| < podlators README | Russ Allbery > Software > podlators | POD Style Guide > |
Fix removal of scripts/pod2man and scripts/pod2text by make realclean, broken in the v6.0.0 release. Thanks to James E Keenan for the report.
Remove autodie from the module build process. When built as part of Perl core, podlators is built before autodie is available. Thanks to James E Keenan for the report and a draft patch. (GitHub #33)
Drop support for Perl 5.10. podlators now requires Perl 5.12 or later.
podlators now uses semantic versioning for the package and module versions, with a v prefix to work with Perl's packaging system.
Pod::Man now translates all "-" characters in the input into *roff "\-" escapes (normally rendered as an ASCII hyphen-minus, U+002D) rather than using fragile heuristics to decide which characters represent true hyphens and which represent ASCII hyphen-minus. The previous heuristics misrendered command names such as apt-get, causing search and cut-and-paste issues. This change may cause line-break issues with long hyphenated phrases. In cases where the intent is a true hyphen, consider using UTF-8 as the POD character set (declared with =encoding) and using true Unicode hyphens instead of the ASCII "-" character.
Pod::Man now disables the special *roff interpretation of "`" and "'" characters as paired quotes everywhere, not just in verbatim text, thus forcing them to be interpreted as the regular ASCII characters. This also disables the use of "``" and "''" for paired double-quotes. The rationale is similar to that for hyphens: there is no way to tell from the POD source that the special interpretation as quotes is intended. To produce paired typographic quotes in the output, use UTF-8 and Unicode paired quote characters.
Man page references in L<> that are detected as such by Pod::Simple are now always formatted as man page references even if our normal heuristic would not detect them. This fixes the formatting of constructions such as @@RXVT_NAME@@perl(3), which are used by packages that format a man page with POD and then substitute variables into it at build time. Thanks to Marco Sirabella for the analysis and an initial patch. (GitHub #21)
Add a workaround to Pod::Man to force persistent ragged-right justification under nroff with groff 1.23.0. Thanks to Guillem Jover for the report and G. Branden Robinson for the analysis. (GitHub #23)
Fix wrapping of text with S<> markup in all subclasses of Pod::Text. Thanks to Jim Avera for the report. (GitHub #24)
Pod::Man now forces a blank line after a nested list contains only =item tags without bodies. In previous versions, the blank line before the next item in the surrounding =over block was not included. Thanks to Julien ÉLIE for the report. (GitHub #26)
Import PerlIO before checking for layers so that PerlIO::F_UTF8 is available, which fixes double-encoding of output when a :utf8 layer is in place and PerlIO is not imported. Thanks to youpong for the bug report, James Keenan for the elaboration, and Graham Knop for the fix. (GitHub #25)
pod2text --help now exits with status 0, not 1, matching normal UNIX command behavior and the behavior of pod2man. (GitHub #19)
Fix tests when NO_COLOR is set in the environment. (GitHub #20)
Guesswork (formatting rules based on heuristics intended for Perl documentation) can now be disabled or selectively enabled in Pod::Text with the guesswork option (--guesswork to pod2text). This is similar to the support added to Pod::Man in 5.00, except there is only one type of guesswork currently (quoting).
Pod::Text suppresses quote marks around some additional cases of Perl code in C<> where the intent had been to suppress the quotes but there were various bugs in the matching regular expressions. This primarily affects function and method calls.
Fix an escaping issue in Pod::Man with complex lquote, rquote, and quotes options containing double quotes that caused some double quotes to be incorrectly duplicated.
Pod::Man now avoids quoting macro arguments when this is unnecessary.
Depend on Pod::Simple 3.26 or later, since Pod::Text now calls its encoding() method.
Drop support for Perl 5.8. The minimum required version is Perl 5.10.
The default output encoding of Pod::Man on non-EBCDIC systems is now UTF-8. The utf8 option (-u or --utf8 to pod2man) is now ignored, since it is the default. See the ENCODING section of its documentation for testing results and further discussion. (#68741)
Pod::Man now supports an encoding option (-e or --encoding to pod2man), to change the output encoding to any encoding recognized by Encode, or the special values groff or roff. Setting it to roff requests the old behavior of using character substitutions and *roff escapes to generate ASCII-only output (the default prior to this version).
Pod::Man now supports the groff output encoding, which replaces all non-ASCII characters with \[uNNNN] escapes. This escape was not supported by the originally *roff implementation, but it is supported by groff and mandoc and allows proper representation of Unicode characters. This output format has no known advantages in portability over UTF-8 on non-EBCDIC systems. It is the default output format on EBCDIC systems, and when the Encode module is not available. (#73804)
Pod::Man no longer does guesswork transformations that only affected troff output. Formatting manual pages with troff is exceptionally rare, and this magic caused constant maintenance issues. This means "--" is no longer changed to an em-dash, strings of capital letters aren't made a bit smaller, no attempt is made to change double quotes to paired quotes, and no special formatting is done for C++. (#132007)
Guesswork (formatting rules based on heuristics intended for Perl documentation) can now be disabled or selectively enabled in Pod::Man with the guesswork option (--guesswork to pod2man). (#143668)
Pod::Text now supports an encoding option (-e or --encoding to pod2text) to force the output encoding, similar to Pod::Man. The utf8 option (-u or --utf8 to pod2text) is still supported and is equivalent to setting encoding to UTF-8.
