#!/home/grinnz/projects/cpandoc-browser/perls/5.38.2/bin/perl package main; use Perl::Tidy; my $arg_string = undef; # give Macs a chance to provide command line parameters if ( $^O =~ /Mac/ ) { $arg_string = MacPerl::Ask( 'Please enter @ARGV (-h for help)', defined $ARGV[0] ? "\"$ARGV[0]\"" : "" ); } # Exit codes returned by perltidy: # 0 = no errors # 1 = perltidy could not run to completion due to errors # 2 = perltidy ran to completion with error messages exit Perl::Tidy::perltidy( argv => $arg_string ); __END__ =head1 NAME perltidy - a perl script indenter and reformatter =head1 SYNOPSIS perltidy [ options ] file1 file2 file3 ... (output goes to file1.tdy, file2.tdy, file3.tdy, ...) perltidy [ options ] file1 -o outfile perltidy [ options ] file1 -st >outfile perltidy [ options ] outfile =head1 DESCRIPTION Perltidy reads a perl script and writes an indented, reformatted script. The formatting process involves converting the script into a string of tokens, removing any non-essential whitespace, and then rewriting the string of tokens with whitespace using whatever rules are specified, or defaults. This happens in a series of operations which can be controlled with the parameters described in this document. Perltidy is a commandline frontend to the module Perl::Tidy. For documentation describing how to call the Perl::Tidy module from other applications see the separate documentation for Perl::Tidy. It is the file Perl::Tidy.pod in the source distribution. Many users will find enough information in L<"EXAMPLES"> to get started. New users may benefit from the short tutorial which can be found at http://perltidy.sourceforge.net/tutorial.html A convenient aid to systematically defining a set of style parameters can be found at http://perltidy.sourceforge.net/stylekey.html Perltidy can produce output on either of two modes, depending on the existence of an B<-html> flag. Without this flag, the output is passed through a formatter. The default formatting tries to follow the recommendations in perlstyle(1), but it can be controlled in detail with numerous input parameters, which are described in L<"FORMATTING OPTIONS">. When the B<-html> flag is given, the output is passed through an HTML formatter which is described in L<"HTML OPTIONS">. =head1 EXAMPLES Here are some example perltidy commands followed by their meanings: perltidy somefile.pl This will produce a file F containing the script reformatted using the default options, which approximate the style suggested in perlstyle(1). The source file F is unchanged. perltidy *.pl Execute perltidy on all F<.pl> files in the current directory with the default options. The output will be in files with an appended F<.tdy> extension. For any file with an error, there will be a file with extension F<.ERR>. perltidy -b file1.pl file2.pl Modify F and F in place, and backup the originals to F and F. If F and/or F already exist, they will be overwritten. perltidy -b -bext='/' file1.pl file2.pl Same as the previous example except that the backup files F and F will be deleted if there are no errors. perltidy -gnu somefile.pl Execute perltidy on file F with a style which approximates the GNU Coding Standards for C programs. The output will be F. perltidy -i=3 somefile.pl Execute perltidy on file F, with 3 columns for each level of indentation (B<-i=3>) instead of the default 4 columns. There will not be any tabs in the reformatted script, except for any which already exist in comments, pod documents, quotes, and here documents. Output will be F. perltidy -i=3 -et=8 somefile.pl Same as the previous example, except that leading whitespace will be entabbed with one tab character per 8 spaces. perltidy -ce -l=72 somefile.pl Execute perltidy on file F with all defaults except use "cuddled elses" (B<-ce>) and a maximum line length of 72 columns (B<-l=72>) instead of the default 80 columns. perltidy -g somefile.pl Execute perltidy on file F and save a log file F which shows the nesting of braces, parentheses, and square brackets at the start of every line. perltidy -dbs -dbl=10 somefile.pl >blocks.csv This will dump a table of comma-separated metrics for subroutines longer than 10 lines to F. perltidy -duv somefile.pl >vars.txt This will dump a list of unused and reused lexical variable names to F. perltidy -html somefile.pl This will produce a file F containing the script with html markup. The output file will contain an embedded style sheet in the section which may be edited to change the appearance. perltidy -html -css=mystyle.css somefile.pl This will produce a file F containing the script with html markup. This output file will contain a link to a separate style sheet file F. If the file F does not exist, it will be created. If it exists, it will not be overwritten. perltidy -html -pre somefile.pl Write an html snippet with only the PRE section to F. This is useful when code snippets are being formatted for inclusion in a larger web page. No style sheet will be written in this case. perltidy -html -ss >mystyle.css Write a style sheet to F and exit. perltidy -html -frm mymodule.pm Write html with a frame holding a table of contents and the source code. The output files will be F (the frame), F (the table of contents), and F (the source code). =head1 OPTIONS - OVERVIEW The entire command line is scanned for options, and they are processed before any files are processed. As a result, it does not matter whether flags are before or after any filenames. However, the relative order of parameters is important, with later parameters overriding the values of earlier parameters. For each parameter, there is a long name and a short name. The short names are convenient for keyboard input, while the long names are self-documenting and therefore useful in scripts. It is customary to use two leading dashes for long names, but one may be used. Most parameters which serve as on/off flags can be negated with a leading "n" (for the short name) or a leading "no" or "no-" (for the long name). For example, the flag to outdent long quotes is B<-olq> or B<--outdent-long-quotes>. The flag to skip this is B<-nolq> or B<--nooutdent-long-quotes> or B<--no-outdent-long-quotes>. Options may not be bundled together. In other words, options B<-q> and B<-g> may NOT be entered as B<-qg>. Option names may be terminated early as long as they are uniquely identified. For example, instead of B<--dump-token-types>, it would be sufficient to enter B<--dump-tok>, or even B<--dump-t>, to uniquely identify this command. =head2 I/O Control The following parameters concern the files which are read and written. =over 4 =item B<-h>, B<--help> Show summary of usage and exit. =item B<-o>=filename, B<--outfile>=filename Name of the output file (only if a single input file is being processed). If no output file is specified, and output is not redirected to the standard output (see B<-st>), the output will go to F. [Note: - does not redirect to standard output. Use B<-st> instead.] =item B<-st>, B<--standard-output> Perltidy must be able to operate on an arbitrarily large number of files in a single run, with each output being directed to a different output file. Obviously this would conflict with outputting to the single standard output device, so a special flag, B<-st>, is required to request outputting to the standard output. For example, perltidy somefile.pl -st >somefile.new.pl This option may only be used if there is just a single input file. The default is B<-nst> or B<--nostandard-output>. =item B<-se>, B<--standard-error-output> If perltidy detects an error when processing file F, its default behavior is to write error messages to file F. Use B<-se> to cause all error messages to be sent to the standard error output stream instead. This directive may be negated with B<-nse>. Thus, you may place B<-se> in a F<.perltidyrc> and override it when desired with B<-nse> on the command line. =item B<-oext>=ext, B<--output-file-extension>=ext Change the extension of the output file to be F instead of the default F (or F in case the -B<-html> option is used). See L<"Specifying File Extensions">. =item B<-opath>=path, B<--output-path>=path When perltidy creates a filename for an output file, by default it merely appends an extension to the path and basename of the input file. This parameter causes the path to be changed to F instead. The path should end in a valid path separator character, but perltidy will try to add one if it is missing. For example perltidy somefile.pl -opath=/tmp/ will produce F. Otherwise, F will appear in whatever directory contains F. If the path contains spaces, it should be placed in quotes. This parameter will be ignored if output is being directed to standard output, or if it is being specified explicitly with the B<-o=s> parameter. =item B<-b>, B<--backup-and-modify-in-place> Modify the input file or files in-place and save the original with the extension F<.bak>. Any existing F<.bak> file will be deleted. See next item for changing the default backup extension, and for eliminating the backup file altogether. B: Writing back to the input file increases the risk of data loss or corruption in the event of a software or hardware malfunction. Before using the B<-b> parameter please be sure to have backups and verify that it works correctly in your environment and operating system. A B<-b> flag will be ignored if input is from standard input or goes to standard output, or if the B<-html> flag is set. In particular, if you want to use both the B<-b> flag and the B<-pbp> (--perl-best-practices) flag, then you must put a B<-nst> flag after the B<-pbp> flag because it contains a B<-st> flag as one of its components, which means that output will go to the standard output stream. =item B<-bext>=ext, B<--backup-file-extension>=ext This parameter serves two purposes: (1) to change the extension of the backup file to be something other than the default F<.bak>, and (2) to indicate that no backup file should be saved. To change the default extension to something other than F<.bak> see L<"Specifying File Extensions">. A backup file of the source is always written, but you can request that it be deleted at the end of processing if there were no errors. This is risky unless the source code is being maintained with a source code control system. To indicate that the backup should be deleted include one forward slash, B, in the extension. If any text remains after the slash is removed it will be used to define the backup file extension (which is always created and only deleted if there were no errors). Here are some examples: Parameter Extension Backup File Treatment <-bext=bak> F<.bak> Keep (same as the default behavior) <-bext='/'> F<.bak> Delete if no errors <-bext='/backup'> F<.backup> Delete if no errors <-bext='original/'> F<.original> Delete if no errors =item B<-bm=s>, B<--backup-method=s> This parameter should not normally be used but is available in the event that problems arise as a transition is made from an older implementation of the backup logic to a newer implementation. The newer implementation is the default and is specified with B<-bm='copy'>. The older implementation is specified with B<-bm='move'>. The difference is that the older implementation made the backup by moving the input file to the backup file, and the newer implementation makes the backup by copying the input file. The newer implementation preserves the file system B value. This may avoid problems with other software running simultaneously. This change was made as part of issue B at github. =item B<-w>, B<--warning-output> Setting B<-w> causes any non-critical warning messages to be reported as errors. These include messages about possible pod problems, possibly bad starting indentation level, and cautions about indirect object usage. The default, B<-nw> or B<--nowarning-output>, is not to include these warnings. =item B<-q>, B<--quiet> Deactivate error messages (for running under an editor). For example, if you use a vi-style editor, such as vim, you may execute perltidy as a filter from within the editor using something like :n1,n2!perltidy -q where C represents the selected text. Without the B<-q> flag, any error message may