Revision history for Spreadsheet-XLSX-Reader-LibXML
Uses the change tracking convention
+ New feature.
- Removed/deprecated feature.
! Bug fix.
* Place holder
NOTE: This software is designated as beta quality until this notice is removed. The goal is to
be method matched to Spreadsheet::ParseExcel prior to the 1.0 release.
v0.38.20 2015-10-10 06:28:21-07:00 America/Los_Angeles
+ Added an Attribute that allows control of whether the errors get Carp's longmess
+ Added better handling of double digit text years in conversion to a date
This assumes that the year is no more than 20 years in the future or 80 years
in the past of the process time
+ Added an attribute 'european_first' to the formatter that prioritizes the DD-MM-YY
style of date string parsing over MM-DD-YY
+ Re-wrote the worksheet parser to have a 'fast-forward' mode.
This reduces the intermediate processing (time) spent on advancing through the worksheet
when you don't want the values, closes #83 in github
! Added a test for bad worksheet dimension values in a worksheet, closes #79 in github
! Added a test for missing row span attributes in a row node, closes #78 in github
+ Aligned dist.ini a bit with the other packages I maintain
v0.38.18 2015-09-24 10:52:42-07:00 America/Los_Angeles
! Several tests added in the background (Not part of the package yet) to ensure a safer release
The (future) goal is to create tests small enough to fit with the package for the same failure modes
! Fixes #82 in github as well as removing the fail for negative decimal rounding numbers inferred from scientific notiation
! Addresses #81 in github with safer string searching
! Addresses #80 in github by Manually building this package on my system and then running the Spreadsheet::Read (master)
tests after. Long term I would like to import more of that testing to this package.
! Addresses #79 in github with by allowing for EOF flags to add this and only use the value where available
Still needs a test added to the test suit
! Addresses #78 in github with by substituting the values from the sheet where no row span is provided
Still needs a test added to the test suit
v0.38.16 2015-09-20 19:21:30-07:00 America/Los_Angeles
! Closes issue #72 in github (hopefully) with safer testing of length in a variable
! Closes issue #75 in github with safer setup of cell instances with empty cell_coercion attributes
! Closes issue #74 in github with safer testing for the apple date flag
! Closes issue #73 in github with safer coercion assesment of date coercions
+ Added caching for the styles sub-file which should noticibly speed up parsing
+ Added a :debug flag tgit hat turns should_warn on. This is mostly useful to get warns on
build fails since the warnings can be turned on after build as desired.
+ Added "5.22"(back) and added "dev" and "5.10.1" to the TravisCI testing
! Closes #71 in github with updates to the import statement
+ fetchrow_array and fetchrow_arrayref no longer report undef for the positions representing
the remaining empty cells in a row ( the list just ends )
+ Treat stored empty stings -''- as empty cells for the 'values_only' attribute
This will still skip the empty string cells even if 'empty_return_type' is set to empty_string
+ Rewrote XMLToPerlData and moved it up one file layer since it is more reader agnostic
This is a first step to integrating a file type agnostic XMLReader #4 (milestone v0.40.2)
with improved XMLtoPerl conversions. This might break implementations that are built on the guts of this thing.
Contact me with questions or concerns. Top level API implementations should be unaffected.
