%%% @TeX-style-file %%% { %%% author = "Philip Taylor", %%% version = "3.05", %%% date = "7 April 2011", %%% time = "15:59:31 BST", %%% filename = "path.sty", %%% license = "Unlimited copying and redistribution of this %%% file are permitted as long as this file is %%% not modified. Modifications are permitted, %%% but only if the resulting file is not named %%% path.sty regardless of case", %%% address = "c/o The Hellenic Institute %%% Royal Holloway, University of London %%% Egham Hull, Egham %%% Surrey TW20 0EX %%% United Kingdom", %%% telephone = "Tel: +44 1622 820 735", %%% email = "P.Taylor@Rhul.Ac.Uk", %%% codetable = "ISO/ASCII", %%% keywords = "file name, filename, path name, pathname, %%% discretionary, discretionaries", %%% supported = "yes", %%% docstring = {Computer filenames, host names, and e-mail %%% addresses tend to be long strings that %%% cause line breaking problems for TeX. %%% Sometimes rather long strings are %%% encountered, such as: %%% %%% mighty-mouse-gw.scrc.symbolics.com %%% %%% This file defines a macro, \path|...|, %%% similar to LaTeX's \verb|...| macro, that %%% sets the text in the typewriter font, %%% allowing hyphen-less line breaks at %%% punctuation characters. %%% %%% The default set of punctuation characters is %%% defined as %%% %%% \discretionaries |~!@$%^&*()_+`-=#{"}[]:;'<>,.?\/| %%% %%% However, you can change it as needed, for %%% example %%% %%% \discretionaries +@%!.+ %%% %%% would assign to it the set @ % ! . which %%% commonly occur in electronic mail addresses. %%% %%% The delimiter characters surrounding the %%% arguments to \discretionaries and \path %%% will normally be a punctuation character not %%% found in the argument, and not otherwise %%% significant to TeX. In particular, backslash %%% cannot be used as a delimiter. In the rare %%% event that this is required, set %%% %%% \specialpathdelimiterstrue %%% %%% This practice is not recommended, because TeX %%% then runs about four times slower while it is %%% typesetting \path\...\ requests. %%% \specialpathdelimitersfalse is the normal %%% setting. %%% %%% This file may be used in Plain TeX or AmSTeX %%% by %%% %%% \input path.sty %%% %%% and in LaTeX by %%% %%% \usepackage {path} %%% %%% The checksum field above, if present, contains %%% a CRC-16 checksum as the first value, followed %%% by the equivalent of the standard UNIX wc (word %%% count) utility output of lines, words, and %%% characters. This is produced by Robert %%% Solovay's checksum utility. %%% } %%% } %%% ==================================================================== % ====================================================================== % % The \path Macro % % An early prototype of this code which was developed by Nelson H. F. % Beebe used a recursive macro to peel off one % character at a time, then with a nest of \ifx statements, test for % each punctuation character, and if one is found, insert discretionary % linebreaks in the form of \- or \penalty0. While this usually works, % it is not entirely satisfactory, because the nested \ifx statements % are ugly, and the recursion can exhaust TeX's macro stack for % arguments longer than about 30 characters. % % A second version by NHFB built upon the ideas used by the LaTeX % \verb|...| macro; this was mostly successful, but was flawed by its % inability to handle at least the three characters ^ [ ], by loss of % use of @ for naming internal private macros, and by its inability to % allow the set of characters at which breaks are permissible. % % The problem was then passed off to Philip Taylor, who wrote this third % version which fully solves the problem. Phil therefore gets the % author credit in the file header above. [Gee, thanks ! ** Phil] % % Revision history (reverse time order): % % 3.05 [7-Apr-2011] % Summary: Use \path instead of \ATcode to save the catcode of @. % Expand \pathafterhook after \path's work is done. % % 3.04 [incorporated as part of Eplain] % % 3.03b [22-Jul-1997] % Summary: No code changes. An out-of-date example removed, the syntax % for LaTeX2e changed to "\usepackage {path}", and the "checksum" % comments modified to indicate that the checksum is not necessarily % present. The initial opening brace re-positioned to align % with the matching closing brace, and the inner closing brace % re-positioned to align with the inner opening brace. A check made % to ensure that no characters are present. % % 3.03a [22-Jul-1997] % Summary: No code changes from V3.03; e-mail addresses, telephone numbers, % etc. amended to reflect current situation. Checksum removed. % % 3.03 [03-Dec-1991] % Problem: A \path|...| entry in a LaTeX index is set with an % preceding blank line. % Solution: Add \leavevmode before \hbox{}; without it, an extra blank % line gets typeset before every invocation of \path in a % LaTeX theindex environment. % Diagnosis: To avoid hyphenation, the expansion of \path starts with % an \hbox{} (c.f. the TeXbook, p. 454); LaTeX indices are % basically set in vertical mode, with each index entry a % paragraph in its own right; when an \hbox {} is % encountered in vmode, it is typeset in vmode, and TeX % continues; the first real character is encountered, and % TeX enters unrestricted horizontal mode. The \hbox {} and % the first character are therefore on separate lines, % separated by \baselineskip. % % 3.02 [23-Oct-1991 13:45:12] % First released version. % ====================================================================== \immediate \write 16 {This is path.sty, Version 3.05 <7-Apr-2011>} %%% The use of `\path' as a temporary control sequence is a kludge, but %%% it's a reasonably simple way to accomplish making @ a letter (to %%% provide ``concealed'' control sequences) without overwriting something %%% (without an `@' in its name) that might already be defined. \edef \path {\the \catcode `\@}% \catcode `\@ = 11 \let \oldc@tcode = \path %%% Make commercial-at a letter to provide concealed control sequences \catcode `\@ = 11 %%% and then declare two \count variables with concealed names \newcount \c@tcode \newcount \c@unter %%% and a boolean variable with an open name, to specify the nature of %%% the delimiters which will be associated with the \path command \newif \ifspecialpathdelimiters %%% If desired, you can define \pathafterhook to take \let \pathafterhook = \relax %%% We need to define control sequences which expand to both %%% active and passive spaces ... \begingroup \catcode `\ = 10 \gdef \passivesp@ce { } \catcode `\ = 13\relax% \gdef\activesp@ce{ }% \endgroup %%% \discretionaries will define a macro \discr@ti@n@ri@s which will %%% make every character between the first and final a %%% discretionary breakpoint. \def \discretionaries %%% % start a group (ended in \discr@ti@naries ), % make every character (apart from space) active, % then transfer control to \discr@tionaries ... {\begingroup \c@tcodes = 13 \discr@tionaries } %%% \discr@tionaries will receive as parameter the initial %%% which will delimit the set of discretionaries; this %%% will be active. \def \discr@tionaries #1%%% % knowing the initial , define \discr@ti@naries with a % delimited parameter structure, the final delimiter being % the second instance of {\def \discr@ti@naries ##1#1%%% % \discr@ti@naries will receive as parameter the set of % characters which are to be made discretionaries; it ends the % group and then defines \discr@ti@n@ries, a tail-recursive % macro, which will quietly process each of the % discretionaries in turn ... {\endgroup \def \discr@ti@n@ries ####1%%% or % by checking to see if it has reached the end of the % list, i.e. comparing its parameter with the known final % ... {\if \noexpand ####1\noexpand #1% % if they match, there is no more to do ... \let \n@xt = \relax \else % otherwise the character becomes active and is % defined as a discretionary with the three texts % , thus suppressing true % hyphenation; the macro then prepares to process % the next character. \catcode `####1 = 13 \def ####1{\discretionary {\char `####1}{}{\char `####1}}% \let \n@xt = \discr@ti@n@ries \fi % here either tail recursion occurs, if there are more % characters to process, or tail recursion ceases if all % have been exhausted \n@xt }% % of course, none of this must take place during the \stress % {call} of \discretionaries; execution is deferred until % another macro such as \path calls the concealed action % macro \discr@ti@n@ri@s; we therefore define this macro to % have the desired effect. \def \discr@ti@n@ri@s {\discr@ti@n@ries ##1#1}% }% % and finally invoke \discr@ti@naries to process the set of % discretionary characters; it will take everything up to the % second instance of as its parameter (and therefore as the % set of characters to be processed). \discr@ti@naries } %%% \path, which is the user interface to \p@th, first checks %%% to see whether \specialpathdelimiters is or ; %%% if , it needs to take special action to ensure that %%% \stress {all} characters (apart from ) are acceptable %%% as delimiters; this is very time-consuming, and should be %%% avoided if at all possible. It also defines \endp@th, to %%% close the appropriate number of groups, and finally transfers %%% control to \p@th \def \path {\ifspecialpathdelimiters \begingroup \c@tcodes = 12 \def \endp@th {\endgroup \endgroup \pathafterhook}% \else \def \endp@th {\endgroup \pathafterhook}% \fi \p@th } %%% \p@th, which has essentially the same syntax as \discretionaries, %%% expects to be followed by a , a , and a second instance %%% of ; it will typeset in the \tt font with hyphenation %%% inhibited --- breaks (but not true hyphenation) are allowed at any %%% or all of the special characters which have %%% previously been declared as \discretionaries. \def \p@th #1%%% % it starts by opening a group (ended in \p@@th) ... % switches to \tt, and inhibits hyphenation; % allows breaks at \discretionaries; saves the % catcode of the initial , because if it has also been % declared within \discretionaries, the initial and final % won't match (which would be a disaster); \c@tcode is overloaded, % but there is no conflict that I can detect ... % invokes \discr@ti@n@ri@s to render active the set of special % characters which have previously been declared as % \discretionaries; each of these characters will expand to a real % \discretionary, with replacement texts ; and % re-instates the catcode of the initial . {\begingroup \tt \c@tcode = \catcode `#1 \discr@ti@n@ri@s \catcode `\ = \active \expandafter \edef \activesp@ce {\passivesp@ce \hbox {}}% \catcode `#1 = \c@tcode % it next defines an inner macro \p@@th with delimited parameter % structure, the final delimiter being the same as the initial % delimiter which it has itself received as #1. \def \p@@th ##1#1% % Within \p@@th, % permissible breakpoints are specified, % the path is typeset, % and the group(s) ended. {\leavevmode \hbox {}##1% \endp@th }% \p@@th } %%% \c@tcodes expects to be followed by the r-h-s of a numeric %%% assignment optionally including the assignment operator; it saves %%% the value of the r-h-s in \c@tcode, and invokes \c@tc@des. \def \c@tcodes {\afterassignment \c@tc@des \c@tcode} %%% \c@tc@des uses the value saved in \c@tcode, and assigns it to the %%% \catcode of all characters with the single exception of , %%% which retains its normal catcode of 10; on exit, every single %%% character apart from will have the catcode which followed %%% \c@tcodes. The code is optimised to avoid unnecessary usage of %%% save-stack space. \def \c@tc@des {\c@unter = 0 \loop \ifnum \catcode \c@unter = \c@tcode \else \catcode \c@unter = \c@tcode \fi \ifnum \c@unter < 255 \advance \c@unter by 1 \repeat \catcode `\ = 10 } %%% We restore the original catcode of commercial-at \catcode `\@ = \oldc@tcode %%% Define a default set of discretionary break characters %%% to include all punctuation characters except vertical %%% bar \discretionaries |~!@$%^&*()_+`-=#{}[]:";'<>,.?\/| \endinput