mess up your screen, so be prepared to use your "undo" key. =item B<-log>, B<--logfile> Save the F<.LOG> file, which has many useful diagnostics. Perltidy always creates a F<.LOG> file, but by default it is deleted unless a program bug is suspected. Setting the B<-log> flag forces the log file to be saved. =item B<-g=n>, B<--logfile-gap=n> Set maximum interval between input code lines in the logfile. This purpose of this flag is to assist in debugging nesting errors. The value of C is optional. If you set the flag B<-g> without the value of C, it will be taken to be 1, meaning that every line will be written to the log file. This can be helpful if you are looking for a brace, paren, or bracket nesting error. Setting B<-g> also causes the logfile to be saved, so it is not necessary to also include B<-log>. If no B<-g> flag is given, a value of 50 will be used, meaning that at least every 50th line will be recorded in the logfile. This helps prevent excessively long log files. Setting a negative value of C is the same as not setting B<-g> at all. =item B<-npro> B<--noprofile> Ignore any F<.perltidyrc> command file. Normally, perltidy looks first in your current directory for a F<.perltidyrc> file of parameters. (The format is described below). If it finds one, it applies those options to the initial default values, and then it applies any that have been defined on the command line. If no F<.perltidyrc> file is found, it looks for one in your home directory. If you set the B<-npro> flag, perltidy will not look for this file. =item B<-pro=filename> or B<--profile=filename> To simplify testing and switching .perltidyrc files, this command may be used to specify a configuration file which will override the default name of .perltidyrc. There must not be a space on either side of the '=' sign. For example, the line perltidy -pro=testcfg would cause file F to be used instead of the default F<.perltidyrc>. A pathname begins with three dots, e.g. ".../.perltidyrc", indicates that the file should be searched for starting in the current directory and working upwards. This makes it easier to have multiple projects each with their own .perltidyrc in their root directories. =item B<-opt>, B<--show-options> Write a list of all options used to the F<.LOG> file. Please see B<--dump-options> for a simpler way to do this. =item B<-f>, B<--force-read-binary> Force perltidy to process binary files. To avoid producing excessive error messages, perltidy skips files identified by the system as non-text. However, valid perl scripts containing binary data may sometimes be identified as non-text, and this flag forces perltidy to process them. =item B<-ast>, B<--assert-tidy> This flag asserts that the input and output code streams are identical, or in other words that the input code is already 'tidy' according to the formatting parameters. If this is not the case, an error message noting this is produced. This error message will cause the process to return a non-zero exit code. The test for this is made by comparing an MD5 hash value for the input and output code streams. This flag has no other effect on the functioning of perltidy. This might be useful for certain code maintenance operations. Note: you will not see this message if you have error messages turned off with the -quiet flag. =item B<-asu>, B<--assert-untidy> This flag asserts that the input and output code streams are different, or in other words that the input code is 'untidy' according to the formatting parameters. If this is not the case, an error message noting this is produced. This flag has no other effect on the functioning of perltidy. =back =head1 FORMATTING OPTIONS =head2 Basic Options =over 4 =item B<--notidy> This flag disables all formatting and causes the input to be copied unchanged to the output except for possible changes in line ending characters and any pre- and post-filters. This can be useful in conjunction with a hierarchical set of F<.perltidyrc> files to avoid unwanted code tidying. See also L<"Skipping Selected Sections of Code"> for a way to avoid tidying specific sections of code. =item B<-i=n>, B<--indent-columns=n> Use n columns per indentation level (default n=4). =item B<-l=n>, B<--maximum-line-length=n> The default maximum line length is n=80 characters. Perltidy will try to find line break points to keep lines below this length. However, long quotes and side comments may cause lines to exceed this length. The default length of 80 comes from the past when this was the standard CRT screen width. Many programmers prefer to increase this to something like 120. Setting B<-l=0> is equivalent to setting B<-l=(a very large number)>. But this is not recommended because, for example, a very long list will be formatted in a single long line. =item B<-vmll>, B<--variable-maximum-line-length> A problem arises using a fixed maximum line length with very deeply nested code and data structures because eventually the amount of leading whitespace used for indicating indentation takes up most or all of the available line width, leaving little or no space for the actual code or data. One solution is to use a very long line length. Another solution is to use the B<-vmll> flag, which basically tells perltidy to ignore leading whitespace when measuring the line length. To be precise, when the B<-vmll> parameter is set, the maximum line length of a line of code will be M+L*I, where M is the value of --maximum-line-length=M (-l=M), default 80, I is the value of --indent-columns=I (-i=I), default 4, L is the indentation level of the line of code When this flag is set, the choice of breakpoints for a block of code should be essentially independent of its nesting depth. However, the absolute line lengths, including leading whitespace, can still be arbitrarily large. This problem can be avoided by including the next parameter. The default is not to do this (B<-nvmll>). =item B<-wc=n>, B<--whitespace-cycle=n> This flag also addresses problems with very deeply nested code and data structures. When the nesting depth exceeds the value B the leading whitespace will be reduced and start at a depth of 1 again. The result is that blocks of code will shift back to the left rather than moving arbitrarily far to the right. This occurs cyclically to any depth. For example if one level of indentation equals 4 spaces (B<-i=4>, the default), and one uses B<-wc=15>, then if the leading whitespace on a line exceeds about 4*15=60 spaces it will be reduced back to 4*1=4 spaces and continue increasing from there. If the whitespace never exceeds this limit the formatting remains unchanged. The combination of B<-vmll> and B<-wc=n> provides a solution to the problem of displaying arbitrarily deep data structures and code in a finite window, although B<-wc=n> may of course be used without B<-vmll>. The default is not to use this, which can also be indicated using B<-wc=0>. =item B Using tab characters will almost certainly lead to future portability and maintenance problems, so the default and recommendation is not to use them. For those who prefer tabs, however, there are two different options. Except for possibly introducing tab indentation characters, as outlined below, perltidy does not introduce any tab characters into your file, and it removes any tabs from the code (unless requested not to do so with B<-fws>). If you have any tabs in your comments, quotes, or here-documents, they will remain. =over 4 =item B<-et=n>, B<--entab-leading-whitespace> This flag causes each B leading space characters produced by the formatting process to be replaced by one tab character. The formatting process itself works with space characters. The B<-et=n> parameter is applied as a last step, after formatting is complete, to convert leading spaces into tabs. Before starting to use tabs, it is essential to first get the indentation controls set as desired without tabs, particularly the two parameters B<--indent-columns=n> (or B<-i=n>) and B<--continuation-indentation=n> (or B<-ci=n>). The value of the integer B can be any value but can be coordinated with the number of spaces used for indentation. For example, B<-et=4 -ci=4 -i=4> will produce one tab for each indentation level and and one for each continuation indentation level. You may want to coordinate the value of B with what your display software assumes for the spacing of a tab. The default is not to use this, which can also be indicated using B<-et=0>. =item B<-t>, B<--tabs> This flag causes one leading tab character to be inserted for each level of indentation. Certain other features are incompatible with this option, and if these options are also given, then a warning message will be issued and this flag will be unset. One example is the B<-lp> option. This flag is retained for backwards compatibility, but if you use tabs, the B<-et=n> flag is recommended. If both B<-t> and B<-et=n> are set, the B<-et=n> is used. =item B<-dt=n>, B<--default-tabsize=n> If the first line of code passed to perltidy contains leading tabs but no tab scheme is specified for the output stream then perltidy must guess how many spaces correspond to each leading tab. This number of spaces B corresponding to each leading tab of the input stream may be specified with B<-dt=n>. The default is B. This flag has no effect if a tab scheme is specified for the output stream, because then the input stream is assumed to use the same tab scheme and indentation spaces as for the output stream (any other assumption would lead to unstable editing). =back =item B<-io>, B<--indent-only> This flag is used to deactivate all whitespace and line break changes within non-blank lines of code. When it is in effect, the only change to the script will be to the indentation and to the number of blank lines. And any flags controlling whitespace and newlines will be ignored. You might want to use this if you are perfectly happy with your whitespace and line breaks, and merely want perltidy to handle the indentation. (This also speeds up perltidy by well over a factor of two, so it might be useful when perltidy is merely being used to help find a brace error in a large script). Setting this flag is equivalent to setting B<--freeze-newlines> and B<--freeze-whitespace>. If you also want to keep your existing blank lines exactly as they are, you can add B<--freeze-blank-lines>. With this option perltidy is still free to modify the indenting (and outdenting) of code and comments as it normally would. If you also want to prevent long comment lines from being outdented, you can add either B<-noll> or B<-l=0>. Setting this flag will prevent perltidy from doing any special operations on closing side comments. You may still delete all side comments however when this flag is in effect. =item B<-enc=s>, B<--character-encoding=s> This flag indicates if the input data stream uses a character encoding. Perltidy does not look for the encoding directives in the source stream, such as B, and instead relies on this flag to determine the encoding. (This is because perltidy often works on snippets of code rather than complete files, so it cannot rely on B directives). Consequently perltidy is likely to encounter problems formatting a file which is only partially encoded. The possible values for B are: -enc=none if no encoding is used, or -enc=utf8 for encoding in utf8 -enc=guess if perltidy should guess between these two possibilities. The value B causes the stream to be processed without special encoding assumptions. This is appropriate for files which are written in single-byte character encodings such as latin-1. The value B causes the stream to be read and written as UTF-8. If the input stream cannot be decoded with this encoding then processing is not done. The value B tells perltidy to guess between either utf8 encoding or no encoding (meaning one character per byte). The B option uses the Encode::Guess module which has been found to be reliable at detecting if a file is encoded in utf8 or not. The current default is B. The abbreviations B<-utf8> or B<-UTF8> are equivalent to B<-enc=utf8>, and the abbreviation B<-guess> is equivalent to B<-enc=guess>. So to process a file named B which is encoded in