Pod::Text now defaults to UTF-8 encoding if it sees a non-ASCII character on a non-EBCDIC system and the input encoding is not specified. This should fix rendering of E<> escapes of non-ASCII characters in POD files that don't specify an input encoding, at the cost of assuming UTF-8 output. Pod::Text also now commits to an encoding the first time it outputs a non-ASCII character and sticks with that encoding for the rest of the file, even if the detected or declared input encoding changes. (#102631)
Stop using a PerlIO encoding layer for Pod::Text output and instead use Encode. If a PerlIO encoding layer is already set, trust it and do no encoding. This fixes encoding problems with output to a string.
Add a coding tag comment to the start of Pod::Man output if the output encoding is not ASCII. groff's preconv program and FreeBSD's mandoc will use this line to determine the input encoding.
Pod::Man now supports a language option (--language to pod2man) that adds groff language configuration commands to the start of the output. This is required for proper line breaking of Japanese and Chinese text when the manual page is not installed in a language-specific directory so that the man program knows to add that configuration automatically. Unfortunately, the commands added when this option is used are groff-specific. Thanks to zynldyx for the bug report and suggested fix.
Pod::Man now converts Unicode zero-width spaces (U+200B) to the *roff escape \:, which indicates a line break point without a space or hyphen. (Unfortunately, groff does not honor U+200B itself as a line break point.) This escape is not part of the language defined in CSTR #54, but is supported by both groff and mandoc, so hopefully use of this escape won't cause problems. (Debian Bug#941980)
Pod::Man and Pod::Text now correctly honor S<> wrapping L<> with an anchor and URL, and make the space between the anchor and URL nonbreaking. (#143768)
Clear the current font before changing fonts in all Pod::Man output, not just headings. groff 1.22.4 does not clear bold or italic when seeing \f(CW font change, which meant bold and italic were extending farther than they should without this change. (#143667)
Honor the quotes, lquote, and rquote configuration parameters to Pod::Man for C<> text inside the special NAME section. (#143967)
Pod::Man suppresses quote marks around some additional cases of Perl code in C<> where the intent had been to suppress the quotes but there were various bugs in the matching regular expressions. This primarily affects method calls and negative numbers.
Avoid non-standard *roff escape in the troff accent mark definition for an acute accent. \h"..." was used instead of \h'...' as seen in the other accent mark definitions. This error appears to have existed since the first version of Pod::Man. Thanks, Paul Evans. (GitHub #14)
Document that nroff adds two spaces after each sentence when reflowing, and therefore if you want formatted Pod::Man output to consistently have one space after each sentence, you will have to avoid ending a sentence at the end of a line in the middle of a paragraph.
Document that parse_lines and parse_string_document expect raw bytes, not decoded characters.
Fix the test suite use of parse_string_document to pass in raw bytes rather than decoded characters, which will hopefully fix test failures with versions of Pod::Simple older than 3.22 and is a more correct test of encoding handling.
Drop support for Perl 5.6. The minimum supported version is now Perl 5.8. Perl 5.6 had not been tested even by CPAN Testers in some time and isn't supported by Travis-CI, so true support is dubious. Dropping that version allows cleaning up some old compatibility code.
Fix a warning when outputting to something without a PerlIO layer, such as when output_string is used.
Fix behavior of S<> with Unicode input in Pod::Text to be consistent with behavior with a default encoding, namely treat all whitespace inside S<> as non-space characters and do not collapse it with adjacent whitespace.
Remove an ancient workaround in Pod::Text::Termcap that set the TERMPATH environment variable whenever a Pod::Text::Termcap object was created in order to add /usr/share/lib/termcap, necessary on some ancient Solaris systems. Setting environment variables is bad behavior for a module, and the Solaris systems requiring this workaround are long obsolete.
Remove the Pod::Text::Termcap fallback to VT100 escape sequences if Term::Cap was not able to find sequences for bold, underline, or normal text, and instead skip that part of the formatting. This will produce more correct behavior on dumb terminals at the possible cost of losing formatting on systems with malfunctioning terminal databases, which seems like an improvement. Thanks, Zenin. (#131124)
Further improve the man/no-encode.t test to not care whether Encode was already loaded or not. Thanks, Martin Becker.
Improve logic for showing large test failures to avoid spurious failures on systems without diff.
Skip the man/no-encode.t test if Encode is already loaded while running the test suite, which seems to happen sometimes with CPAN Testers tests.
Fix the Pod::Text documentation to reflect that the default value of the sentence option is false, not true, and has been since at least 2.0.0. Thanks, eponymous alias. (#124461)
Correctly honor the width option in Pod::Text::Termcap, which was previously ignored due to a bug in interpreting user-supplied options. Thanks, eponymous alias. (#124447)
Fix a subtle wrapping bug in Pod::Text::Color and Pod::Text::Termcap with long =item text that would cause lines to be wrapped when they didn't need to be. Thanks, eponymous alias.
In Pod::Text::Color and Pod::Text::Termcap, clear any text attributes at the end of each line and reinstate them at the start of the next line, since some pagers (less -R in particular) clear all text attributes at the end of each line and were therefore not correctly showing attributes in wrapped text. Thanks, eponymous alias.
Correctly get the terminal width for Pod::Text::Termcap from Term::Cap information when COLUMNS isn't set, instead of getting a string value that later cannot be used numerically.