I also moved the test file to reflect this change
+ Changed the name of the Worksheet class to WorksheetToRow as a first step in moving this to a role
sometime in the future to eliminate inheritance as much as possible. (See the worksheet interface change below)
I changed the name of the test to match
+ Changed the name of the GetCell role to Worksheet with the intent of treating it as an interface
I moved it up one layer in the folder structure
I removed the worksheet pod (now in the same folder) since the documentation was duplicated there
I moved the test and changed the name to reflect this change
+ Re-wrote the worksheet parser to parse whole rows at a time rather than one cell at a time
This unkinks some of the cruft created when adding tracking of hidden_columns (v0.38.12)
+ Added a Row Class to hold and process cell data within a row
+ Added support for .xlsm files closes #68 in github
This does not include any reading or handling of 'vbaProject.bin'
It should be safe even with viruses if the reader doesn't mess with it
! Changed the %ENV test to a caller test run in BUILDARGS ( during ->new )
This should handle all useage by Spreadsheet::Read (enforces the :like_ParseExcel flag)
v0.38.14 2015-08-24 16:49:04-07:00 America/Los_Angeles
! re-closes #52 in github with the addtional testing suggested
! Fixed a bug with 'is_hidden' reporting at the cell level tested by 07-losing_is_hidden_bug.t
! Changed the %INC test to an %ENV test and turn on the like_ParseExcel flag with
$ENV{SPREADSHEET_READ_XLSX} == "Spreadsheet::XLSX::Reader::LibXML"
This is a more accurate flag test for use by Spreadsheet::Read
v0.38.12 2015-08-19 05:19:28-07:00 America/Los_Angeles
+ Fix yet another bug in the hidden formatting (hopefully)
+ Make the decimal vs scientific choice for hidden formatting more robust( issue #65 on github )
+ Adding tests for hidden tabs rows columns and cells in the worksheet and cell
+ Adding the method 'get_merged_areas' to match ~::ParseExcel
+ For compatability purposes this package will now test (during import) to see if it's being
used in Spreadsheet::Read (potential future) and automatically push the :like_ParseExcel
tag on the end of the tag list so that it will behave consistently to ParseExcel for a common
reader. (This is a beta feature and not guaranteed to remain)
+ This package now supports stacked flags (not that there exists so many already)
The purpose of this is as I run into additional use cases the intersection of
the flags I support and the flags that are desired by the user may require some juggling.
Where flags disagree in their defaults the last flag passed will be used. This should allow
for the proliferation of flags for specialty cases without having to explicitly build one
for each use case. (At some point though you may just want to use the attributes).
The test for this is t/Spreadsheet/XLSX/Reader/06-stacked_flags.t
v0.38.10 2015-08-10 21:43:11-07:00 America/Los_Angeles
! Fix TravisCI build fail on test 60-hidden_format_strings.t
! Closes #60 in github
! Exclude text to sprintf number test bug
! Added a return the 'xml_value' only option to the 'group_return_type' attribute
v0.38.9 2015-08-06 13:45:36-07:00 America/Los_Angeles
! Closes #57 in github missing LICENSE file
v0.38.8 2015-08-05 11:15:01-07:00 America/Los_Angeles
! Re-Dist::Zilla build with the version badge corrected
! Closes #51 in github
This includes autoformatting of scientific numbers to the least significant digit in 9 positions
scientific number autoformatting is an excel application function that occurs even in the absence of applied formats
This includes some exclusion of scientific number like strings
+ Added the cell_xml_value attribute and methods to give access to the base scientific notiation data where
Excel would mangle it or make it potentially less precise even in an unformtted raw value thanks @Tux for the test case
+ Added support for the old(old) Excel style single quote in the formula bar representing left justification
! Closes #49 in github
+ Expanded the file type allowance to include capitalized file extentions
v0.38.6 2015-06-30 06:13:36-07:00 America/Los_Angeles
! Some general documentation updates for the FmtDefault class that were not done correctly for the last release
Doh!
+ Added the possibility of passing a closure to the set_headers for header value scrubbing
+ Adjusted the :just_the_data flag
cache_positions => 1, (documented as added but not added last commit)
from_the_edge => 0,
v0.38.4 2015-06-29 17:17:07-07:00 America/Los_Angeles
! The first Alien::LibXML change failed - trying again
v0.38.2 2015-06-29 16:45:47-07:00 America/Los_Angeles
+ Added a cache_positions flag to the Workbook level
For now this only populates to cache built number conversions and the shared strings file
+ 'set_custom_formats' now has built in coercion to change Excel strings to coercion objects
This allows for you to define custom excel column formatting (Not already defined in the file)
using excel strings without building an intermediate coersion before loading it to the worksheet.
