The different overloads of `RooAbsRealLValue::inRange()` implemented
different tolerances when checking if a value `x` falls inside a
specific range. Some overloads checked if the interval
`[x - 1e-6, x + 1e6]` is overlapping with the range, an other overload
checked if the interval `[x - 1e-8*x, x + 1e8*x]` is overlapping.

It's better is this is done consistently and predictably so this commit
suggests to leave out these epsilon margins that were never documented.
For backwards compatibility, one can set a custom relative or absolute
epsilon via the `RooNumber` interface.

The `RooGenFunction` was only a lightweight adaptor that exports a
RooAbsReal as a `ROOT::Math::IGenFunction`. The same can be easily
achieved with the generic `ROOT::Math::Functor1D`, so in the spirit of
not duplicating interfaces, the `RooGenFunction` is removed.

The same goes for the `RooMultiGenFunction`.

The git commit log was analyzed to accurately extract the names of all
people that contributed to RooFit in the 6.28 development cycle.

Integrating https://gitlab.cern.ch/will/xroofit by Will Buttinger to get the RooBrowser.

Hidden behind the ROOT 7 flag until C++17 becomes the default minimum standard.

Co-authored-by: Stephan Hageboeck <stephan.hageboeck@cern.ch>

The `RooSimPdfBuilder` that was deprecated since ROOT 5.20 and got
replaced by the `RooSimWSTool` is removed.

The `RooAbsBinning` interface for bin index lookups was changed to
enable vectorized implementations. Instead of having the override
`RooAbsBinning::binNumber()`, the binning implementations now have to
override the `RooAbsBinning::binNumbers()` function to evaluate the bin
indices of multiple values in one function call.

The interface of `RooAbsBinning::binNumbers()` is designed to facilitate
the accregation of bin indices over multiple dimensions.

The former `RooAbsBinning::binNumber()` method is now implement in terms
of the vectorized version, such that we automatically get wide test
coverage. It was verified that this doesn't come with a performance
overhead.

This will greatly facilitate the vectorization of the RooHistPdf and
RooHistFunc later.

In the previous commit, the wrong old-style implementation of
`RooMultiCategory::evaluate()` was removed.

Now, the only place where the `RooMultiCatIter` is still used is the
`RooSuperCategory::MakeIterator()` function, which is already deprecated
since the RooFit category rewrite for ROOT 6.22 (#5502).

It's not worth to carry around the `RooMultiCatIter` just for that, so
this commit suggests to remove the deprecated
`RooSuperCategory::MakeIterator()` function and get rid of the
RooMultiCatIter.

Copy assignment for RooAbsArgs was implemented in an unexpected and
inconsistent way. While one would expect that the copy assignment is
copying the object, it said in the documentation of
`RooAbsArg::operator=` that it will "assign all boolean and string
properties of the original bject. Transient properties and client-server
links are not assigned." This contradicted with the implementation,
where the server links were actually copied too. Indeed it is
questionable that adding also the servers of the `other` RooAbsArg makes
sense when the proxies are not changed.

There was also a memory leak of the `_ownedComponents`, because
`_ownedComponents` got reset to `nullptr` without deleting the
pointed-to owned collection.

Furthermore, in `RooAbsRealLValue`, the assigment operator was
overloaded by a function that only assigns the value of another
`RooAbsReal`.

With all these inconsistencies, it was deemed safer to disable copy
assignment of RooAbsArgs from now on.

* [DF][NFC] Mention change in behavior of Alias in rel notes

* [skip-ci] make sur the http doc is read by doxygen.

Co-authored-by: Enrico Guiraud <enrico.guiraud@cern.ch>

The default minimizer type was specified both in the RooMinimizer
itself, as well as in the `fitTo` methods. This is not optimal, as one
would have to remember to change it everywhere if the defaults are
changed in RooFit.

In RooStats, there were a few places where the RooMinimizer defaults
were obtained via
`ROOT::Math::MinimizerOptions::DefaultMinimizerType` directly. This
is now done for all RooFit libraries to be consistent.

As printed out by the HistFactory in a warning message for a long time
already, setting the `Const` attribute to the `<NormFactor>` tag is
deprecated and it will be ignored. Instead, add
`<ParamSetting Const="True"> myparam </ParamSetting>`
to your top-level XML's `<Measurement>` entry.

This deprecation implied that the constant parameter flag in the
`RooStats:HistFactory::NormFactor` class had no effect as well. To avoid
ambiguity in the future, the possibility to set and retrieve this flag with
`NormFactor::SetConst()` and `NormFactor::GetConst()` was removed, and the
aforementioned deprecated warning is not printed anymore.

The longtime-deprecated `RooStats::HistFactory::EstimateSummary` class
is removed, including the functions that use it. The information that it
was meant to store is managed by the
`RooStats::HistFactory::Measurement` object since many years.