UTF-8 you can use: perltidy -utf8 file.pl or perltidy -guess file.pl or simply perltidy file.pl since B<-guess> is the default. To process files with an encoding other than UTF-8, it would be necessary to write a short program which calls the Perl::Tidy module with some pre- and post-processing to handle decoding and encoding. =item B<-eos=s>, B<--encode-output-strings=s> This flag was added to resolve an issue involving the interface between Perl::Tidy and calling programs, and in particular B. If you only run the B binary this flag has no effect. If you run a program which calls the Perl::Tidy module and receives a string in return, then the meaning of the flag is as follows: =over 4 =item * The setting B<-eos> means Perl::Tidy should encode any string which it decodes. This is the default because it makes perltidy behave well as a filter, and is the correct setting for most programs. =item * The setting B<-neos> means that a string should remain decoded if it was decoded by Perl::Tidy. This is only appropriate if the calling program will handle any needed encoding before outputting the string. =back The default was changed from B<-neos> to B<-eos> in versions after 20220217. If this change causes a program to start running incorrectly on encoded files, an emergency fix might be to set B<-neos>. Additional information can be found in the man pages for the B module and also in L. =item B<-gcs>, B<--use-unicode-gcstring> This flag controls whether or not perltidy may use module Unicode::GCString to obtain accurate display widths of wide characters. The default is B<--nouse-unicode-gcstring>. If this flag is set, and text is encoded, perltidy will look for the module Unicode::GCString and, if found, will use it to obtain character display widths. This can improve displayed vertical alignment for files with wide characters. It is a nice feature but it is off by default to avoid conflicting formatting when there are multiple developers. Perltidy installation does not require Unicode::GCString, so users wanting to use this feature need set this flag and also to install Unicode::GCString separately. If this flag is set and perltidy does not find module Unicode::GCString, a warning message will be produced and processing will continue but without the potential benefit provided by the module. Also note that actual vertical alignment depends upon the fonts used by the text display software, so vertical alignment may not be optimal even when Unicode::GCString is used. =item B<-ole=s>, B<--output-line-ending=s> where s=C, C, C, or C. This flag tells perltidy to output line endings for a specific system. Normally, perltidy writes files with the line separator character of the host system. The C and C flags have an identical result. =item B<-ple>, B<--preserve-line-endings> This flag tells perltidy to write its output files with the same line endings as the input file, if possible. It should work for B, B, and B line endings. It will only work if perltidy input comes from a filename (rather than stdin, for example). If perltidy has trouble determining the input file line ending, it will revert to the default behavior of using the line ending of the host system. =item B<-atnl>, B<--add-terminal-newline> This flag, which is enabled by default, allows perltidy to terminate the last line of the output stream with a newline character, regardless of whether or not the input stream was terminated with a newline character. If this flag is negated, with B<-natnl>, then perltidy will add a terminal newline to the the output stream only if the input stream is terminated with a newline. Negating this flag may be useful for manipulating one-line scripts intended for use on a command line. =item B<-it=n>, B<--iterations=n> This flag causes perltidy to do B complete iterations. The reason for this flag is that code beautification is an iterative process and in some cases the output from perltidy can be different if it is applied a second time. For most purposes the default of B should be satisfactory. However B can be useful when a major style change is being made, or when code is being beautified on check-in to a source code control system. It has been found to be extremely rare for the output to change after 2 iterations. If a value B is greater than 2 is input then a convergence test will be used to stop the iterations as soon as possible, almost always after 2 iterations. See the next item for a simplified iteration control. This flag has no effect when perltidy is used to generate html. =item B<-conv>, B<--converge> This flag is equivalent to B<-it=4> and is included to simplify iteration control. For all practical purposes one either does or does not want to be sure that the output is converged, and there is no penalty to using a large iteration limit since perltidy will check for convergence and stop iterating as soon as possible. The default is B<-nconv> (no convergence check). Using B<-conv> will approximately double run time since typically one extra iteration is required to verify convergence. No extra iterations are required if no new line breaks are made, and two extra iterations are occasionally needed when reformatting complex code structures, such as deeply nested ternary statements. =back =head2 Code Indentation Control =over 4 =item B<-ci=n>, B<--continuation-indentation=n> Continuation indentation is extra indentation spaces applied when a long line is broken. The default is n=2, illustrated here: my $level = # -ci=2 ( $max_index_to_go >= 0 ) ? $levels_to_go[0] : $last_output_level; The same example, with n=0, is a little harder to read: my $level = # -ci=0 ( $max_index_to_go >= 0 ) ? $levels_to_go[0] : $last_output_level; The value given to B<-ci> is also used by some commands when a small space is required. Examples are commands for outdenting labels, B<-ola>, and control keywords, B<-okw>. When default values are not used, it is recommended that either (1) the value B given with B<-ci=n> be no more than about one-half of the number of spaces assigned to a full indentation level on the B<-i=n> command, or (2) the flag B<-extended-continuation-indentation> is used (see next section). =item B<-xci>, B<--extended-continuation-indentation> This flag allows perltidy to use some improvements which have been made to its indentation model. One of the things it does is "extend" continuation indentation deeper into structures, hence the name. The improved indentation is particularly noticeable when the flags B<-ci=n> and B<-i=n> use the same value of B. There are no significant disadvantages to using this flag, but to avoid disturbing existing formatting the default is not to use it, B<-nxci>. Please see the section L<"B<-pbp>, B<--perl-best-practices>"> for an example of how this flag can improve the formatting of ternary statements. It can also improve indentation of some multi-line qw lists as shown below. # perltidy foreach $color ( qw( AntiqueWhite3 Bisque1 Bisque2 Bisque3 Bisque4 SlateBlue3 RoyalBlue1 SteelBlue2 DeepSkyBlue3 ), qw( LightBlue1 DarkSlateGray1 Aquamarine2 DarkSeaGreen2 SeaGreen1 Yellow1 IndianRed1 IndianRed2 Tan1 Tan4 ) ) # perltidy -xci foreach $color ( qw( AntiqueWhite3 Bisque1 Bisque2 Bisque3 Bisque4 SlateBlue3 RoyalBlue1 SteelBlue2 DeepSkyBlue3 ), qw( LightBlue1 DarkSlateGray1 Aquamarine2 DarkSeaGreen2 SeaGreen1 Yellow1 IndianRed1 IndianRed2 Tan1 Tan4 ) ) =item B<-sil=n> B<--starting-indentation-level=n> By default, perltidy examines the input file and tries to determine the starting indentation level. While it is often zero, it may not be zero for a code snippet being sent from an editing session. To guess the starting indentation level perltidy simply assumes that indentation scheme used to create the code snippet is the same as is being used for the current perltidy process. This is the only sensible guess that can be made. It should be correct if this is true, but otherwise it probably won't. For example, if the input script was written with -i=2 and the current perltidy flags have -i=4, the wrong initial indentation will be guessed for a code snippet which has non-zero initial indentation. Likewise, if an entabbing scheme is used in the input script and not in the current process then the guessed indentation will be wrong. If the default method does not work correctly, or you want to change the starting level, use B<-sil=n>, to force the starting level to be n. =item B using B<--line-up-parentheses>, B<-lp> or B<--extended--line-up-parentheses> , B<-xlp> These flags provide an alternative indentation method for list data. The original flag for this is B<-lp>, but it has some limitations (explained below) which are avoided with the newer B<-xlp> flag. So B<-xlp> is probably the better choice for new work, but the B<-lp> flag is retained to minimize changes to existing formatting. If you enter both B<-lp> and B<-xlp>, then B<-xlp> will be used. In the default indentation method perltidy indents lists with 4 spaces, or whatever value is specified with B<-i=n>. Here is a small list formatted in this way: # perltidy (default) @month_of_year = ( 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec' ); The B<-lp> or B<-xlp> flags add extra indentation to cause the data to begin past the opening parentheses of a sub call or list, or opening square bracket of an anonymous array, or opening curly brace of an anonymous hash. With this option, the above list would become: # perltidy -lp or -xlp @month_of_year = ( 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec' ); If the available line length (see B<-l=n> ) does not permit this much space, perltidy will use less. For alternate placement of the closing paren, see the next section. These flags have no effect on code BLOCKS, such as if/then/else blocks, which always use whatever is specified with B<-i=n>. Some limitations on these flags are: =over 4 =item * A limitation on B<-lp>, but not B<-xlp>, occurs in situations where perltidy does not have complete freedom to choose line breaks. Then it may temporarily revert to its default indentation method. This can occur for example if there are blank lines, block comments, multi-line quotes, or side comments between the opening and closing parens, braces, or brackets. It will also occur if a multi-line anonymous sub occurs within a container since that will impose specific line breaks (such as line breaks after statements). =item * For both the B<-lp> and B<-xlp> flags, any parameter which significantly restricts the ability of perltidy to choose newlines will conflict with these flags and will cause them to be deactivated. These include B<-io>, B<-fnl>, B<-nanl>, and B<-ndnl>. =item * The B<-lp> and B<-xlp> options may not be used together with the B<-t> tabs option. They may, however, be used with the B<-et=n> tab method =back There are some potential disadvantages of this indentation method compared to the default method that should be noted: =over 4 =item * The available line length can quickly be used up if variable names are long. This can cause deeply nested code to quickly reach the line length limit, and become badly formatted, much sooner than would occur with the default indentation method. =item * Since the indentation depends on the lengths of variable names, small changes in variable names can cause changes in indentation over many lines in a file. This means that minor name changes can produce significant file differences. This can be annoying and does not occur with the default indentation method. =back Some things that can be done to minimize these problems are: =over 4 =item * Increase B<--maximum-line-length=n> above the default B characters if necessary. =item * If you use B<-xlp> then long side comments can limit the indentation over multiple lines. Consider adding the flag B<--ignore-side-comment-lengths> to prevent this, or minimizing the use of side comments. =item * Apply this style in a limited way. By default, it applies to all list containers (not just lists in parentheses). The next section describes how to limit this style to, for example, just function calls. The default indentation method will be applied elsewhere. =back =item B<-lpil=s>, B<--line-up-parentheses-inclusion-list> and