+ I changed the custom_formats attribute to _custom_formats in order to indicate that it
should be set via the 'set_custom_formats' method
+ The worksheet class now accepts Excel format strings rather than requireing that the custom coercions be pre-built
(This is worksheet level functionality not workbook level functionality - still)
+ added the ability to only return portions of a node with a specific tag name to Spreadsheet::XLSX::Reader::LibXML::XMLReader::XMLToPerlData
This should be activated for the SharedStrings file whenever the workbook attribute 'group_return_type' is set to 'unformatted' or 'value'
The goal for this is to see a speed boost
Closes #47 in github
+ Added a cache_positions flag to the SharedStrings class
The goal is to see a speed boost for sheets with a large number of text strings
Like perl hash keys, shared strings are not ordered by cell
This means for files with lots of stored text each shared strings position is only read once instead of rewinding and re-searching
The new behaviour is managed by the SharedStrings class attribute 'cache_positions' (default on)
Setting cache_positions => 0 should return the old behaviour
If there is only text in the position then the cache only stores the text to minimize RAM impact
Additionally as a trade off for the increased RAM usage the sharedStrings file is closed after it has been completely read once
I will only read as far into shared strings as you ask (reading one value and then closing the sheet will only read to that value)
It will slow the first values out down but overall speed the reading of lots of values
Sheets with only numbers will not be affected
Closes #46 in github
+ Changed the cell formatter from a Role to a Class to match Spreadsheet::ParseExcel
THIS IS A REALLY BIG CHANGE
The target is to integrate better with Spreadsheet::Read
Closes #45 in github
+ Reduced the implementation of Alien::LibXML in dist.ini to see if that affects the CPAN Testers takeup rate
v0.36.28 2015-06-24 14:41:34-07:00 America/Los_Angeles
+ Surpressed a misleading error when value coercion is not succesful
v0.36.26 2015-06-23 19:01:21-07:00 America/Los_Angeles
! Closed #42 in github
v0.36.24 2015-06-10 19:44:32-07:00 America/Los_Angeles
! Attempting to solve the deep recursion from TravisCI in the new bugfix
+ Added perl 5.22 to TravisCI testing
! Fixed the failing TravisCI build
v0.36.22 2015-06-10 17:11:39-07:00 America/Los_Angeles
! Closed #40 in github thanks @wdbaker54
v0.36.20 2015-04-21 14:30:58-07:00 America/Los_Angeles
! Fixed the "this version" badge
! fixed some uninitialized values thrown by travis-ci
v0.36.18 2015-04-05 21:44:26-07:00 America/Los_Angeles
! Giving up on the javascript if statement (github strips 'script' tags and content)
v0.36_17 2015-04-03 23:37:22-07:00 America/Los_Angeles
! Test of the javascript changes to see if I can get it to be included
v0.36.16 2015-04-03 22:27:06-07:00 America/Los_Angeles
+ Added some javascript to the badges html to distinguish from a cpan domain so
versions only show on github
v0.36.14 2015-04-02 13:22:11-07:00 America/Los_Angeles
+ moved the current version to a badge and tweaked the badges to see what happens
v0.36.12 2015-04-02 11:45:08-07:00 America/Los_Angeles
! Updated the documentation in preparation for a push all the way to CPAN since
TravisCI builds are passing.