Specifying "none" for the errors option of Pod::Man and Pod::Text no longer results in an errata section in the generated documentation, matching the documented behavior. Thanks, Olly Betts.
Fix order of SEE ALSO section in manual pages to match the recommendation of perlpodstyle.
Change manual page references and function names to bold instead of italic in *roff output following the current Linux manual page standard. The previous formatting was taken from Solaris, and it seems safe to say that the Linux manual page formatting conventions are now much more widely followed than Solaris's. Patch from Guillem Jover.
Revert the .IX handling code in Pod::Man to the earlier version from Bjarni Ingi Gislason but add the trailing backslashes that should hopefully avoid blank page issues on HP-UX. This fixes a warning regression when man is run with warnings enabled. (Debian Bug#847972)
In Pod::Man, wrap the output file descriptor in a glob before passing it to PerlIO::get_layers so that the layer check works properly. Previously, this code would throw a warning if given a scalar not wrapped in a glob and not detect layers properly. Patch from Zefram. (#122521)
Produce a proper diagnostic when given empty input on standard input with no other arguments to pod2man or pod2text. Reported by Guillem Jover.
Use Pod::Simple's logic to determine the native code points for NO BREAK SPACE and SOFT HYPHEN in Pod::Text instead of hard-coding the ASCII values. Hopefully fixes the case of mysterious disappearing open brackets on EBCDIC systems. (#118240)
Partially revert change in 4.00 to require the name option (--name to pod2man) when generating manual pages from standard input. Historically, pod2man silently tolerated this, and there turned out to be a lot of software that depended on this, making the change too disruptive. Instead, silently set the manual page title to STDIN in this case, but warn about it in the documentation. (#117990)
Fix rendering bug for "TRUE (1)" in Pod::Man, which was recognized as needing small caps and then erroneously as a manual page reference, resulting in escaped nroff. (Found by Dan Jacobson with the XML::LibXML::Element manual page.) (Debian Bug#836831)
Fix rendering bug causing "\s0(1)" to be mistakenly marked as a man page reference in Pod::Man, later confusing backslash escaping.
Add new lquote and rquote options to Pod::Man (and corresponding --lquote and --rquote flags to pod2man) to set the left and right quotes for C<> text independently. (#103298)
Remove test for nested L<> markup, since an upcoming version of Pod::Simple will drop support for this. (#114075)
Avoid undefined variable warnings when determining the title for a Perl module at the top level of a distribution. Thanks, Dave Mitchell. (#112625)
Fix font resets with nroff when fixed-width fonts are used in the label for an =item. Previously, italic was being ended with \f(CW even in nroff mode, which, with groff, only changes the font to fixed-width and doesn't reset to a non-italic font. Thanks, Paul Townsend. (#98199)
Suppress warnings about a missing Encode module if PERL_CORE is set in the environment. Due to build ordering during Perl core builds, Encode is expected to not yet be available during the build step that sets PERL_CORE. Thanks, Dave Mitchell.
Handle scripts ending in .com on VMS systems and don't generate the man page for perlpodstyle when built as part of Perl core. These are hopefully the last two changes required to fully merge with Perl core without core having to maintain a separate build system. Thanks, Craig A. Berry.
During build, generate the pod2text and pod2man manual pages from the *.PL files rather than the generated scripts. This may be required due to the different script extensions on VMS hosts.
Rename perlpodstyle back to perlpodstyle.pod, since we no longer need the workaround for Module::Build's POD handling.
Switch back to generating pod2man and pod2text from *.PL files. While ExtUtils::MakeMaker can fix the #! line, it can't handle all non-UNIX operating systems, and the *.PL script generation code can. This will hopefully remove the need for Perl core to maintain a separate copy of the *.PL wrapper scripts as well. Thanks, Craig A. Berry.
Fall back (with a warning) to non-utf8 behavior in Pod::Man if the utf8 option is specified but the Encode module is not available. This is useful in some cross-compilation situations. Thanks, Niko Tyni.
Don't include .travis.yml in the distribution so that it isn't picked up by Perl core. Thanks, Karen Etheridge. (#110385)
Add homepage information to the CPAN metadata and change the canonical repository location to GitHub.
Fix tests when POD_MAN_DATE or SOURCE_DATE_EPOCH are already set in the environment. Thanks, Niko Tyni. (Debian Bug#807086)
For versions of Perl prior to 5.11, install the modules into the Perl core module directories, since in those versions site modules did not take precedence over Perl core modules. Thanks, Peter Rabbitson. (#110024)
Do not override the TERMPATH environment variable in Pod::Text::Termcap if it's already set. This should fix the test suite with Term::Cap 1.16 (which has a bug in termcap handling if TERMPATH doesn't point to a valid file). Also document the manipulation of TERMPATH.
Revert the switch to Module::Build as the build system. This creates a circular dependency with Module::Build, since it wants a newer version of Pod::Man than in Perl versions prior to 5.10.1. Instead, add the new metadata to Makefile.PL and stick with a single build system that will also work inside Perl core.
Increase the version number of the package to be larger than any of the previous version numbers of any of the modules, and change all modules to use the same version as the overall podlators package. Switch to a simple decimal version number to avoid complexity with v-strings and portability to old versions of Perl.
podlators now requires Perl 5.006 or later. All modules enable warnings. Please report any unexpected or confusing warnings as bugs in the bug tracker.
