X-Component MAC-UI Suite for NET v2007.10-rG serial key or number

Table of Contents

Christophe Geuzaine and Jean-François Remacle

Gmsh is an automatic 3D finite element mesh generator with build-in pre- and post-processing facilities. This is the Gmsh Reference Manual for Gmsh 4.6.0 (June 22, 2020).

Obtaining Gmsh

The source code and various pre-compiled versions of Gmsh (for Windows, Mac and Unix) can be downloaded from https://gmsh.info. Gmsh is also directly available in pre-packaged form in various Linux and BSD distributions (Debian, Ubuntu, FreeBSD, ...).

If you use Gmsh, we would appreciate that you mention it in your work by citing the following paper: “C. Geuzaine and J.-F. Remacle, Gmsh: a three-dimensional finite element mesh generator with built-in pre- and post-processing facilities. International Journal for Numerical Methods in Engineering, Volume 79, Issue 11, pages 1309-1331, 2009”. A preprint of that paper as well as other references and the latest news about Gmsh development are available on https://gmsh.info.

Copying conditions

Gmsh is “free software”; this means that everyone is free to use it and to redistribute it on a free basis. Gmsh is not in the public domain; it is copyrighted and there are restrictions on its distribution, but these restrictions are designed to permit everything that a good cooperating citizen would want to do. What is not allowed is to try to prevent others from further sharing any version of Gmsh that they might get from you.

Specifically, we want to make sure that you have the right to give away copies of Gmsh, that you receive source code or else can get it if you want it, that you can change Gmsh or use pieces of Gmsh in new free programs, and that you know you can do these things.

To make sure that everyone has such rights, we have to forbid you to deprive anyone else of these rights. For example, if you distribute copies of Gmsh, you must give the recipients all the rights that you have. You must make sure that they, too, receive or can get the source code. And you must tell them their rights.

Also, for our own protection, we must make certain that everyone finds out that there is no warranty for Gmsh. If Gmsh is modified by someone else and passed on, we want their recipients to know that what they have is not what we distributed, so that any problems introduced by others will not reflect on our reputation.

The precise conditions of the license for Gmsh are found in the General Public License that accompanies the source code (see License). Further information about this license is available from the GNU Project webpage https://www.gnu.org/copyleft/gpl-faq.html. Detailed copyright information can be found in Copyright and credits.

If you want to integrate parts of Gmsh into a closed-source software, or want to sell a modified closed-source version of Gmsh, you will need to obtain a different license. Please contact us directly for more information.

1 Overview

Gmsh is a three-dimensional finite element mesh generator with a build-in CAD engine and post-processor. Its design goal is to provide a fast, light and user-friendly meshing tool with parametric input and advanced visualization capabilities.

Gmsh is built around four modules: geometry, mesh, solver and post-processing. All geometrical, mesh, solver and post-processing instructions are prescribed either interactively using the graphical user interface (GUI) or in text files using Gmsh’s own scripting language. Interactive actions generate language bits in the input files, and vice versa. A programming API is also available, for integrating Gmsh in your own C++, C, Python or Julia code: see Gmsh API. A brief description of the four modules is given hereafter.

1.1 Geometry: model entity creation

A model in Gmsh is defined using its Boundary Representation (BRep): a volume is bounded by a set of surfaces, a surface is bounded by a series of curves, and a curve is bounded by two end points. Model entities are topological entities, i.e., they only deal with adjacencies in the model, and are implemented as a set of abstract topological classes. This BRep is extended by the definition of embedded, or internal, model entities: internal points, edges and surfaces can be embedded in volumes; and internal points and curves can be embedded in surfaces.

The geometry of model entities can be provided by different CAD kernels. The two default kernels interfaced by Gmsh are the “Built-in” kernel and the “OpenCASCADE” kernel. Gmsh does not translate the geometrical representation from one kernel to another, or from these kernels to some neutral representation. Instead, Gmsh directly queries the native data for each CAD kernel, which avoids data loss and is crucial for complex models where translations invariably introduce issues linked to slightly different representations.

Gmsh’s scripting language and the Gmsh API allow to parametrize all model entities. The entities can either be built in a “bottom-up” manner (first points, then curves, surfaces and volumes) or in a “Constructive Solid Geometry” fashion (solids on which boolean operations are performed). Both methodologies can also be combined. Finally, groups of model entities (called “physical groups”) can be defined, based on the elementary geometric entities.

1.2 Mesh: finite element mesh generation

A finite element mesh of a model is a tessellation of its geometry by simple geometrical elements of various shapes (in Gmsh: lines, triangles, quadrangles, tetrahedra, prisms, hexahedra and pyramids), arranged in such a way that if two of them intersect, they do so along a face, an edge or a node, and never otherwise. This defines a so-called “conformal” mesh. Gmsh implements several algorithms to generate such meshes automatically. All the meshes produced by Gmsh are considered as “unstructured”, even if they were generated in a “structured” way (e.g., by extrusion). This implies that the mesh elements are completely defined simply by an ordered list of their nodes, and that no predefined ordering relation is assumed between any two elements.

In order to guarantee the conformity of the mesh, mesh generation is performed in a bottom-up flow: curves are discretized first; the mesh of the curves is then used to mesh the surfaces; then the mesh of the surfaces is used to mesh the volumes. In this process, the mesh of an entity is only constrained by the mesh of its boundary, unless entities of lower dimensions are explicitly embedded in entities of higher dimension. For example, in three dimensions, the triangles discretizing a surface will be forced to be faces of tetrahedra in the final 3D mesh only if the surface is part of the boundary of a volume, or if that surface has been explicitly embedded in the volume. This automatically ensures the conformity of the mesh when, for example, two volumes share a common surface. Every meshing step is constrained by a “size field” (also called “characteristic length field”), which prescribes the desired size of the elements in the mesh. This size field can be uniform, specified by values associated with points in the geometry, or defined by general “fields” (for example related to the distance to some boundary, to a arbitrary scalar field defined on another mesh, etc.): see Specifying mesh element sizes. For each meshing step, all structured mesh directives are executed first, and serve as additional constraints for the unstructured parts.

1.3 Solver: external solver interface

Gmsh implements a ONELAB (http://onelab.info) server to pilot external solvers (called “clients”). The ONELAB interface allows to call such clients and have them share parameters and modeling information. The implementation is based on a client-server model, with a server-side database and local or remote clients communicating in-memory or through TCP/IP sockets. Contrary to most solver interfaces, the ONELAB server has no a priori knowledge about any specifics (input file format, syntax, ...) of the clients. This is made possible by having any simulation preceded by an analysis phase, during which the clients are asked to upload their parameter set to the server. The issues of completeness and consistency of the parameter sets are completely dealt with on the client side: the role of ONELAB is limited to data centralization, modification and re-dispatching.

Examples on how to interface solvers are available in the source distribution (see utils/solvers). A full-featured solver interfaced in this manner is GetDP (https://getdp.info), a general finite elements solver using mixed finite elements.

Using the Gmsh API, Gmsh can also be embedded directly in your own solver, and ONELAB parameters can be used to interactively drive it. Examples on how to embed Gmsh in your solver, and build a custom graphical user interface to control it, are available in demos/api. See in particular custom_gui.py and custom_gui.cpp.