B<-lpxl=s>, B<--line-up-parentheses-exclusion-list> The following discussion is written for B<-lp> but applies equally to the newer B<-xlp> version. By default, the B<-lp> flag applies to as many containers as possible. The set of containers to which the B<-lp> style applies can be reduced by either one of these two flags: Use B<-lpil=s> to specify the containers to which B<-lp> applies, or use B<-lpxl=s> to specify the containers to which B<-lp> does NOT apply. Only one of these two flags may be used. Both flags can achieve the same result, but the B<-lpil=s> flag is much easier to describe and use and is recommended. The B<-lpxl=s> flag was the original implementation and is only retained for backwards compatibility. This list B for these parameters is a string with space-separated items. Each item consists of up to three pieces of information in this order: (1) an optional letter code (2) a required container type, and (3) an optional numeric code. The only required piece of information is a container type, which is one of '(', '[', or '{'. For example the string -lpil='(' means use -lp formatting only on lists within parentheses, not lists in square-brackets or braces. The same thing could alternatively be specified with -lpxl = '[ {' which says to exclude lists within square-brackets and braces. So what remains is lists within parentheses. A second optional item of information which can be given for parentheses is an alphanumeric letter which is used to limit the selection further depending on the type of token immediately before the paren. The possible letters are currently 'k', 'K', 'f', 'F', 'w', and 'W', with these meanings for matching whatever precedes an opening paren: 'k' matches if the previous nonblank token is a perl built-in keyword (such as 'if', 'while'), 'K' matches if 'k' does not, meaning that the previous token is not a keyword. 'f' matches if the previous token is a function other than a keyword. 'F' matches if 'f' does not. 'w' matches if either 'k' or 'f' match. 'W' matches if 'w' does not. For example: -lpil = 'f(' means only apply -lp to function calls, and -lpil = 'w(' means only apply -lp to parenthesized lists which follow a function or a keyword. This last example could alternatively be written using the B<-lpxl=s> flag as -lpxl = '[ { W(' which says exclude B<-lp> for lists within square-brackets, braces, and parens NOT preceded by a keyword or function. Clearly, the B<-lpil=s> method is easier to understand. An optional numeric code may follow any of the container types to further refine the selection based on container contents. The numeric codes are: '0' or blank: no check on contents is made '1' exclude B<-lp> unless the contents is a simple list without sublists '2' exclude B<-lp> unless the contents is a simple list without sublists, without code blocks, and without ternary operators For example, -lpil = 'f(2' means only apply -lp to function call lists which do not contain any sublists, code blocks or ternary expressions. =item B<-cti=n>, B<--closing-token-indentation> The B<-cti=n> flag controls the indentation of a line beginning with a C<)>, C<]>, or a non-block C<}>. Such a line receives: -cti = 0 no extra indentation (default) -cti = 1 extra indentation such that the closing token aligns with its opening token. -cti = 2 one extra indentation level if the line looks like: ); or ]; or }; -cti = 3 one extra indentation level always The flags B<-cti=1> and B<-cti=2> work well with the B<-lp> flag (previous section). # perltidy -lp -cti=1 @month_of_year = ( 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec' ); # perltidy -lp -cti=2 @month_of_year = ( 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec' ); These flags are merely hints to the formatter and they may not always be followed. In particular, if -lp is not being used, the indentation for B is constrained to be no more than one indentation level. If desired, this control can be applied independently to each of the closing container token types. In fact, B<-cti=n> is merely an abbreviation for B<-cpi=n -csbi=n -cbi=n>, where: B<-cpi> or B<--closing-paren-indentation> controls B<)>'s, B<-csbi> or B<--closing-square-bracket-indentation> controls B<]>'s, B<-cbi> or B<--closing-brace-indentation> controls non-block B<}>'s. =item B<-icp>, B<--indent-closing-paren> The B<-icp> flag is equivalent to B<-cti=2>, described in the previous section. The B<-nicp> flag is equivalent B<-cti=0>. They are included for backwards compatibility. =item B<-icb>, B<--indent-closing-brace> The B<-icb> option gives one extra level of indentation to a brace which terminates a code block . For example, if ($task) { yyy(); } # -icb else { zzz(); } The default is not to do this, indicated by B<-nicb>. =item B<-nib>, B<--non-indenting-braces> Normally, lines of code contained within a pair of block braces receive one additional level of indentation. This flag, which is enabled by default, causes perltidy to look for opening block braces which are followed by a special side comment. This special side comment is B<#<<<> by default. If found, the code between this opening brace and its corresponding closing brace will not be given the normal extra indentation level. For example: { #<<< a closure to contain lexical vars my $var; # this line does not get one level of indentation ... } # this line does not 'see' $var; This can be useful, for example, when combining code from different files. Different sections of code can be placed within braces to keep their lexical variables from being visible to the end of the file. To keep the new braces from causing all of their contained code to be indented if you run perltidy, and possibly introducing new line breaks in long lines, you can mark the opening braces with this special side comment. Only the opening brace needs to be marked, since perltidy knows where the closing brace is. Braces contained within marked braces may also be marked as non-indenting. If your code happens to have some opening braces followed by '#<<<', and you don't want this behavior, you can use B<-nnib> to deactivate it. To make it easy to remember, the default string is the same as the string for starting a B section. There is no confusion because in that case it is for a block comment rather than a side-comment. The special side comment can be changed with the next parameter. =item B<-nibp=s>, B<--non-indenting-brace-prefix=s> The B<-nibp=string> parameter may be used to change the marker for non-indenting braces. The default is equivalent to -nibp='#<<<'. The string that you enter must begin with a # and should be in quotes as necessary to get past the command shell of your system. This string is the leading text of a regex pattern that is constructed by prepending a '^' and appending a'\s', so you must also include backslashes for characters to be taken literally rather than as patterns. For example, to match the side comment '#++', the parameter would be -nibp='#\+\+' =item B<-olq>, B<--outdent-long-quotes> When B<-olq> is set, lines which is a quoted string longer than the value B will have their indentation removed to make them more readable. This is the default. To prevent such out-denting, use B<-nolq> or B<--nooutdent-long-lines>. =item B<-oll>, B<--outdent-long-lines> This command is equivalent to B<--outdent-long-quotes> and B<--outdent-long-comments>, and it is included for compatibility with previous versions of perltidy. The negation of this also works, B<-noll> or B<--nooutdent-long-lines>, and is equivalent to setting B<-nolq> and B<-nolc>. =item B B<-ola>, B<--outdent-labels> This command will cause labels to be outdented by 2 spaces (or whatever B<-ci> has been set to), if possible. This is the default. For example: my $i; LOOP: while ( $i = ) { chomp($i); next unless $i; fixit($i); } Use B<-nola> to not outdent labels. To control line breaks after labels see L<"-bal=n, --break-after-labels=n">. =item B =over 4 =item B<-okw>, B<--outdent-keywords> The command B<-okw> will cause certain leading control keywords to be outdented by 2 spaces (or whatever B<-ci> has been set to), if possible. By default, these keywords are C, C, C, C, and C. The intention is to make these control keywords easier to see. To change this list of keywords being outdented, see the next section. For example, using C on the previous example gives: my $i; LOOP: while ( $i = ) { chomp($i); next unless $i; fixit($i); } The default is not to do this. =item B B<-okwl=string>, B<--outdent-keyword-list=string> This command can be used to change the keywords which are outdented with the B<-okw> command. The parameter B is a required list of perl keywords, which should be placed in quotes if there are more than one. By itself, it does not cause any outdenting to occur, so the B<-okw> command is still required. For example, the commands C<-okwl="next last redo goto" -okw> will cause those four keywords to be outdented. It is probably simplest to place any B<-okwl> command in a F<.perltidyrc> file. =back =back =head2 Whitespace Control Whitespace refers to the blank space between variables, operators, and other code tokens. =over 4 =item B<-fws>, B<--freeze-whitespace> This flag causes your original whitespace to remain unchanged, and causes the rest of the whitespace commands in this section, the Code Indentation section, and the Comment Control section to be ignored. =item B Here the term "tightness" will mean the closeness with which pairs of enclosing tokens, such as parentheses, contain the quantities within. A numerical value of 0, 1, or 2 defines the tightness, with 0 being least tight and 2 being most tight. Spaces within containers are always symmetric, so if there is a space after a C<(> then there will be a space before the corresponding C<)>. The B<-pt=n> or B<--paren-tightness=n> parameter controls the space within parens. The example below shows the effect of the three possible values, 0, 1, and 2: if ( ( my $len_tab = length( $tabstr ) ) > 0 ) { # -pt=0 if ( ( my $len_tab = length($tabstr) ) > 0 ) { # -pt=1 (default) if ((my $len_tab = length($tabstr)) > 0) { # -pt=2 When n is 0, there is always a space to the right of a '(' and to the left of a ')'. For n=2 there is never a space. For n=1, the default, there is a space unless the quantity within the parens is a single token, such as an identifier or quoted string. Likewise, the parameter B<-sbt=n> or B<--square-bracket-tightness=n> controls the space within square brackets, as illustrated below. $width = $col[ $j + $k ] - $col[ $j ]; # -sbt=0 $width = $col[ $j + $k ] - $col[$j]; # -sbt=1 (default) $width = $col[$j + $k] - $col[$j]; # -sbt=2 Curly braces which do not contain code blocks are controlled by the parameter B<-bt=n> or B<--brace-tightness=n>. $obj->{ $parsed_sql->{ 'table' }[0] }; # -bt=0 $obj->{ $parsed_sql->{'table'}[0] }; # -bt=1 (default) $obj->{$parsed_sql->{'table'}[0]}; # -bt=2 And finally, curly braces which contain blocks of code are controlled by the parameter B<-bbt=n> or B<--block-brace-tightness=n> as illustrated in the example below. %bf = map { $_ => -M $_ } grep { /\.deb$/ } dirents '.'; # -bbt=0 (default) %bf = map { $_ => -M $_ } grep {/\.deb$/} dirents '.'; # -bbt=1 %bf = map {$_ => -M $_} grep {/\.deb$/} dirents '.'; # -bbt=2 To simplify input in the case that all of the tightness flags have the same value , the parameter <-act=n> or B<--all-containers-tightness=n> is an abbreviation for the combination <-pt=n -sbt=n -bt=n -bbt=n>. =item B<-xbt>, B<--extended-block-tightness> The previous section described two controls for spacing within curly braces, namely B<-block-brace-tightness=n> for code block braces and B<-brace-tightness=n> for all other braces. There is a little fuzziness in this division of brace types though because the curly braces considered by perltidy to contain code blocks for formatting purposes, such as highlighting code structure, exclude some of the small code blocks used by Perl mainly for isolating terms. These include curly braces following a keyword where an indirect object might occur, or curly braces following a type symbol. For example, perltidy does not mark the following braces