In previous versions, the -r or --release option to pod2man could be specified without an argument and was interpreted as setting that value to the empty string. That never made a great deal of sense, and the original change to Perl was apparently because no one realized one could pass the empty string as the argument value. The argument is now mandatory, but may be the empty string, which will cause some *roff implementations to use the system default.
Allow any even number of characters to be specified as the quote marks for Pod::Text and Pod::Man (and the corresponding --quotes options of pod2text and pod2man), rather than being artificially limited to one- or two-character quotes. The first half of the string will be used as the left quote and the second half as the right quote. This allows Unicode characters or groff escapes like \(lq and \(rq to be used. (Partly addresses #103298)
Attempt to detect if the Pod::Man input came from a pipe and therefore has a completely unhelpful (and nonreproducible) source file name, and diagnose this as an error. Document that the name option (--name to pod2man) is required when processing POD source from standard input. (Debian Bug#777405)
Honor the environment variable SOURCE_DATE_EPOCH and use it as the timestamp from which to derive the left-hand footer in *roff output if the date option is not set, overriding the timestamp of the input file. Also use the value of the environment variable POD_MAN_DATE as the left-hand footer if it is set. (This is an earlier version of the SOURCE_DATE_EPOCH support.) These changes are primarily useful to ensure reproducible builds of the same output file given the same source and Pod::Man version, even when file timestamps may not be consistent. Thanks, Niko Tyni. (Debian Bug#801621)
The default left-hand footer date in *roff output is now based on UTC rather than the local time zone to make the output more reproducible. Thanks, Chris Lamb. (Debian Bug#780259)
Simplify the *roff preamble code for handling the F register and index entries, and add backslashes after the braces in the preamble code for handling the F register to avoid introducing a spurious page break before at the first page with AT&T *roff. Thanks, Carsten Kunze and Daphne Pfister. (#92979)
Support setting the left-hand footer in *roff output to the empty string.
Fix documentation of the utf8 option to Pod::Man and Pod::Text, and the corresponding -u option to pod2man and pod2text, to reflect that Pod::Simple now autodetects Latin-1 and UTF-8 but warns.
More clearly document the options that set values in the .TH header in the pod2man and Pod::Man documentation. Thanks, Guillem Jover. (#103297)
Fix encoding handling in Pod::Text of documents that start without an encoding declaration and then declare an encoding partway through. Previously, this would result in attempts to print wide characters if there were non-ASCII characters in the document. Thanks, Magnolia K. (#101722)
Change the documentation to not say Pod::Text only generates ASCII text. (#89355)
Switch the preferred module build system to Module::Build, but still provide a Makefile.PL file for backward compatibility and for the use of Perl core. (#108714)
Installation of this package no longer tries to overwrite the Pod::Man and Pod::Text modules that come with Perl core, and instead relies on the normal precedence rules in Perl's module search path that prefer locally-installed modules over core modules.
Work around a bug in Term::Cap 1.16 that caused the test suite to fail by forcing a setting of TERMPATH to a termcap file provided by the test suite while running tests. (#98272)
Fix documentation of the default for the errors constructor parameter.
Skip the empty text and manual page errors tests if Pod::Simple didn't produce any errors, which happens with the version shipped with Perl versions prior to 5.18. Catch warnings as well as exceptions in these tests.
The parse_lines and parse_string_document methods in Pod::Man and Pod::Text now set a default output file handle of STDOUT if none was set.
Perform document initialization even if the document is contentless. Documents with only errors are shown as contentless but then have a POD ERRORS section, and previously this led to internal errors because state variables weren't properly initialized. Thanks, Andreas Koenig. (#88724)
Apply various optimization improvements from Dagfinn Ilmari Mannsåker. There should be no changes in the output. (#83253)
Fix an erroneous output_fh reference in the Pod::Man documentation. Thanks, Andreas Koenig. (#88723)
Fix various comment typos. Thanks, David Steinbrunner. (#85683)
In perlpodstyle, wrap verbatim license line in POD that was over 79 characters after nroff indentation. Thanks, Brian Gottreu and Steve Hay. (#87440)
Adjust the tag width tests and the list handling tests to avoid spurious warnings from Pod::Simple about mismatched =item types.
Support a new errors option in Pod::Man and Pod::Text. Valid values are die, stderr, pod, and none. Convert the stderr option to the errors option with value stderr. Add the corresponding --errors option to pod2man and pod2text. (#39007)
Add a new nourls option to Pod::Man and Pod::Text to suppress the URL from L<> formatting codes that contain anchor text, and add the corresponding --nourls option to pod2man and pod2text. (#62210)
Extend a small-caps section in *roff output through the punctuation that commonly appears in license disclaimers so that small caps isn't turned on and off at the boundaries of every word, producing unreadable *roff.