1.4 Post-processing: scalar, vector and tensor field visualization

Gmsh can load and manipulate multiple post-processing scalar, vector or tensor fields along with the geometry and the mesh. Such fields, together with visualization options, are called “post-processing views” (or simply “views”). Scalar views can be represented by iso-curves, iso-surfaces or color maps, while vector views can be represented by three-dimensional arrows or displacement maps. Post-processing functions include section computation, offset, elevation, boundary and component extraction, color map and range modification, animation, vector graphic output, etc. All the post-processing options can be accessed either interactively, through the input script files or through the API. Various operations on the post-processing data can also be performed through plugins (see Post-processing plugins).

1.5 What Gmsh is pretty good at …

Here is a tentative list of what Gmsh does best:

• quickly describe simple and/or “repetitive” geometries with the built-in scripting language, thanks to user-defined macros, loops, conditionals and includes (see User-defined macros, Loops and conditionals, and General commands). For more advanced geometries, using the Gmsh API (see Gmsh API) in the language of your choice (C++, C, Python or Julia) brings even greater flexibility, the only downside being that you need to either compile your code (for C++ and C) or to configure and install an interpreter (Python or Julia) in addition to Gmsh. A binary Software Development Kit (SDK) is distributed on the Gmsh web site to make the process easier;
• parametrize these geometries. Gmsh’s scripting language or the Gmsh API enable all commands and command arguments to depend on previous calculations (see Expressions, Geometry commands, and Gmsh API). Using the OpenCASCADE geometry kernel, Gmsh gives access to all usual constructive solid geometry operations;
• import geometries from other CAD software in standard exchange formats. Gmsh uses OpenCASCADE to import such files, including label and color information from STEP and IGES files;
• generate 1D, 2D and 3D simplicial (i.e., using line segments, triangles and tetrahedra) finite element meshes (see Mesh module), with fine control over the element size (see Specifying mesh element sizes);
• create simple extruded geometries and meshes (see Geometry commands, and Mesh commands), and allow to automatically couple such structured meshes with unstructured ones (using a layer of pyramids in 3D);
• generate high-order (curved) meshes that conform to the CAD model geometry. High-order mesh optimization tools allow to guarantee the validity of such curved meshes;
• interact with external solvers by defining ONELAB parameters, shared between Gmsh and the solvers and easily modifiable in the GUI (see Solver module);
• visualize and export computational results in a great variety of ways. Gmsh can display scalar, vector and tensor datasets, perform various operations on the resulting post-processing views (see Post-processing module), can export plots in many different formats (see General options list), and can generate complex animations (see General tools, and t8);
• run on low end machines and/or machines with no graphical interface. Gmsh can be compiled with or without the GUI (see Compiling the source code), and all versions can be used either interactively or directly from the command line (see Running Gmsh on your system);
• configure your preferred options. Gmsh has a large number of configuration options that can be set interactively using the GUI, scattered inside script files, changed through the API, set in per-user configuration files and specified on the command-line (see Running Gmsh on your system and Options);
• and do all the above on various platforms (Windows, Mac and Unix), for free (see Copying conditions)!

1.6 … and what Gmsh is not so good at

Here are some known weaknesses of Gmsh:

• Gmsh is not a multi-bloc mesh generator: all meshes produced by Gmsh are conforming in the sense of finite element meshes;
• Gmsh’s user interface is only exposing a limited number of the available features, and many aspects of the interface could be enhanced (especially manipulators).
• Your complaints about Gmsh here :-)

If you have the skills and some free time, feel free to join the project: we gladly accept any code contributions (see Information for developers) to remedy the aforementioned (and all other) shortcomings!

1.7 Bug reports

Please file issues on https://gitlab.onelab.info/gmsh/gmsh/issues. Provide as precise a description of the problem as you can, including sample input files that produce the bug. Don’t forget to mention both the version of Gmsh and the version of your operation system (see Command-line options to see how to get this information).

See Frequently asked questions, and the bug tracking system to see which problems we already know about.

2 How to read this reference manual?

Gmsh can be used at three levels:

1. as a stand-alone application manipulated through its graphical user interface (GUI);
2. as a stand-alone script-driven application;
3. as a library.

You can skip most of this reference manual if you only want to use Gmsh at the first level (i.e., interactively with the GUI). Just read the next chapter (see Running Gmsh on your system) to learn how to launch Gmsh on your system, then go experiment with the GUI and the tutorial files (see Tutorial) provided in the distribution. Screencasts that show how to use the GUI are available here: https://gmsh.info/screencasts/.

The aim of the reference manual is to explain everything you need to use Gmsh at the second level, i.e., using the built-in scripting language. A Gmsh script file is an ASCII text file that contains instructions in Gmsh’s built-in scripting language. Such a file is interpreted by Gmsh’s parser, and can be given any extension (or no extension at all). By convention, Gmsh uses the extension for geometry scripts, and the extension for parsed post-processing datasets. Once you master the tutorial (read the source files: they are heavily commented!), start reading chapter General tools, then proceed with the next four chapters, which detail the syntax of the geometry, mesh, solver and post-processing scripting commands. You will see that most of the interactive actions in the GUI have a direct equivalent in the scripting language. If you want to use Gmsh as a pre- or post-processor for your own software, you will also want to learn about the non-scripting input/output files that Gmsh can read/write. In addition to Gmsh’s native “MSH” file format (see File formats), Gmsh can read/write many standard mesh files, depending on how it was built: check the ‘File->Export’ menu for a list of available formats.

Finally, to use Gmsh at the third level (i.e., to link the Gmsh library with your own code), you will need to learn the Gmsh Application Programming Interface (API). This API is available in C++, C, Python and Julia, and is fully documented in Gmsh API.

2.1 Syntactic rules used in the manual

Here are the rules we tried to follow when writing this reference manual. Note that metasyntactic variable definitions stay valid throughout the manual (and not only in the sections where the definitions appear).

1. Keywords and literal symbols are printed like .
2. Metasyntactic variables (i.e., text bits that are not part of the syntax, but stand for other text bits) are printed like .
3. A colon () after a metasyntactic variable separates the variable from its definition.
4. Optional rules are enclosed in pairs.
5. Multiple choices are separated by .
6. Three dots (…) indicate a possible (multiple) repetition of the preceding rule.

3 Running Gmsh on your system

3.1 Interactive mode

To launch Gmsh in interactive mode, just double-click on the Gmsh icon, or type

at your shell prompt in a terminal. This will open the main Gmsh window, with a tree-like menu on the left, a graphic area on the right, and a status bar at the bottom. (You can detach the tree menu using ‘Window->Attach/Detach Menu’.)

To open the first tutorial file (see Tutorial), select the ‘File->Open’ menu, and choose t1.geo. When using a terminal, you can specify the file name directly on the command line, i.e.:

To perform the mesh generation, go to the mesh module (by selecting ‘Mesh’ in the tree) and choose the dimension (‘1D’ will mesh all the curves; ‘2D’ will mesh all the surfaces—as well as all the curves if ‘1D’ was not called before; ‘3D’ will mesh all the volumes—and all the surfaces if ‘2D’ was not called before). To save the resulting mesh in the current mesh format click on ‘Save’, or select the appropriate format and file name with the ‘File->Export’ menu. The default mesh file name is based on the name of the current active model, with an appended extension depending on the mesh format¹.

