  I had though of
returning a range (begin,end) of the cashflows on the same date, but it
wouldn't work if there was a non-coupon cashflow between two coupons.
Maybe a composite, as you say.

Well, I'll let you know if I get any ideas...

Luigi


-- 

A programming language is low-level when its programs require attention 
to the irrelevant. 
-- Alan Perlis 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17105] trunk/QuantLib/ql/cashflows

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

On Wed, Feb 17, 2010 at 10:21 AM, Luigi Ballabio
<lui...@gm...> wrote:
> On Tue, 2010-02-16 at 19:09 +0000, na...@us... wrote:
>> added more Coupon inspectors
>
> We're duplicating code all over the place---they're all basically the
> same loop.  Why don't you write a nextCoupon() method instead and call
> whatever method you want on the result?

you're right and this is true for both nextCashFlow and nextCoupon,
their "previous" variants, and their CashFlows and Bond flavors

Anyway if bond_ is a shared_ptr<Bond> you would then have to write

(*BondFunctions::nextCashFlow(*bond_))->date();

instead of

bond_->nextCashFlowDate();

You might have multiple CashFlow on the same date, so nextCashFlow
might actually aggregate flows and create a new AggregateCashFlow
instance, then returning a shared_ptr instead of an iterator. In this
case one would have to write:

BondFunctions::nextCashFlow(*bond_)->date();

As for nextCoupon it should return an instance of some AggregateCoupon
class. This class would have pointers to the Coupons to be aggregated
for its own nominal() and rate() method implementations.

What do you think about it ?

ciao -- Nando

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17107] trunk/QuantLib/ql/cashflows

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

On Wed, Feb 17, 2010 at 10:21 AM, Luigi Ballabio
<lui...@gm...> wrote:
> On Tue, 2010-02-16 at 19:20 +0000, na...@us... wrote:
>> added XXX::operator boost::shared_ptr<Leg>() const
>
> Why? Leg is not polymorphic (it's a std::vector), and shared_ptr<Leg> is
> not used anywhere in the library (I'd probably remove it if it was :)

you're right, my fault: reverted.

thank you

ciao -- Nando

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17105] trunk/QuantLib/ql/cashflows

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

On Tue, 2010-02-16 at 19:09 +0000, na...@us... wrote:
> Revision: 17105
>           http://quantlib.svn.sourceforge.net/quantlib/?rev=17105&view=rev
> Author:   nando
> Date:     2010-02-16 19:09:58 +0000 (Tue, 16 Feb 2010)
> 
> Log Message:
> -----------
> added more Coupon inspectors

We're duplicating code all over the place---they're all basically the
same loop.  Why don't you write a nextCoupon() method instead and call
whatever method you want on the result?

Luigi


-- 

These are my principles, and if you don't like them... Well, I have
others. 
-- Groucho Marx 

Re: [Quantlib-dev] [QuantLib-svn] SF.net SVN: quantlib:[17107] trunk/QuantLib/ql/cashflows

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

On Tue, 2010-02-16 at 19:20 +0000, na...@us... wrote:
> Revision: 17107
>           http://quantlib.svn.sourceforge.net/quantlib/?rev=17107&view=rev
> Author:   nando
> Date:     2010-02-16 19:20:57 +0000 (Tue, 16 Feb 2010)
> 
> Log Message:
> -----------
> added XXX::operator boost::shared_ptr<Leg>() const

Why? Leg is not polymorphic (it's a std::vector), and shared_ptr<Leg> is
not used anywhere in the library (I'd probably remove it if it was :)

Luigi


-- 

I am extraordinarily patient, provided I get my own way in the end. 
-- Margaret Thatcher 

[Quantlib-dev] QLXL/QLA/OH 1.0 beta 3 available

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

On behalf of Eric Ehlers, whose Internet is broken:

===================================================

Hi all,

Quoting Luigi Ballabio <luigi.ballabio@...>: 

> Hi all, 
> another beta for QuantLib 1.0 is available at 
> <https://sourceforge.net/projects/quantlib/files/prerelease/>. 

I have uploaded the QLXL/QLA/OH 1.0 beta 3 tarballs to the same   
directory.  There was no beta 1 or 2 release of QLXL/QLA/OH but I   
called it beta 3 to indicate that it's compatible with the QuantLib   
1.0 beta 3 release which is also available at that link. 

I would be grateful to anyone who could test the files and report any   
problems. 

Kind Regards, 
Eric