This commit removes some functions from the RooFit data classes:

  * `RooAbsData::valid()` (virtual method that was overridden in
     RooDataHist but not RooDataSet)

  * `RooDataHist::valid(std::size_t i)` and `RooDataHist::valid()`

  * `RooDataHist::cacheValidEntries()`

The `cacheValidEntries` method was originally intended to be used in
`RooAbsOptTestStatistic` to mask histogram entries out of the variable
range in case of a subrange fit. The reasons why `cacheValidEntries` and
the related `valid()` methods should be removed are:

  1. It is redundant. In a subrange fit, the `RooAbsOptTestStatistic` is
     creating a clone of the dataset with the subrange only using
     `RooAbsData::reduce()`. So all entries are valid by definition.

  2. RooDataHist and RooDataSet have inconistent implementations. For
     the RooDataHist, `valid()` tells you "if bin `i` is considered
     valid within the current range definitions of all observables"
     (according to the documentation). For the RooDataSet, it always
     returns `true`. This inconsistency leaves plenty of room for error,
     and means that a subrange selection relying on
     `RooAbsData::vaild()` would be broken anyway because it wouldn't
     work with RooDataSet.

  3. The masking of out-of-range entries unnessecarily increases the
     `mutable` state of the RooDataHist, which can cause trouble if one
     updates the observable range but then forgets to call
     `cacheValidEntries()`.

  4. Even the documentation said that `RooDataHist::cacheValidEntries()`
     was a "shortcut function only for RooAbsOptTestStatistic". Why keep
     it if `RooAbsOptTestStatistic` doesn't even use it in a meaningful
     way anymore?

* Simple TGraph ctor

* More doc

* add step parameter

In a binned likelihood fit, it is possible to skip the PDF normalization
when the unnormalized binned PDF can be interpreted directly in terms of
event yields. This is now done by default for models HistFactory models,
which results in great speedups for binned fits with many channels. Some
RooFit users like ATLAS were already using this for a long time, so this
feature is battle-tested.

To disable this optimization when using the `hist2workspace` executable,
add the `-disable_binned_fit_optimization` command line argument.
Directly in C++, you can also set the `binnedFitOptimization` to `false`
in the HistFactory configuration as follows:
```C++
RooStats::HistFactory::MakeModelAndMeasurementFast(measurement, {.binnedFitOptimization=false});
```
If your compiler doesn't support aggregate initialization with designators, you
need to create and edit the configuration struct explicitely:
```C++
RooStats::HistFactory::HistoToWorkspaceFactoryFast::Configuration hfCfg;
hfCfg.binnedFitOptimization = false;
RooStats::HistFactory::MakeModelAndMeasurementFast(measurement, hfCfg);
```

The `RooAbsMinimizerFcn` and `RooMinimizerFcn` classes are
implementation details of the `RooMinimizer`.

In one of the recent RooFit meetings, it was decided together with users
that this should be no problem, and doing so will greatly increase the
developers flexibility in updating the RooMinimizer implementation.

This removal is mentioned in the v628 release notes.

The `RooAbsData::createHistogram` has two overloads, one taking RooFit
command arguments and one taking integers for bin numbers (the version
with command arguments is much more common in RooFit code).

They behave inconsistently when using default parameters, e.g. these two
calls result in different binnings:
```C++
// 1. overload with int bin numbers:
data->createHistogram(obs->GetName());

// 2. overload with command arguments:
data->createHistogram(data->GetName(), dataObs);
```

The function with integer bin numbers defaults to computing a binning
automatically from the distribution in the dataset. The function with
command arguments defaults to the binning stored in the variable, which
is much more suitable for RooFit.

To avoid confusion in the future, this commit marks the overload with
the `int` parameters as deprecated so it is removed in ROOT 6.30.

Closes https://github.com/root-project/root/issues/10401.

* New class TGraph2DAsymmErrors

* Update release notes

* fix typo

The original `HistoToWorkspaceFactory` produced models that consisted of
a  Poisson term for each bin.  In this "number counting form" the
dataset has one row and the collumns corresponded to the number of
events for each bin. This led to severe performance problems in
statistical tools that generated pseudo-experiments and evaluated
likelihood ratio test statistics.

Nowadays, everyone uses the faster `HistoToWorkspaceFactoryFast`
implementation that  produces a model in the "standard form" where the
dataset has one row for each event, and the column corresponds to the
value of the observable in the histogram.

Therefore, the original `HistoToWorkspaceFactory` is now removed to
avoid  confusion and maintainance burden.

The string-based fit options were deprecated for a long time (and
officially in ROOT v6.26), and now is a good time to remove the support
for this completely.

This will give us more flexibility when refactoring the RooMinimizer to
integrate new features like the batch mode, the new likelihood classes,
or automatic differentiation.

The string-based fit options are officially deprecated with this commit,
with the intention to finally remove them in the next release (ROOT
v6.28).

This will give us more flexibility when refactoring the RooMinimizer to
integrate new features like the batch mode, the new likelihood classes,
or automatic differentiation.