To create a new geometry or to modify an existing geometry, select ’Geometry’ in the tree. For example, to create a spline, select ‘Elementary entities’, ‘Add’, ‘New’ and ‘Spline’. You will then be asked to select a list of points, and to type to finish the selection (or to abort it). Once the interactive command is completed, a text string is automatically added at the end of the current script file. You can edit the script file by hand at any time by pressing the ‘Edit’ button in the ‘Geometry’ menu and then reloading the model by pressing ‘Reload’. For example, it is often faster to define variables and points directly in the script file, and then use the GUI to define the curves, the surfaces and the volumes interactively.

Several files can be loaded simultaneously in Gmsh. When specified on the command line, the first one defines the active model and the others are ‘merged’ into this model. You can merge such files with the ‘File->Merge’ menu. For example, to merge the post-processing views contained in the files view1.pos and view5.msh together with the geometry of the first tutorial t1, you can type the following command:

In the Post-Processing module (select ‘Post-Processing’ in the tree), three items will appear, respectively labeled ‘A scalar map’, ‘Nodal scalar map’ and ‘Element 1 vector’. In this example the views contain several time steps: you can loop through them with the small “remote-control” icons in the status bar. A mouse click on the view name will toggle the visibility of the selected view, while a click on the arrow button on the right will provide access to the view’s options.

Note that all the options specified interactively can also be directly specified in the script files. You can save the current options of the current active model with the ‘File->Save Model Options’. This will create a new option file with the same filename as the active model, but with an extra extension added. The next time you open this model, the associated options will be automatically loaded, too. To save the current options as your default preferences for all future Gmsh sessions, use the ‘File->Save Options As Default’ menu instead. Finally, you can also save the current options in an arbitrary file by choosing the ‘Gmsh options’ format in ‘File->Export’.

For more information about available options (and how to reset them to their default values), see Options. A full list of options with their current values is also available in the ‘Help->Current Options’ menu.

3.2 Non-interactive mode

Gmsh can be run non-interactively in ‘batch’ mode, without GUI². For example, to mesh the first tutorial in batch mode, just type:

To mesh the same example, but with the background mesh available in the file bgmesh.pos, type:

For the list of all command-line options, see Command-line options. In particular, any complicated workflow can be written in a file, and this file can be executed as a script using

The script can contain e.g. meshing commands, like .

3.3 Command-line options

(Related option names, if any, are given between parentheses)

Geometry:

Output model, then exit

Set geometrical tolerance (Geometry.Tolerance)

Match geometries and meshes

Mesh:

Perform 1D, 2D or 3D mesh generation, then exit

Save mesh, then exit

Specify output file name

Select output mesh format: auto, msh1, msh2, msh22, msh3, msh4, msh40, msh41, msh, unv, vtk, wrl, mail, stl, p3d, mesh, bdf, cgns, med, diff, ir3, inp, ply2, celum, su2, x3d, dat, neu, m, key (Mesh.Format)

Create binary files when possible (Mesh.Binary)

Perform uniform mesh refinement, then exit

Perform barycentric mesh refinement, then exit

Reclassify surface mesh, then exit

Reparametrize surface mesh, then exit

Partition after batch mesh generation (Mesh.NbPartitions)

Weight of a triangle/quad/etc. during partitioning (Mesh.Partition[Tri,Quad,...]Weight)

Save mesh partitions in separate files (Mesh.PartitionSplitMeshFiles)

Create the partition topology (Mesh.PartitionCreateTopology)

Create ghost cells (Mesh.PartitionCreateGhostCells)

Create physical groups for partitions (Mesh.PartitionCreatePhysicals)

Save the partition topology .pro file (Mesh.PartitionTopologyFile)

Preserve element numbering in MSH2 format (Mesh.PreserveNumberingMsh2)

Save all elements (Mesh.SaveAll)

Save nodes with their parametric coordinates (Mesh.SaveParametric)

Save model topology (Mesh.SaveTopology)

Select mesh algorithm: auto, meshadapt, del2d, front2d, delquad, pack, initial2d, del3d, front3d, mmg3d, hxt, initial3d (Mesh.Algorithm and Mesh.Algorithm3D)

Set number of mesh smoothing steps (Mesh.Smoothing)

Set mesh order (Mesh.ElementOrder)

Optimize quality of tetrahedral elements (Mesh.Optimize[Netgen])

Optimize tetrahedral elements that have a quality less than a threshold (Mesh.OptimizeThreshold)

Optimize high order meshes (Mesh.HighOrderOptimize)

High-order optimization parameters (Mesh.HighOrderThreshold[Min,Max], Mesh.HighOrderNumLayers)

Set mesh element size factor (Mesh.CharacteristicLengthFactor)

Set minimum mesh element size (Mesh.CharacteristicLengthMin)

Set maximum mesh element size (Mesh.CharacteristicLengthMax)

Compute mesh element size from curvature, with given minimum number of elements per 2*pi radians (Mesh.CharacteristicLengthFromCurvature and Mesh.MinimumElementsPerTwoPi)

Set maximum anisotropy for bamg (Mesh.AnisoMax)

Set smoothing ration between mesh sizes at nodes of a same edge for bamg (Mesh.SmoothRatio)

Set accuracy of evaluation of mesh size field for 1D mesh (Mesh.LcIntegrationPrecision)

Set the threshold angle (in degrees) between two adjacent faces below which a swap is allowed (Mesh.AllowSwapAngle)

Set random perturbation factor (Mesh.RandomFactor)

Load background mesh from file

Perform various consistency checks on mesh

Ignore periodic boundaries (Mesh.IgnorePeriodicity)

Post-processing:

Select link mode between views (PostProcessing.Link)

Combine views having identical names into multi-time-step views

Solver:

Always listen to incoming connection requests (Solver.AlwaysListen)

Name of Octave interpreter (Solver.OctaveInterpreter)

Name of Python interpreter (Solver.OctaveInterpreter)

Run ONELAB solver(s)

Display:

Hide all meshes and post-processing views on startup (View.Visible, Mesh.[Points,Lines,SurfaceEdges,...])

Disable double buffering (General.DoubleBuffer)

Set num of subdivisions for high order element display (Mesh.NumSubEdges)

Specify the font size for the GUI (General.FontSize)

Specify FLTK GUI theme (General.FltkTheme)

Specify display (General.Display)

Use camera mode view (General.CameraMode)

OpenGL quad-buffered stereo rendering (General.Stereo)

Use gamepad controller if available

Other:

Parse input files, then exit

Create new model before merge next file

Merge next files

Open next files

Log all messages to filename

Start in automatic, geometry, mesh, solver or post-processing mode (General.InitialModule)

Print process id on stdout

Pattern of files to merge as they become available (General.WatchFilePattern)

Load background (image or PDF) file (General.BackgroundImageFileName)

Set verbosity level (General.Verbosity)

Don’t popup dialog windows in scripts (General.NoPopup)

Parse command string at startup

Set constant or option number name=value

Set constant or option string name=value

Parse option file at startup

Convert files into latest binary formats, then exit

Set number of threads (General.NumThreads)

Report CPU times for all operations

Show version number

Show detailed version information

Show command line usage

Show all options

3.4 Mouse actions

Highlight the entity under the mouse pointer and display its properties / Resize a lasso zoom or a lasso (un)selection

Rotate / Select an entity / Accept a lasso zoom or a lasso selection

Start a lasso zoom or a lasso (un)selection

Zoom / Unselect an entity / Accept a lasso zoom or a lasso unselection

Orthogonalize display

Pan / Cancel a lasso zoom or a lasso (un)selection / Pop-up menu on post-processing view button

Reset to default viewpoint

For a 2 button mouse, Middle button = Shift+Left button.