Collapse consecutive whitespace and remove newlines in index term text in *roff output. Thanks, Kevin Ryde. (#82332)
Remove the test of a POD document without an encoding. We previously tested that this interpreted the document as ISO 8859-1, but Pod::Simple behavior has changed so that the test started failing, plus Pod::Simple now warns about a missing =encoding. (#77553)
Fix detection of PerlIO UTF-8 handling by requesting details on PerlIO layers to look for the UTF8 flag, which is not a layer in its own right. Thanks, Leon Timmermans. (#76440)
Fix handling of the F register in *roff output when processing multiple documents at once. .IX will now continue to be defined for documents after the first, and the page number will not be reset at the start of each document. Thanks to Nicholas Clark for the analysis. (perl #103202)
In the pod2man and pod2text driver scripts, report an error and remove the empty output file if the input file had no content (if it did not exist, for example). Exit with non-zero status if there were any errors. Track contentless status inside Pod::Man and Pod::Text. Thanks, Dmitry Smirnov. (#75099)
Override parse_file in Pod::Man and Pod::Text to set output_fh to STDOUT if it is not already set. (#77530)
Format the URL text in Pod::Man before comparing it to the anchor when deciding whether to show separate anchor text. This avoids spurious mismatches between the URL target and anchor text because the anchor text was already formatted and has (for example) hyphens escaped. (#76396)
Define \*(C` and \*(C' to the empty string in *roff output when processed through troff to avoid groff warnings. Avoid warnings from checking the F register (used to enable index output) when running under groff. Patch from Bjarni Ingi Gislason. (#75434)
Fix the ASCII fallback string for the AE ligature in *roff output to use the string that was actually defined.
Stop removing pod2man and pod2text on make realclean, left over from when they were generated from *.PL scripts. (#74848)
Embed the PID in file names generated by the test suite to avoid conflicts when running the test suite in parallel. (#62083)
Switch UTF-8 output encoding to use Encode directly instead of adding a PerlIO layer. Probe the PerlIO layers with protection for Perl versions without PerlIO and set a flag indicating whether we also need to encode, to avoid double-encoding when writing to a file handle that is already doing UTF-8 encoding via PerlIO.
Do not strip escaped trailing whitespace in *roff output such as that created by S<> at the end of a line, since the backslash is then taken by *roff as escaping the newline. Thanks, Kevin Ryde. (#61781)
Add perlpodstyle, a new style guide for POD documentation, split mostly from the NOTES section of the pod2man manual page. Remove the NOTES section of pod2man's documentation.
Convert pod2man and pod2text from scripts generated from *.PL files to simple scripts, relying on ExtUtils::MakeMaker to handle replacing the #! path during the package build.
Increase $VERSION in Pod::Text::Color and Pod::Text::Termcap, missed in the previous release.
Support anchor text for L<> links of type URL by rendering the anchor text and then the URL in angle brackets. Now requires Pod::Simple 3.06 or later.
When formatting item tags in Pod::Text, use the width of the tag without formatting codes. This fixes formatting issues with Pod::Text::Color, Pod::Text::Termcap, and Pod::Text::Overstrike.
Suppress all formatting in the NAME section in *roff output to avoid confusing lexgrog and fix mishandling of C<> markup in NAME. Clarify in the pod2man documentation that no markup should be used in the NAME section of a manual page. Thanks, Niko Tyni.
Escape backslashes in the quoted text of .IX macros generated from X<> formatting code in *roff output.
Avoid using POSIX::strftime in Pod::Man because POSIX requires Fcntl, which is an XS module, and hence can't build in miniperl. This allows ExtUtils::MakeMaker to build as a normal module in Perl core. Thanks, Michael G Schwern.
Allow anchor text for URLs in Pod::ParseLink. Fix the check of the anchor text to not think no text was provided when the text was "0".
Remove the temporary files created by the test suite in a loop to ensure that all versions are deleted on VMS. Thanks, John E. Malmberg.
In Pod::Text, correctly handle indentation of verbatim paragraphs that contain lines with only whitespace. Thanks, Renee Baecker.
In the legacy pod2text method, properly initialize the output file handle when called with only one argument. Thanks, Michael G Schwern.
Fix the t/text-encoding.t test on Windows by setting raw encoding on the output file handle. Thanks, Steve Hay.
Try to preserve the previous Pod::Text behavior of setting the output encoding to match the input encoding if utf8 is not set, but support forcing an output encoding of utf8 with the utf8 option. Add a corresponding --utf8 option to pod2text. Document the PerlIO limitations of the current utf8 support.
Quote all module version numbers to preserve any trailing zeroes.
Skip UTF-8 tests for versions of Perl prior to 5.8.
Add a stderr option to Pod::Man and Pod::Text that sends POD errors to standard error instead of adding a POD ERRORS section to the generated documentation. Add a corresponding --stderr option to pod2man and pod2text.
Stop remapping the code point for non-breaking space in Pod::Man. This should not be necessary and was wrong when the string from Pod::Simple was a character string and not a byte string. It was papering over a bug in setting the encoding of an input POD file.
Use .SS instead of a local .Sh macro for subheadings in *roff output, and stop defining .Sh.
Remap ISO 8850-1 non-breaking spaces produced by Pod::Simple to the corresponding UTF-8 code point for UTF-8 *roff output.
Fix multiple spelling and markup errors.
Do not include the accent mark definitions in generated *roff if the output is in UTF-8.
Fix the test for S<> handling with all whitespace to not give a spurious failure with Pod::Simple 3.06.
Add a new utf8 option to Pod::Man. If set, do not convert non-ASCII characters to *roff escapes or X, and instead output literal UTF-8 characters. Add a new --utf8 option to pod2man.
In Pod::Man, match text between \f(CW and \fP or \fR in headings non-greedily to get the fonts right with multiple C<> formatting codes.
Protect .Sh text against leading *roff control characters since some *roff implementations apparently "look through" font escapes at the beginning of lines.
