In this chapter, we discuss the general editing facilities present in TeXmacs. Besides classical features, like “cut and paste”, “search and replace”, etc., TeXmacs also supports additional functionality which exploits the document structure. Examples of such features are “structured cursor movement” and “structured variants”. It should be noticed that more traditional functions like “search and replace” also attempt to exploit the structure. For instance, when searching in math mode, you will only find matches which are also in math mode.

1.Cut and paste

You can select text and formulas by maintaining the left mouse button. In order to delete the selected region, use EditCut. In order to copy the selected region, first click on EditCopy. Next, paste it as many times as you want to the location of your cursor, using EditPaste. Alternatively, you may copy a selected region using the middle mouse button.

It is also possible to the change text properties of a selected region. For instance, in order to transform some black text in red, you select it using the left mouse button and click on FormatColorRed. Similarly, if you select a formula and you click on InsertFraction, then the formula becomes the numerator of some fraction.

When using the copy and paste mechanism to communicate with other applications, text is copied and pasted using the TeXmacs data format. You may specify other import and export formats using EditImport resp. EditExport. By default, copying and pasting uses the primary text buffer. Using EditCopy to and EditPaste from, you may specify as many other buffers as you like.

There are two ways to make selections using the keyboard. When using the cursor keys left, right, etc. while holding down the button, you may select text while moving around the cursor. Alternatively, you may press once to fix a starting position. When moving around using the cursor keys, the text between the starting position and the current position keeps being selected. The selection gets cleared when pressing .

Notice that the shortcut also allows you to make structured selections. You may select the current word you are in by pressing twice. Each additional time you press results in the selection of the smallest structure which englobes the current selection. Ultimately, when the entire document gets selected, pressing once more clears the selection.

2.Search and replace

You can start searching text by pressing C-s or EditSearch. During a search, the “search string” is displayed at the left hand side of the footer. Each character you type is appended to this search string and the next occurrence of it is surrounded by a red box. When pressing C-s a second time during a search, the next occurrence is being searched. A beep indicates that no more occurrences were found in the document; pressing C-s will continue the search at the beginning of your document. You may press backspace in order to undo key presses during a search.

Usually, text is being searched for in a forward manner, starting from the current cursor position. You may also search backwards, using C-r. During a search, only text in the same mode and the same language will be found, as those which are active at the position where you started your search. In other words, when searching an in math-mode, you will not find any x's in the ordinary text. As a current limitation, the search string can only contain ordinary text and no math-symbols or more complicated structured text.

A query replace is started by pressing C-= or EditReplace. You are prompted for a string which is to be replaced and the string by which to replace. At each occurrence of the string to be replaced you are prompted and you have to choose between replacing the string (y), not replacing it (n) and replace this and all further occurrences (a). Like in the case of searching, the query-replace command is mode and language sensitive.

The current search and replace system is still quite rudimentary in the sense that it only deals with plain text in a simple way. In the future, we plan to implement a more powerful search and replace mechanism for structured text.

For the time being, you may nevertheless search and replace arbitrary document fragments as follows: somewhere inside your document or another window, select the fragment you want to search using EditCopy toSearch. Next hit C-s twice or more in order to search all occurences of this fragment inside your document. Similarly, using EditCopy toReplace, you may select a second fragment by which you want to replace the first one. In order to start the replacement, hit C-= and specify the empty string as the arguments for “Replace” and “Replace by”.

3.Spell checking

If the program ispell has been installed on your system, then you may use it to check your text for misspelled words by pressing or EditSpell. Notice that you might have to verify that the dictionaries corresponding to the languages in which your texts have been written have been installed on your system; this is usually the case for English.

When you launch the spell checker (either on the whole text or a selected region), you will be prompted at each misspelled word and the footer displays the available options:

a)

Accepts the misspelled word and all its future occurrences in the text.

r)

Replace the misspelled word by a correction you have to enter.

i)

Indicate that the “misspelled” word is actually correct and that it has to be inserted in your personal dictionary.

1-9)

Several suggested corrections for your misspelled word.

Notice that ispell just checks for misspelled words. No grammatical faults will be detected.

When starting the spell checker, it will use the dictionary of the language which is active at the current cursor position (or the start of a selection). Only text in that language will be checked for. If your document contains text in several languages, then you will have to launch the spell checker once for each language being used.

4.Undo and redo

It is possible to gradually undo the changes you made in a document from the moment that you launched TeXmacs. This can be done via EditUndo or using the keystrokes or . Undone changes can be “redone” using EditRedo or .