! Changed the MakeMaker call from setting CFLAGS to setting CCFLAGS
v0.36.10 2015-04-02 11:02:47-07:00 America/Los_Angeles
! Moved the Alien::LibXML requirement to the configure stage to try and pass a TavisCI build
v0.36.8 2015-04-02 10:58:17-07:00 America/Los_Angeles
! Rewrote the dist.ini so that Makefile.PL should implement Alien::LibXML correctly
v0.36.6 2015-03-31 19:48:28-07:00 America/Los_Angeles
+ Added a version header to the primary pod
+ Added the cpan version badge and a badge header
v0.36.2 2015-03-31 08:13:56-07:00 America/Los_Angeles
+ Added badges (Thanks @GARU)
+ Added Alien::LibXML as a requirement in MakeFile.PL to see if there was any impact
! Removed TODO #6 (I already did it)
v0.34.6 2015-03-13 22:44:41-07:00 America/Los_Angeles
! Closed #6 in github by writing POD for Spreadsheet::XLSX::Reader::LibXML::XMLReader
! Closed #7 in github by writing POD for Spreadsheet::XLSX::Reader::LibXML::XMLReader::SharedStrings
! Closed #8 in github by writing POD for Spreadsheet::XLSX::Reader::LibXML::XMLReader::Worksheet
! Closed #9 in github by writing POD for Spreadsheet::XLSX::Reader::LibXML::XMLReader::XMLToPerlData
+ added an 'alt_default' flag so that behaviour different than the Spreadsheet::ParseExcel would be easy
and closed #31 in github
v0.34.4 2015-02-20 09:41:28-08:00 America/Los_Angeles
! Closed #32 in github by creating a separation of chartsheet reading from worksheet
reading. I also found a read sheets more than once bug and fixed it along the way (thanks @morungos)
! Wrote POD for Types.pm module and closed #5 in github
v0.34.2 2015-02-14 08:28:33-08:00 America/Los_Angeles
! Scrubbed exising pod for cruft and closed #13 issue in github
! Improved the contributors section
! Re-orderd the changes section from bottom to top rather than top to bottom
! Moved the 'next' row function for empty requests for fetchrow_arrayref to
Spreadsheet::XLSX::Reader::LibXML::XMLReader::Worksheet from ~LibXML::GetCell
+ moved the xml_reader attribute from ~XMLReader to private and changed the
file_name attribute to coerce to a file handle and then build the xml_reader
in the background
+ Closed #27 in github by removing some cruft in the worksheet build (Thank you Frank Maas)
+ Closed #26 in github by manageing the variable out rather that messing with the variable in
This includes adding a possible test for the old failure in 10-get_cell.t (Thank you Frank Maas)
+ Added test 04-no_pivot_bug.t in response to issue #24 (pull request)
This test actually uncovered two issues with the simple implementation and both are fixed
The second issue is with a Styles sheet with no number formats listed
The third issue is with worksheet ranges that only have one cell. (Thank you Frank Maas)
! Fixed the bug where there are no shared strings in the excel file (Thank you Frank Maas)
! Closed #29 in github with a substantial rework of XMLToPerlData.pm (Thank you Frank Maas)
The module API and target results didn't change
+ Closed #30 in github by adding the 'values_only' attribute
documentation is found in the top level README
! Cleaned up a few more bugs found along the way
! Wrote POD for Styles.pod module and closed #12 in github
! Wrote POD for SharedStrings.pod module and closed #11 in github
! Wrote POD for XMLReader::Styles.pm module and closed #10 in github
v0.34.0 2014-12-29 08:49:14-08:00 America/Los_Angeles
! Closed #22 issue in github by adding the requirement for Type::Tiny 1.000 (vs 0.046)
to implement make_immutable in the type library
+ Scrubbed Spreadsheet::XLSX::Reader::LibXML POD to comply better with updates
v0.32.2 2014-12-26 16:25:30-08:00 America/Los_Angeles
! Closed #21 issue in github with error objects vs error strings - solved with a type coercion
v0.30.2 2014-12-25 21:25:32-08:00 America/Los_Angeles
! Closed #20 in github (pull request for undef FileType message specificity)
+ Closed #3 in github Move Temp::Dir methods to Archive::Zip in github
This is a big change since it changes some of the extensibility API's
It also changes the way the initial file is locked. It used to unzip the file to a
temporory directory and then unlock the file. Now it holds a lock on the file during the whole read.
! Closed #18 in github. The temp dir management is very different
The failing solo SciTE GUI test 03-temp_dir_bug.t now passes.
v0.30.0 2014-12-15 12:51:23-08:00 America/Los_Angeles
+ Closed #16 in github by moving the role Spreadsheet::XLSX::Reader::LibXML::LogSpace to Log::Shiras::LogSpace
and moving all the attribute callouts behind ###LogSD comments
+ Closed #19 in github by expanding the allowed width of the match data reference pointer when checking the
answer for specific errors
v0.28.2 2014-12-13 20:30:26-08:00 America/Los_Angeles
! Fixed #15 issue in github (Return precedence error)
- Closed #2 issue in github (Remove the OLE::Storage_lite dependancy)
! Closes #17 in github (Allow parse to accept either a file name or a file handle)
+ Added a file_handle attribute to support passing either file handles or file names
+ Improved the test file 01-types.t to include coercions and error message checking
v0.26.2 2014-12-11 08:57:26-08:00 America/Los_Angeles
! Fixed the 'requires missing role' bug (github issue -> missing role?)