In Pod::Man, escape backslashes separately from processing non-ASCII characters and do that, dash escaping, and underscore adjustment before processing non-ASCII characters. Otherwise, we escape the hyphen in eth characters.
Escape apostrophes and backquotes in verbatim and C<> text in *roff output.
Define the IX macro to empty in *roff output rather than leaving it undefined when indexing is not requested to eliminate warnings when groff warnings are enabled.
Simplify the logic to skip lib directories in Pod::Man when constructing page titles to avoid Perl warnings and unnecessary checks.
Accept and mostly ignore a hash of options as the first option to parse_from_file. Support an option of -cutting and configure Pod::Simple to assume the POD has already started. This is for backward compatibility with Pod::Parser.
Recognize more uses of hyphens in regular English text and allow them to be regular hyphens in *roff output.
Turn off hyphenation and, for nroff, justification after the .TH macro since that's where groff turns them on.
Stop mapping vertical bar to \(bv in *roff output, since it produces Unicode characters where they aren't desirable. Remove the preamble reference to the Tr string, which was never defined.
Pod::Simple's source_filename method returns garbage if we're parsing from a file handle, so use the current time if stating the returned input file fails in Pod::Man.
Add parse_from_filehandle methods to Pod::Man and Pod::Text for backward compatibility with the earlier versions based on Pod::Parser.
In the test suite, pass in a file handle for Pod::Simple output and then close it afterwards. This works around Pod::Simple leaving file handles open and preventing removal of temporary files on Windows. This is temporary until a new Pod::Simple release offers a better approach.
In the parse_from_file method, flush the output file handle rather than closing it. Closing it is unexpected and could break callers.
In Pod::Text::Termcap and Pod::Text::Color, use additional temporary variables to avoid ${char}{0,$width} in regexes, which only works in very recent Perls.
Call reinit before calling the Pod::Simple parse_from_file method to preserve the previous capability of reusing the same Pod::Man object for multiple documents. Close the output file handle after Pod::Simple returns to force the output to flush.
The legacy pod2text method was broken because Pod::Simple's parse_file method only takes one argument. Pass the second argument to output_fh instead.
Fix portability issues with Perl 5.005.
Use a single object for all conversions in pod2man and pod2text for a minor speedup.
Rewrite all modules and driver scripts to use Pod::Simple instead of Pod::Parser. The output should be identical except that C<> with no content outputs quotes for Pod::Text-based parsers, E<> handling is improved, and various small bugs have been fixed. Thanks, Sean Burke.
Create a new parser for each file in pod2man since Pod::Simple parsers are not reusable.
Add pod2text support for multiple pairs of input and output files, similar to pod2man.
Strip vendor_perl as well as site_perl when determining the manual page title. Thanks, Alexey Tourbin.
Count text lengths correctly when wrapping in Pod::Text::Color and Pod::Text::Termcap when there are multiple adjacent escape sequences. Use a temporary variable to make the regex clearer.
Change section ordering in some documentation following perl5-porters discussion.
Remove obsolete documentation caution against enclosing URLs in L<>.
In Pod::Text::Termcap, handle the case where the HOME environment variable isn't set, mostly for Windows.
Make sure Pod::Man returns 1 to keep Perl 5.8.0 happy.
In Pod::Man, track the type of items in an =over list and only map asterisk to a real bullet if the item type is bullet. Fix a bug where =item 0 was treated the same as =item with no tag.
Support a margin option in Pod::Text and use it to set the initial indentation level. Fix handling of the colon in the margin when the alt format is enabled. Add a new -m option to pod2text to set the margin.
Clean up some old-style L<> links in pod2text that were workarounds for fixed bugs in Pod::Man and Pod::Text.
Add a pointer to the module web site in the documentation.
Tweak the regex for matching numbers in C<> to not consider a single period to be a number, which affects whether surrounding quotes are added.
Fix the regex for wrapping lines in Pod::Text::Overstrike to use a non-backtracking section for each character to avoid exponential backtracking on lines with a lot of markup.
In Pod::Text::Overstrike, use [\b] instead of \cH in regexes to match backspaces, for platforms that use EBCDIC where \b and \cH aren't the same character.
Do not apply Pod::Man guesswork to the text inside the NAME section, since it may confuse programs like catman. Do not output .UC after the .TH macro. catman doesn't like anything between the NAME section and .TH, and .UC doesn't appear to actually do anything on any modern platform.
Correctly handle a verbatim paragraph right before a heading in Pod::Man.
Fix error reporting in Pod::Text for unknown sequences and unknown commands to be more consistent and update the documentation to match.
Change terminal speed to a number in Pod::Text::Termcap. Also fall back on a hard-coded terminal speed if getospeed doesn't work.
Fall back on a hard-coded terminal speed for Pod::Text::Termcap if POSIX::Termios doesn't work, such as on VMS.
Escape L<> in the NAME section of the Pod::ParseLink documentation.
In Pod::Man, return references to arrays rather than references to scalars for already-formatted text. There are too many odd bugs with scalar references in older versions of Perl. No longer bless references since the current Pod::Parser doesn't require it. Now requires Pod::Parser 1.13 or later.
Change all documentation references from interior sequences to formatting codes to match the terminology of perlpodspec.
Use an @INC path of ../lib and a new function to find source files for when the module test suite is being run as part of the Perl core tests.