In order to save memory, the number of successive actions which can be undone is limited to 100 (by default). It is possible to increase this number by adding a command like

    (set-maximal-undo-depth 1000)

in our personal initialization file (see HelpScheme). When specifying a negative number as your maximal undo depth, any number of actions can be undone.

5.Structured editing

As a general rule, the behaviour of most structured editing operations is conditioned by the current focus. By default, the innermost tag which contains the cursor. Whenever some selection is active, the current focus is the innermost tag which contains the selection. During structured operations, such as navigating among similar tags, the current focus may temporarily be set to something else. The current focus is visually indicated by the innermost cyan box around the cursor.

For instance, the structured insertion commands , , and have a particular meaning both inside tables and trees. Inside tables, they allow you to insert new rows and columns (see figure ?). Inside trees, they correspond to the insertion of new nodes (see figure ?). Whenever you inside a tree inside a table, then the innermost tag is a tree, and node insertions will take precedence over the insertion of new rows and columns.

In many cases, a “default particular behaviour” has been defined for all tags minus some exceptional ones. In our example of structured insertion, the default behaviour of and is to insert a new argument to the tag at the left or at the right (when allowed).

Figure 1. Assume that the cursor is at the position of | inside the left-most matrix. Then the four other matrices respectively correspond to the insertion of a new column at the left () or right (), resp. a new row above () or below ().

Figure 2. Assume that the cursor is at the position of | inside the left-most tree. Then the four other trees respectively correspond to the insertion of a new node at the left (), at the right (), above () or below ().

Similarly, in the case of matrices, the keys , , and can be used for inserting a new first or last column, resp. a new first or last row. The keys and are mapped to the commands for backward resp. forward structured deletion. In the case of matrices, this will result in the removal of the column before or after the cursor (see figure ?). In order to remove the enclosing environment you may use or .

Figure 3. Assume that the cursor is at the position of | inside the left-most matrix. Then pressing the keys and respectively result in the next two matrices. Pressing either or replaces the matrix by the content of the cell in which you are, leaving you with the at the right-hand side.

6.Structured cursor movement

TeXmacs implements the three main mechanisms for structured cursor movement:

1. Traversal of the entire structure of the document.
2. Traversal of tags which are similar to the innermost tag.
3. Movements inside the innermost tag.

Most keyboard shortcuts for structured cursor movements can be used in combination with the -key so as to similtaneously select text while moving around.

To do: customizing the behaviour

Structured traversal of the document

The , , and keys are used for the structured traversal of the entire document. Inside plain text, and allow you to move in a word-by-word manner, while and correspond to paragraph-by-paragraph motion.

In the presence of other markup, the and keys allow you to visit all accessible cursor positions of the document, except that we keep moving in a word-by-word manner inside plain text. The behaviour of the and keys is more context-dependent. Inside matrices, they typically allow you to move one row up or down.

Traversal of tags which are similar to the innermost tag

This type of cursor movement allows you to quickly visit all other tags in the document which are similar to the innermost tag. The and keys allow you move to the previous or next similar tags, whereas and directly jump to the first or last similar tags.

For instance, when you are inside a section title, you may move to the previous sectional title (which could also be the title of a subsection or a chapter, for instance) using . Notice that you may use C-§ to jump to the previous section title.

Movements inside the innermost tag

It is also possible to quickly move inside the innermost tag without quitting it. The shortcuts , , and provide a way to move to the previous, next, first or last argument of the innermost tag. Furthermore, the shortcuts and may be used to exit the innermost tag on the left or on the right.

This default behaviour may be overridden in special contexts. For instance, inside tables or trees, they rather correspond to cell-by-cell or node-by-node cursor movement. In addition, these cases associate vertical cursor movements to , , and .

7.Structured variants

When creating an environment like a theorem, an equation or a list, it frequently happens that one would like to change the environment a posteriori. This can be done using the keyboard shortcuts and for cycling through the list of structured variants of the innermost tag in a direct or inverse manner.

For instance, assuming that you are inside a theorem, pressing several times will change the theorem into a proposition, a lemma, a corollary, a conjecture and back into a theorem. The key allows you to cycle in the inverse direction: theorem conjectur corollary lemma proposition theorem.

In the case of mathematical formulas, the shortcuts allows you to change an inline formula like into a displayed formula like

while taking care of potential “trailing spaces and punctuation signs”.

TeXmacs also provides the shortcut for turning numbered environments into unnumbered environments and vice versa. This works for most common environments like theorems, remarks, tables, equations, etc. Notice that also turns an unnumbered itemize environment into an enumeration and vice versa, whereas allows you to cycle between the available kinds of list items (bullets, dashes, arrows, etc.).