For a 1 button mouse, Middle button = Shift+Left button, Right button = Alt+Left button.

3.5 Keyboard shortcuts

(On Mac Ctrl is replaced by Cmd (the ‘Apple key’) in the shortcuts below.)

Go to previous time step

Go to next time step

Make previous view visible

Make next view visible

Reload geometry

Reload full project

Mesh lines

Mesh surfaces

Mesh volumes

Cancel lasso zoom/selection, toggle mouse selection ON/OFF

End/accept selection in geometry creation mode

Go to geometry module

Go to mesh module

Go to post-processing module

Abort selection in geometry creation mode

Go to solver module

Toogle x coordinate freeze in geometry creation mode

Toogle y coordinate freeze in geometry creation mode

Toogle z coordinate freeze in geometry creation mode

Bring all windows to front

Show geometry options

Show mesh options

Show general options

Show post-processing options

Show solver options

Show post-processing view plugins

Show post-processing view options

Move only along x coordinate in geometry creation mode

Move only along y coordinate in geometry creation mode

Move only along z coordinate in geometry creation mode

Enable full mouse selection

Attach/detach menu

Export project

Enter full screen

Show statistics window

Save model options

Show message console

Minimize window

Create new project file

Open project file

Quit

Rename project file

Save mesh in default format

Show clipping plane window

Show current options and workspace window

Save options as default

Show manipulator window

Show option window

Merge file(s)

Show plugin window

Show visibility window

Loop through axes modes

Hide/show bounding boxes

Loop through predefined color schemes

Hide/Show element outlines for visible post-pro views

Change redraw mode (fast/full)

Hide/show all post-processing views

Hide/show all post-processing view scales

Hide/show geometry lines

Toggle visibility of all mesh entities

Hide/show all post-processing view annotations

Change projection mode (orthographic/perspective)

Hide/show geometry points

Loop through range modes for visible post-pro views

Hide/show geometry surfaces

Loop through interval modes for visible post-pro views

Hide/show geometry volumes

Enable/disable all lighting

Set X view

Set Y view

Set Z view

Hide/show small axes

Hide/show mesh volume faces

Loop through predefined colormaps

Hide/show mesh surface faces

Hide/show mesh lines

Hide/show mesh points

Hide/show mesh surface edges

Same as Alt+t, but with numeric mode included

Hide/show mesh volume edges

Set -X view

Set -Y view

Set -Z view

4 General tools

This chapter describes the general commands and options that can be used in Gmsh’s script files. By “general”, we mean “not specifically related to one of the geometry, mesh, solver or post-processing modules”. Commands peculiar to these modules will be introduced in Geometry module, Mesh module, Solver module, and Post-processing module, respectively.

If you plan to use Gmsh through its API (see Gmsh API) instead of the built-in scripting language, you can skip this chapter entirely.

4.1 Comments

Gmsh script files support both C and C++ style comments:

1. any text comprised between and pairs is ignored;
2. the rest of a line after a double slash is ignored.

These commands won’t have the described effects inside double quotes or inside keywords. Also note that ‘white space’ (spaces, tabs, new line characters) is ignored inside all expressions.

4.2 Expressions

The two constant types used in Gmsh scripts are and (there is no integer type). These types have the same meaning and syntax as in the C or C++ programming languages.

4.2.1 Floating point expressions

Floating point expressions (or, more simply, “expressions”) are denoted by the metasyntactic variable (remember the definition of the syntactic rules in Syntactic rules), and are evaluated during the parsing of the script file:

Such s are used in most of Gmsh’s scripting commands. When is appended to a string , the result is a new string formed by the concatenation of , (an underscore) and the value of the . This is most useful in loops (see Loops and conditionals), where it permits to define unique strings automatically. For example,

is the same as

The brackets permit to extract one item from a list (parentheses can also be used instead of brackets). The permits to get the size of a list. The operators , , , and are defined in Operators. For the definition of s, see Built-in functions. The various s are listed in Options. searches for occurrences of the first expression in the second (both of which can be lists). searches the first for any occurrence of the second . compares the two strings (returns an integer greater than, equal to, or less than 0, according as the first string is greater than, equal to, or less than the second string). returns the length of the string. creates attributes for text strings. checks if a variable with the given name exists (i.e., has been defined previously), and checks if the file with the given name exists. creates a name from the provided string. allows to get the value of a ONELAB variable (the optional second argument is the default value returned if the variable does not exist). allows to ask the user for a value interactively (the second argument is the value returned in non-interactive mode). For example, inserting in an input file will query the user for the value of a certain parameter alpha, assuming the default value is 5.76. If the option is set (see General options list), no question is asked and the default value is automatically used.

allows to define a ONELAB variable in-line. The given as the first argument is the default value; this is followed by the various ONELAB options. See the ONELAB tutorial wiki for more information.

List of expressions are also widely used, and are defined as:

with

The second case in this last definition permits to create a list containing the range of numbers comprised between two s, with a unit incrementation step. The third case also permits to create a list containing the range of numbers comprised between two s, but with a positive or negative incrementation step equal to the third . The fourth, fifth and sixth cases permit to reference an expression list (parentheses can also be used instead of brackets). sorts the entries in the list and removes all duplicates. takes the absolute value of all entries in the list. reads a list of numbers from a file. and construct lists using linear or logarithmic spacing. The next two cases permit to reference an expression sublist (whose elements are those corresponding to the indices provided by the ). The next cases permit to retrieve the indices of entities created through geometrical transformations, extrusions and boolean operations (see Transformations, Extrusions and Boolean operations).

The next two cases allow to retrieve entities in a given bounding box, or get the bounding box of a given entity. The last five cases permit to retrieve the mass or the center of mass of an entity, the coordinates of a given geometry point (see Points), the elementary entities making up physical groups, and the tags of all (physical or elementary) points, curves, surfaces or volumes in the model. These operations all trigger a synchronization of the CAD model with the internal Gmsh model.

To see the practical use of such expressions, have a look at the first couple of examples in Tutorial. Note that, in order to lighten the syntax, you can omit the braces enclosing an if this only contains a single item. Also note that a braced can be preceded by a minus sign in order to change the sign of all the s.

For some commands it makes sense to specify all the possible expressions in a list. This is achieved with , defined as:

The meaning of “all” () depends on context. For example, will get the ids of all the existing curves in the model, while will get the ids of all existing surfaces.

4.2.2 Character expressions

Character expressions are defined as:

returns the current date. returns the current ONELAB action (e.g. or ). returns the full path of the Gmsh executable. (or ) and return the directory and file name of the script being parsed. and take the prefix (e.g. to remove the extension) or the relative path of a given file name. and concatenate character expressions ( adds a newline character after each string except the last). returns the first or second depending on the value of . returns the portion of the string that starts at the character position given by the first and spans the number of characters given by the second or until the end of the string (whichever comes first; or always if the second is not provided). converts the to upper case. returns the absolute path of a file. returns the directory of a file. is equivalent to the C function (where is a format string that can contain floating point formatting characters: , , etc.) The various s are listed in Options. gets the value of an environment variable from the operating system. allows to get a ONELAB string value (the second optional argument is the default value returned if the variable does not exist). asks the user for a value interactively (the second argument is the value used in non-interactive mode). ’s arguments are: input string, old substring, new substring (brackets can be used instead of parentheses in and ). , etc., or , etc., retrieve the name of the physical or elementary entity, if any. converts a variable name into a string.

allows to define a ONELAB variable in-line. The given as the first argument is the default value; this is followed by the various ONELAB options. See the ONELAB tutorial wiki for more information.