Wrap the call to Term::Cap in Pod::Text::Termcap with eval because it throws exceptions if the terminal can't be found. Fall back on the ANSI escape sequences rather than dying if the termcap entry is incomplete. Note the fallback in the documentation.
Delete the lax option in pod2man before calling Pod::Man and document that it is obsolete and podchecker should be used instead.
Improve the pod2man and Pod::Man documentation to refer to podchecker, add discussion of guesswork and 8-bit character handling, and mention the fragility of the heuristics.
Interpolate before formatting in Pod::Text::Overstrike to prevent the formatting codes from ending up in the output, and strip any existing formatting before applying new formatting.
Use font escapes in Pod::Man rather than .I to avoid strange problems with quoting, at least for =head3. =head1 and =head2 likely still have troubles with repeated double-quotes. Fix all fixed-width font changes for nroff, not just the simple ones, and don't hard-code the value of any fixed-width font.
Improve and simplify the handling of indentation shifts in Pod::Man.
When intuiting the manual page name for a module, also strip $^O by itself as a directory component even when not preceded or followed by a dash and other text.
Fix handling of =for or =begin/=end in =item paragraphs in Pod::Text. Default to a tag of "*" if none is given. Insert some whitespace for empty item paragraphs to keep them from blending into subsequent text.
Fix a Pod::ParseLink bug in the handling of link text that's entirely in quotes. Double quotes are now only removed around sections, not names. Text enclosed entirely in double quotes is interpreted as a link to a section.
Fix various -w warnings.
Fix -w warnings with hyphen handling.
Add a new module, Pod::ParseLink, to parse the contents of an L<> sequence. Use it everywhere. Defer expansion of formatting escapes inside L<> until after L<> is processed. Surround URLs with angle brackets in the output.
Remove the special handling of consecutive L</section> links.
Support E<apos>, E<nbsp>, and E<shy>.
Completely rewrite the name parsing code for modules in Pod::Man to use File::Spec. In the process, fix a bug in dealing with the new three-component version number directories. Swap the order of date and release in the .TH line to better comply with the man macro documentation.
Rewrite the handling of dashes and hyphens in Pod::Man. Be much more conservative about which hyphens are turned into dashes, and make all hyphens non-breaking unless we can be fairly sure that they're inside normal words.
Handle indentation of =item-less =over/=back blocks in Pod::Man.
Include the version of Pod::Parser in the header of *roff output.
Only try to determine a module name from the path for the manual page name if the manual page we're generating is in section 3.
No longer insert a timestamp into the generated manual page. It just causes unnecessary differences and merge conflicts.
Inside S<> in Pod::Man, convert all whitespace to non-breaking spaces, not just spaces.
Add the --name option to pod2man and document the name option in Pod::Man.
Use L<> for all manual page references in the documentation that should be highlighted. Switch the rest to bold versions of the program name. Change func(n) to func(3) in the example of things that are automatically formatted so that it will be formatted. Remove from BUGS the note that some of the path mangling assumes Unix directory separators. Don't give anchor text for L<> links that no longer require it.
Update the documentation in Pod::Text and subclasses to use now-allowable POD constructs like C<< >>. Don't escape angle brackets that don't require escaping. Don't give anchor text for L<> links that no longer require it.
Add the code option to Pod::Text to include the non-POD text of the input file and document it. Add the corresponding --code flag to pod2text.
Convert warnings for unknown escapes, unknown sequences, and unmatched =back into warnings from carps, and include the file and line number of the POD data instead of the Perl code.
In Pod::Text::Overstrike, better handle the case where a highlighted portion of text ends a line.
Add --verbose flag to pod2man to print out each output file as it is generated.
Fix *roff syntax error from using .if with .el during quote handling.
Fix output for X<> sequences in Pod::Man.
Add heuristics to decide whether to quote the argument of C<>.
Remove font changes for nroff with C<> to work around a bug in the Solaris 2.6 version of nroff's handling of \fP in headings. No longer add an extra level of quoting for =item; it isn't necessary.
Remove the Pod::Man logic turning PI into a pretty pi character. It produces too many false positives.
Remove the definition .Ip from the preamble of *roff output. Remove .bd B 3 from the preamble; this isn't part of the accent mark definitions but instead changes the way bolding is done, confusing some other translators. Use .IP instead of .Ip everywhere.
In the POD style section of pod2man, add a description of the COPYRIGHT AND LICENSE section, add a mention of a mailing list in SEE ALSO, and mention that large logs are better kept separate from HISTORY in the description of a standard manual page.
Fine-tune *roff formatting guesswork. Don't allow colons after sequences to put in small caps since they're already handled by being rolled into the sequence and were causing weird things to happen in references to functions. Allow small caps before an open paren. Teach the handling of functions and manual page references about small caps escapes, and be pickier about what constitutes a manual page reference.
Fix again the incorrect mappings for E<Iacute> and E<iacute> in Pod::Text, and this time for E<Igrave> and E<igrave> too. Thanks, Sean Burke.
Output anything that looks like a URL verbatim rather than interpreting it as a manual page reference.
Remove newlines from heading contents in Pod::Man.
Quote the file name in the manual page header if it contains spaces.
New Pod::Text::Overstrike contributed by Joe Smith. Pass -o or --overstrike to pod2text to use it for formatting.
Add another level of quoting of double quotes for =item text in *roff output. This was already present but not working for C<> text because it was in the wrong order.
Change the default quote character for C<> to be double quotes rather than matched left/right single quotes.