as code block braces: print {*STDERR} $message; return ${$foo}; Consequently, the spacing within these small braced containers by default follows the flag B<--brace-tightness=n> rather than B<--block-brace-tightness=n>, as one might expect. If desired, small blocks such as these can be made to instead follow the spacing defined by the B<--block-brace-tightness=n> flag by setting B<--extended-block-tightness>. The specific types of small blocks to which this parameter applies is controlled by a companion control parameter, described in the next section. Note that if the two flags B<-bbt=n> and B<-bt=n> have the same value B then there would be no reason to set this flag. =item B<-xbtl=s>, B<--extended-block-tightness-list=s> The previous parameter B<-xbt> can be made to apply to curly braces preceded by any of the keywords print printf exec system say and/or the special symbols $ @ % & * $# The parameter string B may contain a selection of these keywords and symbols to indicate the brace types to which B<-xbt> applies. For convenience, all of the keywords can be selected with 'k', and all of the special symbols can be selected with 't'. The default is equivalent to B<-xbtl='k'>, which selects all of the keywords. Examples: -xbtl='k' # selects just the keywords [DEFAULT] -xbtl="t" # selects just the special type symbols -xbtl="k t" # selects all keywords and symbols, or more simply -xbtl="kt" # selects all keywords and symbols -xbtl="print say" # selects just keywords B and B: Here are some formatting examples using the default values of B<-bt=n> and B<-bbt=n>. Note that in these examples B<$ref> is in block braces but B<$key> is not. # default formatting print {*STDERR} $message; my $val = ${$ref}{$key}; # perltidy -xbt or # perltidy -xbt -xbtl=k print { *STDERR } $message; my $val = ${$ref}{$key}; # perltidy -xbt -xbtl=t print {*STDERR} $message; my $val = ${ $ref }{$key}; # perltidy -xbt -xbtl=kt print { *STDERR } $message; my $val = ${ $ref }{$key}; Finally, note that this parameter merely changes the way that the parameter B<--extended-block-tightness> works. It has no effect unless B<--extended-block-tightness> is actually set. =item B<-tso>, B<--tight-secret-operators> The flag B<-tso> causes certain perl token sequences (secret operators) which might be considered to be a single operator to be formatted "tightly" (without spaces). The operators currently modified by this flag are: 0+ +0 ()x!! ~~<> ,=> =( )= For example the sequence B<0 +>, which converts a string to a number, would be formatted without a space: B<0+> when the B<-tso> flag is set. This flag is off by default. =item B<-sts>, B<--space-terminal-semicolon> Some programmers prefer a space before all terminal semicolons. The default is for no such space, and is indicated with B<-nsts> or B<--nospace-terminal-semicolon>. $i = 1 ; # -sts $i = 1; # -nsts (default) =item B<-sfs>, B<--space-for-semicolon> Semicolons within B loops may sometimes be hard to see, particularly when commas are also present. This option places spaces on both sides of these special semicolons, and is the default. Use B<-nsfs> or B<--nospace-for-semicolon> to deactivate it. for ( @a = @$ap, $u = shift @a ; @a ; $u = $v ) { # -sfs (default) for ( @a = @$ap, $u = shift @a; @a; $u = $v ) { # -nsfs =item B<-asc>, B<--add-semicolons> Setting B<-asc> allows perltidy to add any missing optional semicolon at the end of a line which is followed by a closing curly brace on the next line. This is the default, and may be deactivated with B<-nasc> or B<--noadd-semicolons>. =item B<-dsm>, B<--delete-semicolons> Setting B<-dsm> allows perltidy to delete extra semicolons which are simply empty statements. This is the default, and may be deactivated with B<-ndsm> or B<--nodelete-semicolons>. (Such semicolons are not deleted, however, if they would promote a side comment to a block comment). =item B<-aws>, B<--add-whitespace> Setting this option allows perltidy to add certain whitespace to improve code readability. This is the default. If you do not want any whitespace added, but are willing to have some whitespace deleted, use B<-naws>. (Use B<-fws> to leave whitespace completely unchanged). =item B<-dws>, B<--delete-old-whitespace> Setting this option allows perltidy to remove some old whitespace between characters, if necessary. This is the default. If you do not want any old whitespace removed, use B<-ndws> or B<--nodelete-old-whitespace>. =item B For those who want more detailed control over the whitespace around tokens, there are four parameters which can directly modify the default whitespace rules built into perltidy for any token. They are: B<-wls=s> or B<--want-left-space=s>, B<-nwls=s> or B<--nowant-left-space=s>, B<-wrs=s> or B<--want-right-space=s>, B<-nwrs=s> or B<--nowant-right-space=s>. These parameters are each followed by a quoted string, B, containing a list of token types. No more than one of each of these parameters should be specified, because repeating a command-line parameter always overwrites the previous one before perltidy ever sees it. To illustrate how these are used, suppose it is desired that there be no space on either side of the token types B<= + - / *>. The following two parameters would specify this desire: -nwls="= + - / *" -nwrs="= + - / *" (Note that the token types are in quotes, and that they are separated by spaces). With these modified whitespace rules, the following line of math: $root = -$b + sqrt( $b * $b - 4. * $a * $c ) / ( 2. * $a ); becomes this: $root=-$b+sqrt( $b*$b-4.*$a*$c )/( 2.*$a ); These parameters should be considered to be hints to perltidy rather than fixed rules, because perltidy must try to resolve conflicts that arise between them and all of the other rules that it uses. One conflict that can arise is if, between two tokens, the left token wants a space and the right one doesn't. In this case, the token not wanting a space takes priority. It is necessary to have a list of all token types in order to create this type of input. Such a list can be obtained by the command B<--dump-token-types>. Also try the B<-D> flag on a short snippet of code and look at the .DEBUG file to see the tokenization. B Be sure to put these tokens in quotes to avoid having them misinterpreted by your command shell. =item B The various parameters controlling whitespace within a program are requests which perltidy follows as well as possible, but there are a number of situations where changing whitespace could change program behavior and is not done. Some of these are obvious; for example, we should not remove the space between the two plus symbols in '$x+ +$y' to avoid creating a '++' operator. Some are more subtle and involve the whitespace around bareword symbols and locations of possible filehandles. For example, consider the problem of formatting the following subroutine: sub print_div { my ($x,$y)=@_; print $x/$y; } Suppose the user requests that / signs have a space to the left but not to the right. Perltidy will refuse to do this, but if this were done the result would be sub print_div { my ($x,$y)=@_; print $x /$y; } If formatted in this way, the program will not run (at least with recent versions of perl) because the $x is taken to be a filehandle and / is assumed to start a quote. In a complex program, there might happen to be a / which terminates the multiline quote without a syntax error, allowing the program to run, but not as intended. Related issues arise with other binary operator symbols, such as + and -, and in older versions of perl there could be problems with ternary operators. So to avoid changing program behavior, perltidy has the simple rule that whitespace around possible filehandles is left unchanged. Likewise, whitespace around barewords is left unchanged. The reason is that if the barewords are defined in other modules, or in code that has not even been written yet, perltidy will not have seen their prototypes and must treat them cautiously. In perltidy this is implemented in the tokenizer by marking token following a B keyword as a special type B. When formatting is being done, whitespace following this token type is generally left unchanged as a precaution against changing program behavior. This is excessively conservative but simple and easy to implement. Keywords which are treated similarly to B include B, B, B, B. Changes in spacing around parameters following these keywords may have to be made manually. For example, the space, or lack of space, after the parameter $foo in the following line will be unchanged in formatting. system($foo ); system($foo); To find if a token is of type B you can use B. For the first line above the result is 1: system($foo ); 1: kkkkkk{ZZZZb}; which shows that B is type B (keyword) and $foo is type B. =item B Despite these precautions, it is still possible to introduce syntax errors with some asymmetric whitespace rules, particularly when call parameters are not placed in containing parens or braces. For example, the following two lines will be parsed by perl without a syntax error: # original programming, syntax ok my @newkeys = map $_-$nrecs+@data, @oldkeys; # perltidy default, syntax ok my @newkeys = map $_ - $nrecs + @data, @oldkeys; But the following will give a syntax error: # perltidy -nwrs='-' my @newkeys = map $_ -$nrecs + @data, @oldkeys; For another example, the following two lines will be parsed without syntax error: # original programming, syntax ok for my $severity ( reverse $SEVERITY_LOWEST+1 .. $SEVERITY_HIGHEST ) { ... } # perltidy default, syntax ok for my $severity ( reverse $SEVERITY_LOWEST + 1 .. $SEVERITY_HIGHEST ) { ... } But the following will give a syntax error: # perltidy -nwrs='+', syntax error: for my $severity ( reverse $SEVERITY_LOWEST +1 .. $SEVERITY_HIGHEST ) { ... } To avoid subtle parsing problems like this, it is best to avoid spacing a binary operator asymmetrically with a space on the left but not on the right. =item B When an opening paren follows a Perl keyword, no space is introduced after the keyword, unless it is (by default) one of these: my local our and or xor eq ne if else elsif until unless while for foreach return switch case given when These defaults can be modified with two commands: B<-sak=s> or B<--space-after-keyword=s> adds keywords. B<-nsak=s> or B<--nospace-after-keyword=s> removes keywords. where B is a list of keywords (in quotes if necessary). For example, my ( $a, $b, $c ) = @_; # default my( $a, $b, $c ) = @_; # -nsak="my local our" The abbreviation B<-nsak='*'> is equivalent to including all of the keywords in the above list. When both B<-nsak=s> and B<-sak=s> commands are included, the B<-nsak=s> command is executed first. For example, to have space after only the keywords (my, local, our) you could use B<-nsak="*" -sak="my local our">. To put a space after all keywords, see the next item. =item B When an opening paren follows a function or keyword, no space is introduced after the keyword except for the keywords noted in the previous item. To always put a space between a function or keyword and its opening paren, use the command: B<-skp> or B<--space-keyword-paren> You may also want to use the flag B<-sfp> (next item) too. =item B When an opening paren follows a function the default and recommended formatting is not to introduce a space. To cause a space to be introduced use: B<-sfp> or B<--space-function-paren> myfunc( $a, $b, $c ); # default myfunc ( $a, $b, $c ); # -sfp You will probably also want to use the flag B<-skp> (previous item) too. The parameter is not recommended because spacing a function paren can make a program vulnerable to parsing problems by Perl. For example, the following two-line program will run as written but will have a syntax error if reformatted with -sfp: if ( -e filename() ) { print "I'm here\n"; } sub filename { return $0 } In this particular case the syntax error can be removed if the line order is reversed, so that Perl parses 'sub filename' first. =item B<-fpva> or B<--function-paren-vertical-alignment> A side-effect of using the B<-sfp> flag is that the parens may become vertically aligned. For example, # perltidy -sfp myfun ( $aaa, $b, $cc ); mylongfun ( $a, $b, $c ); This