Character expressions are mostly used to specify non-numeric options and input/output file names. See t8, for an interesting usage of s in an animation script.

List of character expressions are defined as:

4.2.3 Color expressions

Colors expressions are hybrids between fixed-length braced s and s:

The first case permits to use the X Windows names to refer to colors, e.g., , , , … (see Common/Colors.h in the source code for a complete list). The second case permits to define colors by using three expressions to specify their red, green and blue components (with values comprised between 0 and 255). The third case permits to define colors by using their red, green and blue color components as well as their alpha channel. The last case permits to use the value of a as a . The various s are listed in Options.

See t3, for an example of the use of color expressions.

Safari is a graphicalweb browser developed by Apple, based on the WebKit engine. First released on desktop in 2003 with Mac OS X Panther, a mobile version has been bundled with iOS devices since the iPhone's introduction in 2007. Safari is the default browser on Apple devices. A Windows version was available from 2007 to 2012.^[6]

History and development[edit]

Until 1997, Apple's Macintosh computers shipped with the Netscape Navigator and Cyberdog web browsers only. Internet Explorer for Mac was later included as the default web browser for Mac OS 8.1 and later,^[7] as part of a five-year agreement between Apple and Microsoft. During that time, Microsoft released three major versions of Internet Explorer for Mac that were bundled with Mac OS 8 and Mac OS 9, though Apple continued to include Netscape Navigator as an alternative. Microsoft ultimately released a Mac OS X edition of Internet Explorer for Mac, which was included as the default browser in all Mac OS X releases from Mac OS X DP4^[8] up to and including Mac OS X v10.2.^[9]

Safari 1[edit]

On January 7, 2003, at Macworld San Francisco, Steve Jobs announced that Apple had developed its own web browser, called Safari. It was based on Apple's internal fork of the KHTMLrendering engine, called WebKit.^[10] The company released the first beta version, available only for Mac OS X, later that day. A number of official and unofficial beta versions followed, up until version 1.0 was released on June 23, 2003.^[11] Initially only available as a separate download for Mac OS X 10.2, Safari was bundled with Mac OS X v10.3 on October 24, 2003, as the default browser, with Internet Explorer for Mac included only as an alternative browser. Version 1.0.3, released on August 13, 2004, was the last version to support Mac OS X 10.2, while 1.3.2, released on January 12, 2006, was the last version to support Mac OS X 10.3. However, 10.3 received security updates through 2007.

Safari 2[edit]

In April 2005, Dave Hyatt, one of the Safari developers at Apple, documented his study by fixing specific bugs in Safari, thereby enabling it to pass the Acid2 test developed by the Web Standards Project. On April 27, 2005, he announced that his development version of Safari now passed the test, making it the first web browser to do so.^[12]

Safari 2.0 was released on April 29, 2005, as the only web browser included with Mac OS X 10.4. This version was touted by Apple as possessing a 1.8x speed boost over version 1.2.4, but did not yet include the Acid2 bug fixes. The necessary changes were initially unavailable to end-users unless they downloaded and compiled the WebKit source code themselves or ran one of the nightly automated builds available at OpenDarwin.org.^[13] Apple eventually released version 2.0.2 of Safari, which included the modifications required to pass Acid2, on October 31, 2005.

In June 2005, after some criticism from KHTML developers over lack of access to change logs, Apple moved the development source code and bug tracking of WebCore and JavaScriptCore to OpenDarwin.org. WebKit itself was also released as open source. The source code for non-renderer aspects of the browser, such as its GUI elements, remains proprietary.

The final stable version of Safari 2, Safari 2.0.4, was released on January 10, 2006, for Mac OS X. It was only available as part of Mac OS X Update 10.4.4. This version addressed layout and CPU usage issues, among other improvements.^[14] Safari 2.0.4 was the last version to be released exclusively on Mac OS X.

Safari 3[edit]

On January 9, 2007, at Macworld SF, Jobs announced the iPhone. The device's operating system (later called iPhone OS and subsequently renamed to iOS) used a mobile version of the Safari browser and was able to display full, desktop-class websites.^[15]

On June 11, 2007, at the Apple Worldwide Developers Conference, Jobs announced Safari 3 for Mac OS X 10.5, Windows XP, and Windows Vista. During the announcement, he ran a benchmark based on the iBench browser test suite comparing the most popular Windows browsers,^[16] hence claiming that Safari was the fastest browser. Later third-party tests of HTTP load times would support Apple's claim that Safari 3 was indeed the fastest browser on the Windows platform in terms of initial data loading over the Internet, though it was found to be only negligibly faster than Internet Explorer 7 and Mozilla Firefox when loading static content from local cache.^[17]

The initial Safari 3 beta version for Windows, released on the same day as its announcement at WWDC 2007, had several known bugs^[18] and a zero day exploit that allowed remote execution.^[19] The addressed bugs were then corrected by Apple three days later on June 14, 2007, in version 3.0.1 for Windows. On June 22, 2007, Apple released Safari 3.0.2 to address some bugs, performance issues and other security issues. Safari 3.0.2 for Windows handles some fonts that are missing in the browser but already installed on Windows computers, such as Tahoma, Trebuchet MS, and others.

The iPhone was formally released on June 29, 2007. It included a version of Safari based on the same WebKit rendering engine as the desktop version, but with a modified feature set better suited for a mobile device. The version number of Safari as reported in its user agent string is 3.0,^[20] in line with the contemporary desktop versions of Safari.

The first stable, non-beta release of Safari for Windows, Safari 3.1, was offered as a free download on March 18, 2008. In June 2008, Apple released version 3.1.2,^[21][22] addressing a security vulnerability in the Windows version where visiting a malicious web site could force a download of executable files and execute them on the user's desktop.^[23]

Safari 3.2, released on November 13, 2008, introduced anti-phishing features using Google Safe Browsing and Extended Validation Certificate support. The final version of Safari 3 is 3.2.3, released on May 12, 2009.

Safari 4[edit]

On June 2, 2008, the WebKit development team announced SquirrelFish,^[24] a new JavaScript engine that vastly improves Safari's speed at interpreting scripts.^[25] The engine is one of the new features in Safari 4, released to developers on June 11, 2008. The new JavaScript engine quickly evolved into SquirrelFish Extreme, featuring even further improved performance over SquirrelFish,^[26] and was eventually marketed as Nitro. A public beta of Safari 4 was released on February 24, 2009, with new features such as the Top Sites tool (similar to Opera's Speed Dial feature), which displays the user's most visited sites on a 3D wall.^[27]Cover Flow, a feature of Mac OS X and iTunes, was also implemented in Safari. In the public beta versions, tabs were placed in the title bar of the window, similar to Google Chrome. The tab bar was moved back to its original location, below the URL bar, in the final release.^[28] The Windows version adopted a native Windows theme, rather than the previously employed Mac OS X-style interface. Also, Apple removed the blue progress bar located in the address bar (later reinstated in Safari 5). Safari 4.0.1 was released for Mac on June 17, 2009, and fixed problems with Faces in iPhoto '09. Safari 4 in Mac OS X v10.6 "Snow Leopard" has 64-bit support, which can make JavaScript loading up to 50% faster. It also has built-in crash resistance unique to Snow Leopard; crash resistance will keep the browser intact if a plug-in like Flash player crashes, such that the other tabs or windows will be unaffected.^[29] Safari 4.0.4, released on November 11, 2009, for both OS X and Windows, further improves JavaScript performance.^[30]

Safari was one of the twelve browsers offered to EU users of Microsoft Windows in 2010. It was one of the five browsers displayed on the first page of browser choices along with Chrome, Firefox, Internet Explorer and Opera.^[31][32]

Safari 4 features[edit]

Beginning with Safari 4, the address bar has been completely revamped:

• The blue inline progress bar is replaced with a spinning bezel and a loading indicator attached to it.
• The button to add a bookmark is now attached to the address bar by default.
• The reload/stop button is now superimposed on the right end of the address bar.