Add support for =head3 and =head4.
Allow pod2man to take multiple pairs of input and output files on the command line to decrease the time that it takes to process all of Perl's documentation.
Switch \*C` and \*C' sequences in *roff output from C<> as well as literal double-quotes if the quote character contains double quotes. Not doing this was causing weird output on some systems in some circumstances. Use a separate quote mapping function for text blocks to work around a Solaris 2.6 nroff bug.
Use \fP to switch back to the default font in *roff output rather than changing back to \fR so that font changes work correctly in headings using a different font. Sprinkle \fP through all font changes so that the default font is always the "previous" font so that the above works.
Output .PD 0 and .PD around repeated =item tags in *roff output so that they're formatted without intervening blank lines, improving formatting of, e.g., perlfunc.pod.
Fix incorrect mappings for E<Iacute> and E<iacute> in Pod::Text. Thanks, Sean Burke.
Support configuration of what quote characters to use around C<> text. Add a new --quotes option to pod2man and pod2text.
Report nicer errors when encountering an unknown paragraph command.
Add support for E<sol> and E<verbar>.
Fix the regex for stripping bullets from index entries in *roff output so that it doesn't strip a leading "o".
In the prelude of *roff output, terminate the .IX definition with ".." instead of ". ." for groff.
Avoid assigning to $_[0] in the pod2text method.
Fix hyphens and underscores in *roff output only in literal C<> content to avoid mangling of hyphens and underscores that are the result of other sequence processing.
Install the modules in the Perl core area if the Perl version is 5.6.0 or higher.
Strip a leading lib/ from a file name for module manual pages when constructing the manual page name, needed for ExtUtils::MakeMaker.
This has now been incorporated into Perl core as pod2man and pod2text. Rename pod2roff to pod2man accordingly.
Hide "-" arguments to the driver scripts from Getopt::Long so that Pod::Parser will interpret them as STDIN or STDOUT.
Protect any line in *roff output that starts with a backlash and leading periods following font escapes. Replace embedded newlines in titles with spaces.
Use "perl v5.6.0" instead of "perl 5.6, patch 0" for the default release string in *roff output, and handle both pre-5.6 and post-5.6 version numbering schemes. Zero-pad the month and day in the modification date. Avoid warnings when center, date, or release aren't set.
Allow for two-character fonts in *roff output.
Work around a Perl 5.6 bug affecting L<> text generation in *roff output. Fix Z<> handling with current Perl.
Make filename munging safe in roff output even when $ is set and the filenames contain embedded newlines.
Fix the regex to concatenate multiple L<> section links in *roff output, and fix whitespace handling for it around "and".
Add the remaining ISO 8859-1 HTML entities to Pod::Text. Thanks, Tim Jenness.
Change Getopt::Long config in pod2man from bundling to bundling_override so that options like -center work for backwards compatibility.
Don't default to Pod::Text::Termcap in pod2text even if STDOUT is a tty until it works right on Windows, VMS, etc.
Add support for numeric E<> escapes.
Fix doubled quotes in links to sections in *roff output.
Export pod2text for backwards compatibility.
Fix argument passing to Pod::Parser in pod2roff to use an expanded hash instead of a hash reference.
Change the parsing model in Pod::Man so that, rather than deferring E<> escapes until just before output, *roff output is generated by the interior sequence parsing and the result is passed up the parse trees as Pod::Man::String objects instead of scalars to mark the output as already processed. In the process, clean up what *roff escaping and guesswork is applied where, and clean up the whole process of applying guesswork. Improve the escaping of dashes and hyphens to use a single pass.
Improve the small caps guesswork for *roff output to allow for more cases, including several adjacent all caps words.
Fix some bugs with the link text generation for manual page references in *roff output.
Improve the index generation in *roff output slightly.
Fix several places in Pod::Man that were clobbering the caller's $_.
Add pod2roff and Pod::Man, which convert POD to manual pages.
Rename pod2txt to pod2text and Pod::PlainText to Pod::Text.
Output =begin text blocks verbatim in Pod::Text rather than interpreting them as POD.
Document the Pod::Text oddity with Ctrl-As as a restriction.
Always treat =for paragraphs as verbatim text in Pod::Text.
Add a BUGS note to Pod::Text::Color that the implementation is rather incomplete, and document the reliance on Term::ANSIColor.
Add an explicit check for Term::ANSIColor in pod2text if -c was given.
Add pod2text documentation of -h and expand the DIAGNOSTICS section to include directly-generated error messages and the most common Getopt::Long message.
Rename Pod::SimpleText to Pod::PlainText in one more place in the documentation.
Use File::Spec during the build to build file paths for portability, and remove the dist setting since current Perls get this right.
Fix the #! line in pod2txt during the build.
Rename Pod::SimpleText to Pod::PlainText.
Document that Pod::Text::Termcap is used by default in pod2txt if STDOUT is a tty. Clarify the documentation of --loose.
Rename the package itself from Pod::SimpleText to podlators.
Add a pod2text function for backwards compatibility.
Properly wrap multiline =item tags.
Fix a spurious space with =for text commands.
Check the content of sequences against the empty string specifically rather than testing truth so that it does the right thing with 0.
Process sequences for =head headings.
Initial release with pod2txt and Pod::SimpleText.
| < podlators README | Russ Allbery > Software > podlators | POD Style Guide > |