is the default behavior. To prevent this alignment use B<-nfpva>: # perltidy -sfp -nfpva myfun ( $aaa, $b, $cc ); mylongfun ( $a, $b, $c ); =item B<-spp=n> or B<--space-prototype-paren=n> This flag can be used to control whether a function prototype is preceded by a space. For example, the following prototype does not have a space. sub usage(); This integer B may have the value 0, 1, or 2 as follows: -spp=0 means no space before the paren -spp=1 means follow the example of the source code [DEFAULT] -spp=2 means always put a space before the paren The default is B<-spp=1>, meaning that a space will be used if and only if there is one in the source code. Given the above line of code, the result of applying the different options would be: sub usage(); # n=0 [no space] sub usage(); # n=1 [default; follows input] sub usage (); # n=2 [space] =item B<-ssp=n> or B<--space-signature-paren=n> This flag is analogous to the previous except that it applies to the space before the opening paren of a sub B rather than a sub B. For example, consider the following line: sub circle( $xc, $yc, $rad ) This space before the opening paren can be controlled with integer B which may have the value 0, 1, or 2 with these meanings: -ssp=0 means no space before the paren -ssp=1 means follow the example of the source code [DEFAULT] -ssp=2 means always put a space before the paren The default is B<-ssp=1>, meaning that will be a space in the output if, and only if, there is one in the input. Given the above line of code, the result of applying the different options would be: sub circle( $xc, $yc, $rad ) # n=0 [no space] sub circle( $xc, $yc, $rad ) # n=1 [default; same as input] sub circle ( $xc, $yc, $rad ) # n=2 [space] =item B<-kpit=n> or B<--keyword-paren-inner-tightness=n> The space inside of an opening paren, which itself follows a certain keyword, can be controlled by this parameter. The space on the inside of the corresponding closing paren will be treated in the same (balanced) manner. This parameter has precedence over any other paren spacing rules. The values of B are as follows: -kpit=0 means always put a space (not tight) -kpit=1 means ignore this parameter [default] -kpit=2 means never put a space (tight) To illustrate, the following snippet is shown formatted in three ways: if ( seek( DATA, 0, 0 ) ) { ... } # perltidy (default) if (seek(DATA, 0, 0)) { ... } # perltidy -pt=2 if ( seek(DATA, 0, 0) ) { ... } # perltidy -pt=2 -kpit=0 In the second case the -pt=2 parameter makes all of the parens tight. In the third case the -kpit=0 flag causes the space within the 'if' parens to have a space, since 'if' is one of the keywords to which the -kpit flag applies by default. The remaining parens are still tight because of the -pt=2 parameter. The set of keywords to which this parameter applies are by default are: if elsif unless while until for foreach These can be changed with the parameter B<-kpitl=s> described in the next section. =item B<-kpitl=string> or B<--keyword-paren-inner-tightness=string> This command can be used to change the keywords to which the the B<-kpit=n> command applies. The parameter B is a required list either keywords or functions, which should be placed in quotes if there are more than one. By itself, this parameter does not cause any change in spacing, so the B<-kpit=n> command is still required. For example, the commands C<-kpitl="if else while" -kpit=2> will cause the just the spaces inside parens following 'if', 'else', and 'while' keywords to follow the tightness value indicated by the B<-kpit=2> flag. =item B<-lop> or B<--logical-padding> In the following example some extra space has been inserted on the second line between the two open parens. This extra space is called "logical padding" and is intended to help align similar things vertically in some logical or ternary expressions. # perltidy [default formatting] $same = ( ( $aP eq $bP ) && ( $aS eq $bS ) && ( $aT eq $bT ) && ( $a->{'title'} eq $b->{'title'} ) && ( $a->{'href'} eq $b->{'href'} ) ); Note that this is considered to be a different operation from "vertical alignment" because space at just one line is being adjusted, whereas in "vertical alignment" the spaces at all lines are being adjusted. So it is sort of a local version of vertical alignment. Here is an example involving a ternary operator: # perltidy [default formatting] $bits = $top > 0xffff ? 32 : $top > 0xff ? 16 : $top > 1 ? 8 : 1; This behavior is controlled with the flag B<--logical-padding>, which is set 'on' by default. If it is not desired it can be turned off using B<--nological-padding> or B<-nlop>. The above two examples become, with B<-nlop>: # perltidy -nlop $same = ( ( $aP eq $bP ) && ( $aS eq $bS ) && ( $aT eq $bT ) && ( $a->{'title'} eq $b->{'title'} ) && ( $a->{'href'} eq $b->{'href'} ) ); # perltidy -nlop $bits = $top > 0xffff ? 32 : $top > 0xff ? 16 : $top > 1 ? 8 : 1; =item B quotes> B<-tqw> or B<--trim-qw> provide the default behavior of trimming spaces around multi-line C quotes and indenting them appropriately. B<-ntqw> or B<--notrim-qw> cause leading and trailing whitespace around multi-line C quotes to be left unchanged. This option will not normally be necessary, but was added for testing purposes, because in some versions of perl, trimming C quotes changes the syntax tree. =item B<-sbq=n> or B<--space-backslash-quote=n> lines like $str1=\"string1"; $str2=\'string2'; can confuse syntax highlighters unless a space is included between the backslash and the single or double quotation mark. this can be controlled with the value of B as follows: -sbq=0 means no space between the backslash and quote -sbq=1 means follow the example of the source code -sbq=2 means always put a space between the backslash and quote The default is B<-sbq=1>, meaning that a space will be used if there is one in the source code. =item B B<-trp> or B<--trim-pod> will remove trailing whitespace from lines of POD. The default is not to do this. =back =head2 Comment Controls Perltidy has a number of ways to control the appearance of both block comments and side comments. The term B here refers to a full-line comment, whereas B will refer to a comment which appears on a line to the right of some code. Perltidy does not do any word wrapping of commented text to match a selected maximum line length. This is because there is no way to determine if this is appropriate for the given content. However, an interactive program named B is available in the B folder of the perltidy distribution which can assist in doing this. =over 4 =item B<-ibc>, B<--indent-block-comments> Block comments normally look best when they are indented to the same level as the code which follows them. This is the default behavior, but you may use B<-nibc> to keep block comments left-justified. Here is an example: # this comment is indented (-ibc, default) if ($task) { yyy(); } The alternative is B<-nibc>: # this comment is not indented (-nibc) if ($task) { yyy(); } See also the next item, B<-isbc>, as well as B<-sbc>, for other ways to have some indented and some outdented block comments. =item B<-isbc>, B<--indent-spaced-block-comments> If there is no leading space on the line, then the comment will not be indented, and otherwise it may be. If both B<-ibc> and B<-isbc> are set, then B<-isbc> takes priority. =item B<-olc>, B<--outdent-long-comments> When B<-olc> is set, lines which are full-line (block) comments longer than the value B will have their indentation removed. This is the default; use B<-nolc> to prevent outdenting. =item B<-msc=n>, B<--minimum-space-to-comment=n> Side comments look best when lined up several spaces to the right of code. Perltidy will try to keep comments at least n spaces to the right. The default is n=4 spaces. =item B<-fpsc=n>, B<--fixed-position-side-comment=n> This parameter tells perltidy to line up side comments in column number B whenever possible. The default, n=0, will not do this. =item B<-iscl>, B<--ignore-side-comment-lengths> This parameter causes perltidy to ignore the length of side comments when setting line breaks. The default, B<-niscl>, is to include the length of side comments when breaking lines to stay within the length prescribed by the B<-l=n> maximum line length parameter. For example, the following long single line would remain intact with -l=80 and -iscl: perltidy -l=80 -iscl $vmsfile =~ s/;[\d\-]*$//; # Clip off version number; we can use a newer version as well whereas without the -iscl flag the line will be broken: perltidy -l=80 $vmsfile =~ s/;[\d\-]*$// ; # Clip off version number; we can use a newer version as well =item B<-ipc>, B<--ignore-perlcritic-comments> Perltidy, by default, will look for side comments beginning with B<## no critic> and ignore their lengths when making line break decisions, even if the user has not set B<-iscl>. The reason is that an unwanted line break can make these special comments ineffective in controlling B. Setting B<--ignore-perlcritic-comments> tells perltidy not to look for these B<## no critic> comments. =item B<-hsc>, B<--hanging-side-comments> By default, perltidy tries to identify and align "hanging side comments", which are something like this: my $IGNORE = 0; # This is a side comment # This is a hanging side comment # And so is this A comment is considered to be a hanging side comment if (1) it immediately follows a line with a side comment, or another hanging side comment, and (2) there is some leading whitespace on the line. To deactivate this feature, use B<-nhsc> or B<--nohanging-side-comments>. If block comments are preceded by a blank line, or have no leading whitespace, they will not be mistaken as hanging side comments. =item B A closing side comment is a special comment which perltidy can automatically create and place after the closing brace of a code block. They can be useful for code maintenance and debugging. The command B<-csc> (or B<--closing-side-comments>) adds or updates closing side comments. For example, here is a small code snippet sub message { if ( !defined( $_[0] ) ) { print("Hello, World\n"); } else { print( $_[0], "\n" ); } } And here is the result of processing with C: sub message { if ( !defined( $_[0] ) ) { print("Hello, World\n"); } else { print( $_[0], "\n" ); } } ## end sub message A closing side comment was added for C in this case, but not for the C and C blocks, because they were below the 6 line cutoff limit for adding closing side comments. This limit may be changed with the B<-csci> command, described below. The command B<-dcsc> (or B<--delete-closing-side-comments>) reverses this process and removes these comments. Several commands are available to modify the behavior of these two basic commands, B<-csc> and B<-dcsc>: =over 4 =item B<-csci=n>, or B<--closing-side-comment-interval=n> where C is the minimum number of lines that a block must have in order for a closing side comment to be added. The default value is C. To illustrate: # perltidy -csci=2 -csc sub message { if ( !defined( $_[0] ) ) { print("Hello, World\n"); } ## end if ( !defined( $_[0] )) else { print( $_[0], "\n" ); } ## end else [ if ( !defined( $_[0] )) } ## end sub message Now the C and C blocks are commented. However, now this has become very cluttered. =item B<-cscp=string>, or B<--closing-side-comment-prefix=string> where string is the prefix used before the name of the block type. The default prefix, shown above, is C<## end>. This string will be added to closing side comments, and it will also be used to recognize them in order to update, delete, and format them. Any comment identified as a closing side comment will be placed just a single space to the right of its closing brace. =item B<-cscl=string>, or B<--closing-side-comment-list> where C is a list of block types to be tagged with closing side comments. By default, all code block types preceded by a keyword or label (such as C, C, and so on) will be tagged. The B<-cscl> command changes the default list to be any selected block types; see L<"Specifying Block Types">. For example, the following command requests that only C's, labels, C, and C blocks be affected by any B<-csc> or B<-dcsc> operation: -cscl="sub : BEGIN END" =item B<-csct=n>, or B<--closing-side-comment-maximum-text=n> The text appended to certain block types, such as an C block, is whatever lies between the keyword introducing the block, such as C, and the opening brace. Since this might be too much text for a side comment, there needs to be a limit, and that is the purpose of this parameter. The default value is C, meaning that no additional tokens will be appended to this text after its length reaches 20 characters. Omitted text is indicated with C<...