Safari on Mac OS X and Windows was made to look more similar to Safari on iPhone than previous versions.

Safari 4 also includes the following new features:

• Completely passes the Acid3 standards test
• Cover Flow browsing for History and Bookmarks
• Improved developer tools, including Web Inspector, CSS element viewing, JavaScript debugger and profiler, offline table and database management with SQL support, and resource graphs
• Nitro JavaScript engine that executes JavaScript up to eight times faster than Internet Explorer 8 and more than four times faster than Firefox 3^[33]
• Native Windows look on Windows (Aero, Luna, Classic, etc., depending on OS and settings) with standard Windows font rendering and optional Apple font rendering
• Support for CSS image retouching effects
• Support for CSS Canvas
• Speculative loading, where Safari loads the documents, scripts, and style information that are required to view a web page ahead of time
• Support for HTML5
• Top Sites, which displays up to 24 thumbnails of a user's most frequently visited pages on startup

Safari 5[edit]

Apple released Safari 5 on June 7, 2010, featuring the new Safari Reader for reading articles on the web without distraction (based on Arc90's Readability tool^[34]), and a 30 percent JavaScript performance increase over Safari 4. Safari 5 includes improved developer tools and supports more than a dozen new HTML5 technologies, focused on interoperability. Since Safari 5, developers can create secure Safari Extensions to customize and enhance the browsing experience.^[35] Apple also re-added the progress bar behind the address bar in this release. Safari 5.0.1 enabled the Extensions PrefPane by default; previously, users had to enable it via the Debug menu.

Apple also released Safari 4.1 concurrently with Safari 5, exclusively for Mac OS X Tiger. The update included the majority of the features and security enhancements found in Safari 5. It did not, however, include Safari Reader or Safari Extensions. Together with Mac OS X 10.7 Lion, Apple released Safari 5.1 for both Windows and Mac on July 20, 2011, with the new function 'Reading List' and a faster browsing experience. Apple simultaneously released Safari 5.0.6 for Mac OS X 10.5 Leopard, excluding Leopard users from the new functions in Safari 5.1.

Safari 5.1.7 has become the last version of Safari developed for Windows.

Safari 5 features[edit]

Safari 5 includes the following new features:

• Full-text search through the browser history^[36]
• Safari Reader, which removes formatting and ads from webpages.^[37]
• Smarter address field, where the address bar autocomplete will match against titles of web page in history or bookmarks.
• Extensions, which are add-ons that customize the web browsing experience.^[38]
• Improved support for HTML5, including full screen video, closed caption, geolocation, EventSource, and a now obsolete early variant of the WebSocket protocol.
• Improved Web Inspector.
• Faster Nitro JavaScript Engine.
• DNS prefetching, where Safari finds links and looks up addresses on the web page ahead of time.
• Bing search.
• Improved graphics hardware acceleration on Windows.

Additionally, the blue inline progress bar has returned to the address bar, in addition to the spinning bezel and loading indicator introduced in Safari 4. Top Sites view now has a button to switch to Full History Search. Other features include Extension builder for developers of Safari Extensions, which are built using web standards such as HTML5, CSS3, and JavaScript.

Safari 6[edit]

Safari 6.0 was previously known as Safari 5.2 until Apple announced the change at WWDC 2012. The stable release of Safari 6 coincided with the release of OS X Mountain Lion on July 25, 2012, and is integrated into the OS.^[39] As Apple integrated it with Mountain Lion, it is no longer available for download from the Apple website or other sources. Apple released Safari 6 via Software Update for users of OS X Lion. It has not been released for OS X versions prior to Lion or for Windows. Regarding the unavailability of Safari 6 on Windows, Apple has stated "Safari 6 is available for Mountain Lion and Lion. Safari 5 continues to be available for Windows."^[40] Microsoft removed Safari from its BrowserChoice page.

On June 11, 2012, Apple released a developer preview of Safari 6.0 with a feature called iCloud Tabs, which allows users to 'sync' their open tabs with any iOS or other OS X device running the latest software. Safari 6 also included new privacy features, including an "Ask websites not to track me" preference, and the ability for websites to send OS X 10.8 Mountain Lion users notifications, although it removed RSS support.^[41] Safari 6 has the Share Sheets capability in OS X Mountain Lion. The Share Sheet options are: Add to Reading List, Add Bookmark, Email this Page, Message, Twitter and Facebook. Tabs with full-page previews were added, too.^[42]

Safari 6 features[edit]

Safari 6 introduced the following features, many of which are only available on OS X 10.8 Mountain Lion:^[43]

• Unified smart search field, which combines the web address and search fields, similar to Chrome's Omnibox and Firefox's Awesome Bar.
• Tab view (Mountain Lion only), which enables movement between tabs using multi-touch gestures.
• iCloud tabs (Mountain Lion only) synchronizes recent websites across OS X and iOS devices.
• Built-in sharing (Mountain Lion only) to email, Messages, Twitter and Facebook.
• Improved performance
• Support for -webkit-calc()

Additionally various features were removed, including, but not limited to, Activity Window, separate Download Window, direct support for RSS feeds in the URL field and bookmarks. The separate search field is also no longer available as a toolbar configuration option.

Safari 7[edit]

Announced at Apple's Worldwide Developer Conference (WWDC) on June 10, 2013, the Safari 7/6.1^[44] developer preview brought improvements in JavaScript performance and memory usage, as well as a new look for Top Sites and the Sidebar, and a new Shared Links feature. Additionally, a new Power Saver feature pauses Plugins which are not in use.^[45] Safari 7 for OS X Mavericks and Safari 6.1 (for Lion and Mountain Lion) were released along with OS X Mavericks in an Apple special event on October 22, 2013.^[46]

Safari 8[edit]

Safari 8 was announced at WWDC 2014 and released with OS X Yosemite. It included WebGL support, stronger privacy features, increased speed and efficiency, enhanced iCloud integration, and updated design.^[47][48]

Safari 8 features[edit]

Safari 8 introduced the following features, available on OS X Yosemite:^[49]

Safari 9[edit]

Safari 9 was announced at WWDC 2015 and released with OS X El Capitan. It included muting tabs and pinned tabs.