8.Positioning and resizing objects

The prefix may be used for positioning and resizing objects. For instance, inside a cell of a table, you may use structured:geometry right to align the cell more to the right. Behind a space introduced via FormatSpace, the same key allows you to increase the width of space. More generally, the following shortcuts are implemented:

Decrease the horizontal size of an object, or align more to the left.

Increase the horizontal size of an object, or align more to the right.

Decrease/increase the vertical size of an object, or align more to the bottom.

Increase/decrease the vertical size of an object, or align more to the top.

Decrease the horizontal offset of an object, or left align.

Increase the horizontal offset of an object, or right align.

Decrease the vertical offset of an object, or align at the bottom.

Increase the vertical offset of an object, or align at the top.

Revert the geometry (size, position, alignment) to the defaults.

,

Circulate among available length units for specifying the geometry.

,

Decrease or increase the step size when positioning or resizing.

Particular tags to which the shortcuts apply are the following

Spaces

Both horizontal and vertical spaces from the FormatSpace menu. You should put the cursor just after the space tag for the shortcuts to apply.

Box modifiers

The tags move, shift, resize and clipped from the FormatTransform menu.

Animations

The durations of animations can be modified using and .

Images

The size and alignment of images can be changed.

9.Versioning tools

When writing documents in collaboration with other authors, it frequently arises that one wants to go through changes made by the other authors, and either accept, discard or further correct them. After enabling the versioning tool through EditPreferencesUtilitiesVersioning tool, a special menu Version appears in the main menu bar, which makes this process automatic. Let us describe in more detail how this tool works.

For the moment, automatic version control systems such as Subversion are not yet supported. In the future, we intend to include support for such systems and the operation of merging two different new versions.

Comparing two versions

Assume that we have two versions old.tm and new.tm of the same document. In order to see the changes, first load the newest version new.tm, then click on VersionFileCompare and select the oldest version old.tm. The buffer will still be named new.tm, and the changes between both versions will be indicated by special markup. If there are any changes, then the cursor will be positioned at the first difference.

It is possible to go through all the differences between the old and new versions either from the items in the submenu VersionMove, or using the keyboard shortcuts and . One may also use the more general structured navigation shortcuts , , and .

Visualization of the differences

Differences between the two versions can be displayed in three ways: by showing only the old version, only the new version, or both versions simultaneously. In all cases, the old version is displayed in dark red and the new version in dark green.

The visualization style can be specified individually for each individual change, via VersionShow or the keyboard shortcuts (old version), (new version) and (both versions). One may also cycle through the different style using the structured variant key . If you selected some text, then the above actions will apply to the whole selection. The visualization style may also be specified globally, using VersionFileShow old version, VersionFileShow new version and VersionFileShow both versions.

Retaining a specific version

It often occurs that we want to go through the changes between two versions and progressively retain either one or the other version for each individual difference. Assuming that the cursor is inside a given difference, this can be done from entries in the submenu VersionRetain. Alternatively, one may use the shortcuts , and to retain the old, new and currently displayed version, respectively. If both versions are displayed, then retains the new version. After retaining one of the versions, we automatically jump to the next difference, which can then be processed. If you selected some text, then any of the above action will retain the appropriate version for each of the differences in the selection.

It is also possible to globally select the old, new or current version using VersionFileRetain old version, VersionFileRetain new version, resp. VersionFileRetain current version. A convenient alternative way to process all differences is to use and to go through the differences, use and to select the preferred version, and then click on VersionFileRetain current version as soon as all differences have been processed.

Grain control and reactualizing the differences

The entries in the submenu VersionGrain allow you to control the grain with which differences between versions are computed. By default, we use the finest grain Detailed. It is also possible to compute differences on a paragraph-based level, using Block. In that case, the entire paragraphs in which a change occurs will be highlighted. The roughest grain Rough will highlight the entire text, if a change occurs within it.

The grain is used when comparing two documents using VersionFileCompare, but it is also possible to recompare the differences in a selected portion of text using VersionReactualize. The latter possibility is especially useful after a change the grain.

Similarly, the cursor is inside a difference, then you may recompare the two versions VersionReactualize. This may be useful, if you made some changes to one of the versions. For instance, assume that the old version contained a theorem and that we changed it into a lemma in the new version and also modified part of its inside text. When visualizing the changes, the whole theorem will be highlighted, since there is no appropriate markup to indicate that we just changed from a theorem to a lemma. Nevertheless, if we want to compare the inside texts, we may turn the old theorem into a lemma and then click on VersionReactualize.