>. (Tokens, including sub names, are never truncated, however, so actual lengths may exceed this). To illustrate, in the above example, the appended text of the first block is C< ( !defined( $_[0] )...>. The existing limit of C caused this text to be truncated, as indicated by the C<...>. See the next flag for additional control of the abbreviated text. =item B<-cscb>, or B<--closing-side-comments-balanced> As discussed in the previous item, when the closing-side-comment-maximum-text limit is exceeded the comment text must be truncated. Older versions of perltidy terminated with three dots, and this can still be achieved with -ncscb: perltidy -csc -ncscb } ## end foreach my $foo (sort { $b cmp $a ... However this causes a problem with editors which cannot recognize comments or are not configured to do so because they cannot "bounce" around in the text correctly. The B<-cscb> flag has been added to help them by appending appropriate balancing structure: perltidy -csc -cscb } ## end foreach my $foo (sort { $b cmp $a ... }) The default is B<-cscb>. =item B<-csce=n>, or B<--closing-side-comment-else-flag=n> The default, B, places the text of the opening C statement after any terminal C. If B is used, then each C is also given the text of the opening C statement. Also, an C will include the text of a preceding C statement. Note that this may result some long closing side comments. If B is used, the results will be the same as B whenever the resulting line length is less than the maximum allowed. =item B<-cscb>, or B<--closing-side-comments-balanced> When using closing-side-comments, and the closing-side-comment-maximum-text limit is exceeded, then the comment text must be abbreviated. It is terminated with three dots if the B<-cscb> flag is negated: perltidy -csc -ncscb } ## end foreach my $foo (sort { $b cmp $a ... This causes a problem with older editors which do not recognize comments because they cannot "bounce" around in the text correctly. The B<-cscb> flag tries to help them by appending appropriate terminal balancing structures: perltidy -csc -cscb } ## end foreach my $foo (sort { $b cmp $a ... }) The default is B<-cscb>. =item B<-cscw>, or B<--closing-side-comment-warnings> This parameter is intended to help make the initial transition to the use of closing side comments. It causes two things to happen if a closing side comment replaces an existing, different closing side comment: first, an error message will be issued, and second, the original side comment will be placed alone on a new specially marked comment line for later attention. The intent is to avoid clobbering existing hand-written side comments which happen to match the pattern of closing side comments. This flag should only be needed on the first run with B<-csc>. =back B =over 4 =item * Closing side comments are only placed on lines terminated with a closing brace. Certain closing styles, such as the use of cuddled elses (B<-ce>), preclude the generation of some closing side comments. =item * Please note that adding or deleting of closing side comments takes place only through the commands B<-csc> or B<-dcsc>. The other commands, if used, merely modify the behavior of these two commands. =item * It is recommended that the B<-cscw> flag be used along with B<-csc> on the first use of perltidy on a given file. This will prevent loss of any existing side comment data which happens to have the csc prefix. =item * Once you use B<-csc>, you should continue to use it so that any closing side comments remain correct as code changes. Otherwise, these comments will become incorrect as the code is updated. =item * If you edit the closing side comments generated by perltidy, you must also change the prefix to be different from the closing side comment prefix. Otherwise, your edits will be lost when you rerun perltidy with B<-csc>. For example, you could simply change C<## end> to be C<## End>, since the test is case sensitive. You may also want to use the B<-ssc> flag to keep these modified closing side comments spaced the same as actual closing side comments. =item * Temporarily generating closing side comments is a useful technique for exploring and/or debugging a perl script, especially one written by someone else. You can always remove them with B<-dcsc>. =back =item B Static block comments are block comments with a special leading pattern, C<##> by default, which will be treated slightly differently from other block comments. They effectively behave as if they had glue along their left and top edges, because they stick to the left edge and previous line when there is no blank spaces in those places. This option is particularly useful for controlling how commented code is displayed. =over 4 =item B<-sbc>, B<--static-block-comments> When B<-sbc> is used, a block comment with a special leading pattern, C<##> by default, will be treated specially. Comments so identified are treated as follows: =over 4 =item * If there is no leading space on the line, then the comment will not be indented, and otherwise it may be, =item * no new blank line will be inserted before such a comment, and =item * such a comment will never become a hanging side comment. =back For example, assuming C<@month_of_year> is left-adjusted: @month_of_year = ( # -sbc (default) 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', ## 'Dec', 'Nov' 'Nov', 'Dec'); Without this convention, the above code would become @month_of_year = ( # -nsbc 'Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', ## 'Dec', 'Nov' 'Nov', 'Dec' ); which is not as clear. The default is to use B<-sbc>. This may be deactivated with B<-nsbc>. =item B<-sbcp=string>, B<--static-block-comment-prefix=string> This parameter defines the prefix used to identify static block comments when the B<-sbc> parameter is set. The default prefix is C<##>, corresponding to C<-sbcp=##>. The prefix is actually part of a perl pattern used to match lines and it must either begin with C<#> or C<^#>. In the first case a prefix ^\s* will be added to match any leading whitespace, while in the second case the pattern will match only comments with no leading whitespace. For example, to identify all comments as static block comments, one would use C<-sbcp=#>. To identify all left-adjusted comments as static block comments, use C<-sbcp='^#'>. Please note that B<-sbcp> merely defines the pattern used to identify static block comments; it will not be used unless the switch B<-sbc> is set. Also, please be aware that since this string is used in a perl regular expression which identifies these comments, it must enable a valid regular expression to be formed. A pattern which can be useful is: -sbcp=^#{2,}[^\s#] This pattern requires a static block comment to have at least one character which is neither a # nor a space. It allows a line containing only '#' characters to be rejected as a static block comment. Such lines are often used at the start and end of header information in subroutines and should not be separated from the intervening comments, which typically begin with just a single '#'. =item B<-osbc>, B<--outdent-static-block-comments> The command B<-osbc> will cause static block comments to be outdented by 2 spaces (or whatever B<-ci=n> has been set to), if possible. =back =item B Static side comments are side comments with a special leading pattern. This option can be useful for controlling how commented code is displayed when it is a side comment. =over 4 =item B<-ssc>, B<--static-side-comments> When B<-ssc> is used, a side comment with a static leading pattern, which is C<##> by default, will be spaced only a single space from previous character, and it will not be vertically aligned with other side comments. The default is B<-nssc>. =item B<-sscp=string>, B<--static-side-comment-prefix=string> This parameter defines the prefix used to identify static side comments when the B<-ssc> parameter is set. The default prefix is C<##>, corresponding to C<-sscp=##>. Please note that B<-sscp> merely defines the pattern used to identify static side comments; it will not be used unless the switch B<-ssc> is set. Also, note that this string is used in a perl regular expression which identifies these comments, so it must enable a valid regular expression to be formed. =back =back =head2 Skipping Selected Sections of Code Selected lines of code may be passed verbatim to the output without any formatting by marking the starting and ending lines with special comments. There are two options for doing this. The first option is called B<--format-skipping> or B<-fs>, and the second option is called B<--code-skipping> or B<-cs>. In both cases the lines of code will be output without any changes. The difference is that in B<--format-skipping> perltidy will still parse the marked lines of code and check for errors, whereas in B<--code-skipping> perltidy will simply pass the lines to the output without any checking. Both of these features are enabled by default and are invoked with special comment markers. B<--format-skipping> uses starting and ending markers '#<<<' and '#>>>', like this: #<<< format skipping: do not let perltidy change my nice formatting my @list = (1, 1, 1, 1, 2, 1, 1, 3, 3, 1, 1, 4, 6, 4, 1,); #>>> B<--code-skipping> uses starting and ending markers '#<>V', like this: #< | _ | ] | <''> ] }; #>>V Additional text may appear on the special comment lines provided that it is separated from the marker by at least one space to highlight the sign, as in the above examples. Any number of code-skipping or format-skipping sections may appear in a file. If an opening code-skipping or format-skipping comment is not followed by a corresponding closing comment, then skipping continues to the end of the file. If a closing code-skipping or format-skipping comment appears in a file but does not follow a corresponding opening comment, then it is treated as an ordinary comment without any special meaning. It is recommended to use B<--code-skipping> only if you need to hide a block of an extended syntax which would produce errors if parsed by perltidy, and use B<--format-skipping> otherwise. This is because the B<--format-skipping> option provides the benefits of error checking, and there are essentially no limitations on which lines to which it can be applied. The B<--code-skipping> option, on the other hand, does not do error checking and its use is more restrictive because the code which remains, after skipping the marked lines, must be syntactically correct code with balanced containers. These features should be used sparingly to avoid littering code with markers, but they can be helpful for working around occasional problems. Note that it may be possible to avoid the use of B<--format-skipping> for the specific case of a comma-separated list of values, as in the above example, by simply inserting a blank or comment somewhere between the opening and closing parens. See the section L<"Controlling List Formatting">. The following sections describe the available controls for these options. They should not normally be needed. =over 4 =item B<-fs>, B<--format-skipping> As explained above, this flag, which is enabled by default, causes any code between special beginning and ending comment markers to be passed to the output without formatting. The code between the comments is still checked for errors however. The default beginning marker is #<<< and the default ending marker is #>>>. Format skipping begins when a format skipping beginning comment is seen and continues until a format-skipping ending comment is found. This feature can be disabled with B<-nfs>. This should not normally be necessary. =item B<-fsb=string>, B<--format-skipping-begin=string> This and the next