Safari 10[edit]

Safari 10 was released alongside macOS Sierra 10.12 for OS X Yosemite and OS X El Capitan.^[50] It does not include all of the new features available in macOS Sierra, like Apple Pay on the web and picture-in-picture support for videos, but the update includes the following new functions:

• Safari Extensions such as 1Password, Save to Pocket, and DuckDuckGo
• New Bookmarks sidebar, including double-click to focus in on a folder
• Redesigned Bookmarks and History views
• Site-specific zoom: Safari remembers and re-applies your zoom level to websites
• Improved AutoFill from Contacts card
• Reader improvements, including in-line sub-headlines, bylines, and publish dates
• Legacy plug-ins are turned off by default in favor of HTML5 versions of websites
• Allow reopening of recently closed tabs through the History menu, holding the "+" button in the tab bar, and using Shift-Command-T
• When a link opens in a new tab, it is now possible to hit the back button or swipe to close it and go back to the original tab
• Improved ranking of Frequently Visited Sites
• Web Inspector Timelines Tab
• Debugging using Web Inspector

Safari 10 also includes a number of security updates, including fixes for six WebKit vulnerabilities and issues related to Reader and Tabs. The first version of Safari 10 was released on September 20, 2016, and the last version (10.1.2) was released on July 19, 2017.

Safari 11[edit]

Safari 11 was released as a part of macOS High Sierra but was also made available for OS X El Capitan and macOS Sierra on September 19, 2017.^[51] Safari 11 included several new features such as Intelligent Tracking Prevention^[52] which aims to prevent cross-site tracking by placing limitations on cookies and other website data.^[53]

Safari 12[edit]

Safari 12 was released in the lead up to macOS Mojave but was also made available for macOS Sierra and macOS High Sierra on September 17, 2018. Safari 12 includes several new features such as Icons in tabs, Automatic Strong Passwords, and Intelligent Tracking Prevention 2.0.^[54] An updated Safari version 12.0.1 was released on October 30, 2018, as part of macOS Mojave 10.14.1 release, and Safari 12.0.2 was released on December 5, 2018, alongside macOS 10.14.2.

Support for developer-signed classic Safari Extensions has been dropped. This version will also be the last one that supports the official Extensions Gallery, and Apple encourages extension authors to switch to Safari App Extensions. This move triggered negative feedback in the community.^[55][56]

Safari 13[edit]

Safari 13 was announced alongside macOS Catalina at WWDC 2019 on June 3, 2019. Safari 13 includes several new features such as prompting users to change weak passwords, FIDO2 USB security key authentication support, Sign in with Apple support, Apple Pay on the Web support, and increased speed and security.^[57] Safari 13 was released on September 20, 2019, on macOS Mojave and macOS High Sierra.^[58]

Safari 14[edit]

In June 2020 it was announced that macOS Big Sur will include Safari 14.^[59] Safari 14 introduces new privacy features, including Privacy Report, which shows blocked content and privacy information on web pages. Users will also receive a monthly report on trackers that Safari has blocked. Extensions can also be enabled or disabled on a site-by-site basis.^[60] Safari 14 introduced support for the WebExtension API used in Google Chrome, Microsoft Edge, Firefox, and Opera, making it easier for developers to port their extensions from those web browsers to Safari.^[61] Support for Adobe Flash Player will also be dropped from Safari.^[62] Safari 14 was released as a standalone update to macOS Catalina and Mojave users on September 16, 2020.^[63]

Safari Technology Preview[edit]

Safari Technology Preview was first released alongside OS X El Capitan 10.11.4. Safari Technology Preview releases include the latest version of WebKit, incorporating Web technologies to be incorporated in future stable releases of Safari, so that developers and users can install the Technology Preview release on a Mac, test those features, and provide feedback.^[64]

Other features[edit]

On macOS, Safari is a Cocoa application.^[65] It uses Apple's WebKit for rendering web pages and running JavaScript. WebKit consists of WebCore (based on Konqueror's KHTML engine) and JavaScriptCore (originally based on KDE's JavaScript engine, named KJS). Like KHTML and KJS, WebCore and JavaScriptCore are free software and are released under the terms of the GNU Lesser General Public License. Some Apple improvements to the KHTML code are merged back into the Konqueror project. Apple also releases additional code under an open source 2-clause BSD-like license.

Until Safari 6.0, it included a built-in web feedaggregator that supported the RSS and Atom standards. Current features include Private Browsing (a mode in which no record of information about the user's web activity is retained by the browser),^[66] the ability to archive web content in WebArchive format, the ability to email complete web pages directly from a browser menu, the ability to search bookmarks, and the ability to share tabs between all Mac and iOS devices running appropriate versions of software via an iCloud account.

iOS-specific features[edit]

This section needs to be updated. Please update this article to reflect recent events or newly available information.(July 2020)

Safari on an iPad running iOS 12 in Landscape view

iOS-specific features for Safari enable:

New in iOS 4[edit]

iOS 4.2[edit]

iOS 4.3[edit]

• Integration of the Nitro JavaScript engine for faster page loads. This feature was expanded to home-screen web applications in iOS 5.0.^[69]

New in iOS 5[edit]

• True tabbed browsing, similar to the desktop experience, only for iPads.^[70]
• Reading List, a bookmarking feature that allows tagging of certain sites for reading later, which syncs across all Safari browsers (mobile and desktop) via Apple's iCloud service.^[70]
• Reader, a reading feature that can format text and images from a web page into a more readable format, similar to a PDF document, while stripping out web advertising and superfluous information.^[70]
• Private browsing, like in most desktop browsers a feature that does not save the user's cookies and history, or allow anything to be written into local storage or webSql databases.

New in iOS 6[edit]

• iCloud Tabs, linking the desktop and iOS versions of Safari.
• Offline Reading Lists allow users to read pages stored previously without remaining connected to the internet.^[71]
• Full-screen landscape view for iPhone and iPod touch users hides most of the Safari controls except back and forward buttons and the status bar when in landscape mode.

New in iOS 7[edit]

• New icon
• 64-bit build on supported devices using the A7 processor.
• iCloud Keychain: iCloud can remember passwords, account names and credit card numbers. Safari can also autofill them as well. Requires devices that run iOS 7.0.3 and later and OS X Mavericks or later.
• Password Generator: When creating a new account, Safari can suggest the user a long, more secure, hard to guess password and Safari will also automatically remember the password.
• Shared Links
• Do Not Track
• Parental controls
• Tab limit increased from 9 to 36
• New Tab view (iPhone and iPod touch only)
• Unified smart search field
• Sync Bookmarks with Google Chrome^[72] and Firefox^[73] on Windows.

New in iOS 8[edit]

• The Tab view from iPhone is now available on iPads.
• A search function to search through all open tabs has been added in Tab view on iPad and select iPhones.
• Two-finger pinch to reveal Tab view on iPads and select iPhones.
• New Sidebar that slides out to reveal bookmarks, Reading List, and Shared Links on iPads and select iPhones in landscape view.
• Address bar now hides when scrolling down on iPads.
• Spotlight Search is now available from Safari's address bar.
• Option to “Scan Credit Card” when filling out credit card info on a web form.
• WebGL support.
• APNG support.
• Private browsing per tab.
• RSS feeds in Shared Links.
• DuckDuckGo support.
• Option to Request the desktop site while entering a web address.
• Option to add a website to Favorites while entering a web address.
• Swipe to close iCloud tabs from other devices.
• Hold the "+" (new tab button) in tab view to list recently closed tabs is now available on iPhone.
• Can delete individual items from History.
• Safari now blocks ads from automatically redirecting to the App Store without user interaction.
• Bookmark icon updated.
• Improved, iPad-like interface available on select iPhones in landscape view.

New in iOS 9[edit]

Safari on iOS 12, on the Wikipedia mobile landing page

• The option to add content blocking extensions is available to block specific web content.
• Safari view controller can be used to display web content from within an app, sharing cookies and other website data with Safari.
• Improved reader view, allowing the user to choose from different fonts and themes as well as hiding the controls

New in iOS 10[edit]

New in iOS 11[edit]

• More rounded search bar
• Redesigned video player
• Modified scrolling speed and momentum

New in iOS 12[edit]

• Support for stronger password suggestion
• Support for auto-fill from third-party provider
• Third-party can suggest strong password
• Auto-fill of 2FA code sent by email
• Fullscreen Support

WebKit2[edit]

WebKit2 has a multiprocess API for WebKit, where the web-content is handled by a separate process than the application using WebKit. Apple announced WebKit2 in April 2010.^[75] Safari for OS X switched to the new API with version 5.1.^[76] Safari for iOS switched to WebKit2 with iOS 8.^[77]

Security[edit]

Plugins[edit]

Apple maintains a plugin blacklist that it can remotely update to prevent potentially dangerous or vulnerable plug-ins from running on Safari. Initially, Apple had blocked versions of Flash and Java in earlier versions of Safari. Since Safari 12 support for NPAPI plugins (except for Flash) have been completely dropped. Starting with the release of Safari 14, support for Adobe Flash Player will be dropped altogether.^[62]

License[edit]

The license has common terms against reverse engineering, copying and sub-licensing, except parts that are open source, and it disclaims warranties and liability.^[78]

Apple tracks use of the browser. Windows users may not opt out of tracking since their license omits the opening If clause.^[79] Other users may opt out, and all users can opt out of location tracking by not using location services. "If you choose to allow diagnostic and usage collection, you agree that Apple and its subsidiaries and agents may collect... usage and related information... to provide ... services to you (if any) related to the Apple Software... in a form that does not personally identify you... Apple may also provide any such partner or third party developer with a subset of diagnostic information that is relevant to that partner’s or developer’s software... Apple and its partners, licensees, third party developers and website may transmit, collect, maintain, process and use your location data... and location search queries... in a form that does not personally identify you ... You may withdraw this consent at any time..."^[78]

Apple thinks "personal" does not cover "unique device identifiers" such as serial number, cookie number, or IP address, so they use these where allowed by law.^[80] "We may collect, use, transfer, and disclose non-personal information for any purpose. The following are some examples of non-personal information that we collect ... unique device identifier... We treat information collected by cookies and other technologies as non‑personal information. However, to the extent that Internet Protocol (IP) addresses or similar identifiers are considered personal information by local law, we also treat these identifiers as personal information."^[80]

In September 2017 Apple announced that it will use artificial intelligence (AI) to reduce the ability of advertisers to track Safari users as they browse the web. Cookies used for tracking will be allowed for 24 hours, then disabled, unless AI judges the user wants the cookie.^[81] Major advertising groups objected, saying it will reduce the free services supported by advertising, while other experts praised the change.^[82]

Browser exploits[edit]

An overview and detailed information about Safari exploits are listed by CVE Details.

In the PWN2OWN contest at the 2008 CanSecWest security conference in Vancouver, British Columbia, an exploit of Safari caused Mac OS X to be the first OS to fall in a hacking competition. Participants competed to find a way to read the contents of a file located on the user's desktop in one of three operating systems: Mac OS X Leopard, Windows Vista SP1, and Ubuntu 7.10. On the second day of the contest, when users were allowed to physically interact with the computers (the prior day permitted only network attacks), Charlie Miller compromised Mac OS X through an unpatched vulnerability of the PCRE library used by Safari.^[83] Miller was aware of the flaw before the conference and worked to exploit it unannounced, as is the common approach in these contests.^[83] The exploited vulnerability and other flaws were patched in Safari 3.1.1.^[84]

In the 2009 PWN2OWN contest, Charlie Miller performed another exploit of Safari to hack into a Mac. Miller again acknowledged that he knew about the security flaw before the competition and had done considerable research and preparation work on the exploit.^[85][86] Apple released a patch for this exploit and others on May 12, 2009 with Safari 3.2.3.^{[87][88][permanent dead link]}

System requirements[edit]

This section needs to be updated. Please update this article to reflect recent events or newly available information.(October 2019)

Safari 6.0 requires a Mac running Mac OS X v10.7.4 or later.^[89] Safari 5.1.7 requires a Mac running Mac OS X v10.6.8 or any PC running Windows XP Service Pack 2 or later, Windows Vista, or Windows 7.^[90][91] Safari 5.0.6 requires a Mac running on Mac OS X 10.5.8.^[92]

64-bit builds[edit]

The version of Safari included in Mac OS X v10.6 (and later versions) is compiled for 64-bit architecture. Apple claims that running Safari in 64-bit mode will increase rendering speeds by up to 50%.

On 64-bit devices, iOS and its stock apps are 64-bit builds including Safari.^[93][94]

Criticism[edit]

Distribution through Apple Software Update[edit]

An earlier version of Apple Software Update (bundled with Safari, QuickTime, and iTunes for Microsoft Windows) selected Safari for installation from a list of Apple programs to download by default, even when an existing installation of Safari was not detected on a user's machine. John Lilly, former CEO of Mozilla, stated that Apple's use of its updating software to promote its other products was "a bad practice and should stop." He argued that the practice "borders on malware distribution practices" and "undermines the trust that we're all trying to build with users."^[95] Apple spokesman Bill Evans sidestepped Lilly's statement, saying that Apple was only "using Software Update to make it easy and convenient for both Mac and Windows users to get the latest Safari update from Apple."^[96] Apple also released a new version of Apple Software Update that puts new software in its own section, though still selected for installation by default.^[97] By late 2008, Apple Software Update no longer selected new installation items in the new software section by default.^{[citation needed]}

Security updates for Snow Leopard and Windows platforms[edit]

Software security firm Sophos detailed how Snow Leopard and Windows users were not supported by the Safari 6 release at the time,^[98] while there were over 121 vulnerabilities left unpatched on those platforms.^[99] Since then, Snow Leopard has had only three minor version releases (the most recent in September 2013^[100]), and Windows has had none.^[101] While no official word has been released by Apple, the indication is that these are the final versions available for these operating systems, and both retain significant security issues.^[102][103]

Failure to adopt modern standards[edit]

While Safari pioneered several now standard HTML5 features (such as the Canvas API) in its early years, it has come under attack<^{[citation needed]} for failing to keep pace with some modern web technologies. Since 2015, iOS has allowed third party web browsers to be installed, including Chrome, Firefox, Opera and Edge; however, they are all forced to use the underlying WebKit browser engine, and inherit its limitations.^[104]

Intentionally limiting ad blockers and tracking protection[edit]

Beginning in 2018, Apple made technical changes to Safari's content blocking functionality which prompted backlash from users^[105] and developers^[106] of ad blocking extensions, who said the changes made it impossible to offer a similar level of user protection found in other browsers. Internally, the update limited the number of blocking rules^[107] which could be applied by third-party extensions, preventing the full implementation of community-developed blocklists. In response, several developers of popular ad and tracking blockers announced their products were being discontinued^[108], as they were now incompatible with Safari's newly-limited content blocking features. As a matter of policy, Apple requires the use of WebKit,