parameter allow the special beginning and ending comments to be changed. However, it is recommended that they only be changed if there is a conflict between the default values and some other use. If they are used, it is recommended that they only be entered in a B<.perltidyrc> file, rather than on a command line. This is because properly escaping these parameters on a command line can be difficult. If changed comment markers do not appear to be working, use the B<-log> flag and examine the F<.LOG> file to see if and where they are being detected. The B<-fsb=string> parameter may be used to change the beginning marker for format skipping. The default is equivalent to -fsb='#<<<'. The string that you enter must begin with a # and should be in quotes as necessary to get past the command shell of your system. It is actually the leading text of a pattern that is constructed by appending a '\s', so you must also include backslashes for characters to be taken literally rather than as patterns. Some examples show how example strings become patterns: -fsb='#\{\{\{' becomes /^#\{\{\{\s/ which matches #{{{ but not #{{{{ -fsb='#\*\*' becomes /^#\*\*\s/ which matches #** but not #*** -fsb='#\*{2,}' becomes /^#\*{2,}\s/ which matches #** and #***** =item B<-fse=string>, B<--format-skipping-end=string> The B<-fse=string> is the corresponding parameter used to change the ending marker for format skipping. The default is equivalent to -fse='#<<<'. The beginning and ending strings may be the same, but it is preferable to make them different for clarity. =item B<-cs>, B<--code-skipping> As explained above, this flag, which is enabled by default, causes any code between special beginning and ending comment markers to be directly passed to the output without any error checking or formatting. Essentially, perltidy treats it as if it were a block of arbitrary text. The default beginning marker is #<>V. This feature can be disabled with B<-ncs>. This should not normally be necessary. =item B<-csb=string>, B<--code-skipping-begin=string> This may be used to change the beginning comment for a B<--code-skipping> section, and its use is similar to the B<-fsb=string>. The default is equivalent to -csb='#<, B<--code-skipping-end=string> This may be used to change the ending comment for a B<--code-skipping> section, and its use is similar to the B<-fse=string>. The default is equivalent to -cse='#>>V'. =back =head2 Formatting a Limited Range of Lines A command B<--line-range-tidy=n1:n2> is available to process just a selected range of lines of an input stream with perltidy. This command is mainly of interest for programming interactive code editors. When it is used, the entire input stream is read but just the selected range of lines of the input file are processed by the perltidy tokenizer and formatter, and then the stream is reassembled for output. The selected lines need to contain a complete statement or balanced container. Otherwise, a syntax error will occur and the code will not be tidied. There are a couple of limitations on the use of this command: (1) it may not be applied to multiple files, and (2) it only applies to code tidying and not, for example, html formatting. =over 4 =item B<-lrt=n1:n2>, B<--line-range-tidy=n1:n2> The range of lines is specified by integers B and B, where B is the first line number to be formatted (start counting with 1) and B is the last line number to be formatted. If B is not given, or exceeds the actual number of lines, then formatting continues to the end of the file. Examples: -line-range-tidy=43:109 # tidy lines 43 through 109 -line-range-tidy=' 43 : 109' # tidy lines 43 through 109 (spaces ok in quotes) -line-range-tidy=1: # tidy all lines -line-range-tidy=0:90 # ERROR (n1 must be >= 1) =back =head2 Line Break Control The parameters in this and the next sections control breaks after non-blank lines of code. Blank lines are controlled separately by parameters in the section L<"Blank Line Control">. =over 4 =item B<-dnl>, B<--delete-old-newlines> By default, perltidy first deletes all old line break locations, and then it looks for good break points to match the desired line length. Use B<-ndnl> or B<--nodelete-old-newlines> to force perltidy to retain all old line break points. =item B<-anl>, B<--add-newlines> By default, perltidy will add line breaks when necessary to create continuations of long lines and to improve the script appearance. Use B<-nanl> or B<--noadd-newlines> to prevent any new line breaks. This flag does not prevent perltidy from eliminating existing line breaks; see B<--freeze-newlines> to completely prevent changes to line break points. =item B<-fnl>, B<--freeze-newlines> If you do not want any changes to the line breaks within lines of code in your script, set B<-fnl>, and they will remain fixed, and the rest of the commands in this section and sections L<"Controlling List Formatting">, L<"Retaining or Ignoring Existing Line Breaks">. You may want to use B<-noll> with this. Note: If you also want to keep your blank lines exactly as they are, you can use the B<-fbl> flag which is described in the section L<"Blank Line Control">. =back =head2 Controlling Breaks at Braces, Parens, and Square Brackets =over 4 =item B<-ce>, B<--cuddled-else> Enable the "cuddled else" style, in which C and C are follow immediately after the curly brace closing the previous block. The default is not to use cuddled elses, and is indicated with the flag B<-nce> or B<--nocuddled-else>. Here is a comparison of the alternatives: # -ce if ($task) { yyy(); } else { zzz(); } # -nce (default) if ($task) { yyy(); } else { zzz(); } In this example the keyword B is placed on the same line which begins with the preceding closing block brace and is followed by its own opening block brace on the same line. Other keywords and function names which are formatted with this "cuddled" style are B, B, B, B. Other block types can be formatted by specifying their names on a separate parameter B<-cbl>, described in a later section. Cuddling between a pair of code blocks requires that the closing brace of the first block start a new line. If this block is entirely on one line in the input file, it is necessary to decide if it should be broken to allow cuddling. This decision is controlled by the flag B<-cbo=n> discussed below. The default and recommended value of B<-cbo=1> bases this decision on the first block in the chain. If it spans multiple lines then cuddling is made and continues along the chain, regardless of the sizes of subsequent blocks. Otherwise, short lines remain intact. So for example, the B<-ce> flag would not have any effect if the above snippet is rewritten as if ($task) { yyy() } else { zzz() } If the first block spans multiple lines, then cuddling can be done and will continue for the subsequent blocks in the chain, as illustrated in the previous snippet. If there are blank lines between cuddled blocks they will be eliminated. If there are comments after the closing brace where cuddling would occur then cuddling will be prevented. If this occurs, cuddling will restart later in the chain if possible. =item B<-cb>, B<--cuddled-blocks> This flag is equivalent to B<-ce>. =item B<-cbl>, B<--cuddled-block-list> The built-in default cuddled block types are B. Additional block types to which the B<-cuddled-blocks> style applies can be defined by this parameter. This parameter is a character string, giving a list of block types separated by commas or spaces. For example, to cuddle code blocks of type sort, map and grep, in addition to the default types, the string could be set to -cbl="sort map grep" or equivalently -cbl=sort,map,grep Note however that these particular block types are typically short so there might not be much opportunity for the cuddled format style. Using commas avoids the need to protect spaces with quotes. As a diagnostic check, the flag B<--dump-cuddled-block-list> or B<-dcbl> can be used to view the hash of values that are generated by this flag. Finally, note that the B<-cbl> flag by itself merely specifies which blocks are formatted with the cuddled format. It has no effect unless this formatting style is activated with B<-ce>. =item B<-cblx>, B<--cuddled-block-list-exclusive> When cuddled else formatting is selected with B<-ce>, setting this flag causes perltidy to ignore its built-in defaults and rely exclusively on the block types specified on the B<-cbl> flag described in the previous section. For example, to avoid using cuddled B and B, which are among the defaults, the following set of parameters could be used: perltidy -ce -cbl='else elsif continue' -cblx =item B<-cbo=n>, B<--cuddled-break-option=n> Cuddled formatting is only possible between a pair of code blocks if the closing brace of the first block starts a new line. If a block is encountered which is entirely on a single line, and cuddled formatting is selected, it is necessary to make a decision as to whether or not to "break" the block, meaning to cause it to span multiple lines. This parameter controls that decision. The options are: cbo=0 Never force a short block to break. cbo=1 If the first of a pair of blocks is broken in the input file, then break the second [DEFAULT]. cbo=2 Break open all blocks for maximal cuddled formatting. The default and recommended value is B. With this value, if the starting block of a chain spans multiple lines, then a cascade of breaks will occur for remaining blocks causing the entire chain to be cuddled. The option B can produce erratic cuddling if there are numerous one-line blocks. The option B produces maximal cuddling but will not allow any short blocks. =item B<-bl>, B<--opening-brace-on-new-line>, or B<--brace-left> Use the flag B<-bl> to place an opening block brace on a new line: if ( $input_file eq '-' ) { ... } By default it applies to all structural blocks except B and anonymous subs. The default is B<-nbl> which places an opening brace on the same line as the keyword introducing it if possible. For example, # default if ( $input_file eq '-' ) { ... } When B<-bl> is set, the blocks to which this applies can be controlled with the parameters B<--brace-left-list> and B<-brace-left-exclusion-list> described in the next sections. =item B<-bll=s>, B<--brace-left-list=s> Use this parameter to change the types of block braces for which the B<-bl> flag applies; see L<"Specifying Block Types">. For example, B<-bll='if elsif else sub'> would apply it to only C and named sub blocks. The default is all blocks, B<-bll='*'>. =item B<-blxl=s>, B<--brace-left-exclusion-list=s> Use this parameter to exclude types of block braces for which the B<-bl> flag applies; see L<"Specifying Block Types">. For example, the default settings B<-bll='*'> and B<-blxl='sort map grep eval asub'> mean all blocks except B and anonymous sub blocks. Note that the lists B<-bll=s> and B<-blxl=s> control the behavior of the B<-bl> flag but have no effect unless the B<-bl> flag is set. =item B<-sbl>, B<--opening-sub-brace-on-new-line> The flag B<-sbl> provides a shortcut way to turn on B<-bl> just for named subs. The same effect can be achieved by turning on B<-bl> with the block list set as B<-bll='sub'>. For example, perltidy -sbl produces this result: sub message { if (!defined($_[0])) { print("Hello, World\n"); } else { print($_[0], "\n"); } } This flag is negated with B<-nsbl>, which is the default. =item B<-asbl>, B<--opening-anonymous-sub-brace-on-new-line> The flag B<-asbl> is like the B<-sbl> flag except that it applies to anonymous sub's instead of named subs. For example perltidy -asbl produces this result: $a = sub { if ( !defined( $_[0] ) ) { print("Hello, World\n"); } else { print( $_[0], "\n" ); } }; This flag is negated with B<-nasbl>, and the default is B<-nasbl>. =item B<-bli>, B<--brace-left-and-indent> The flag B<-bli> is similar to the B<-bl> flag but in addition it causes one unit of continuation indentation ( see B<-ci> ) to be placed before an opening and closing block braces. For example, perltidy -bli gives if ( $input_file eq '-' ) { important_function(); } By default, this extra indentation occurs for block types: B, B, B, B, B, B, B, B, and also B and blocks preceded by a B