v0.24.2 2014-12-09 05:18:13-08:00 America/Los_Angeles
+ added the empty_return_type attribute so that this package will emulate ParseExcel
You can still choose to return undef by setting this attribute to 'undef_string'
! added another fix in the cleanup since it forced a fail when fails weren't needed
v0.22.2 2014-12-04 15:46:15-08:00 America/Los_Angeles
! Fixed a bug in the error reporting for the value method in the cell
! Fixed a bug where the sheet position wasn't set correctly when calling for worksheet generation by name
added a test file for this case
! Improved temp dir cleanup in error situations - with less code
this inclues a new explicit DEMOLISH at the cell level
v0.20.4 2014-12-03 13:27:23-08:00 America/Los_Angeles
+ A round of POD scrubbing for POD completed to this point
v0.20.2 2014-12-02 12:56:46-08:00 America/Los_Angeles
+ Finished the POD for ParseExcelFormatStrings.pm
+ Merged UtilFunctions with ParseExcelFormatStrings since there was no reason
for them to be separate
+ Improved the POD for GetCell.pm
+ made the methods for clearing the row reading start and end column definitions public
v0.18.2 2014-11-30 13:54:00-08:00 America/Los_Angeles
! Although the last update had more passes on CPAN testers this is another update
attempting to reach 100% pass rate
v0.16.2 2014-11-29 16:43:29-08:00 America/Los_Angeles
+ Some tweaks to the LibXML POD
! Another re-write of the ParseExcelFormatStrings.pm to try and fix CPAN Testers fails
v0.14.2 2014-11-26 21:52:22-08:00 America/Los_Angeles
! Fixed a 32bit precision issue
v0.12.4 2014-11-26 06:45:17-08:00 America/Los_Angeles
! Removed the regex match within a regex test to pass older perls in test
v0.12.2 2014-11-26 06:06:30-08:00 America/Los_Angeles
+ Improved the FmtDefault POD
+ Improved the LibXML POD
+ Extracted several elements of cruft from Types
+ Reworked the ParseExcelFormatStrings to run on Dispatch Tables for improved consistency
and to hopefully address CPAN testers fails
+ moved the change_output_encoding out of ParseExcelFormatStrings an into the cell and value
v0.10.6 2014-11-18 16:46:07-08:00 America/Los_Angeles
! Fixed several errors in the CellToColumnRow POD
+ Improved the Error POD
+ More tweaks to the LibXML POD
! Fixed a precision bug in ParseExcelFormatStrings
calls for more universal implementation.
v0.10.4 2014-11-16 09:35:36-08:00 America/Los_Angeles
+ Improved CellToColumnRow POD
+ Improved the top level POD
+ Improved the Worksheet POD
! Removed the SVPerlTalk.odp from the package (as bloat)
Look in the github repository if you want to review it
+ Added a new POD file for the SharedStrings class (to be universal across Reader and DOM)
+ Added a new POD file for the Styles class (to be universal across Reader and DOM)
+ Added a new POD file for the CalcChain class (to be universal across Reader and DOM)
v0.10.2 2014-11-13 10:29:27-08:00 America/Los_Angeles
+ Fixed and updated the POD based on a MetaCPAN read
v0.10.1 2014-11-13 05:37:32-08:00 America/Los_Angeles
+ Finished the Worksheet POD
+ Completed the Cell POD
+ Completed the Error POD
+ Completed the CellToColumnRow POD
+ Another (but not final0 proofread of the Spreadsheet::XLSX::Reader::LibXML POD
+ Release with a Beta rev to try and Index on MetaCPAN
v0.7_1 2014-11-12 07:41:39-08:00 America/Los_Angeles
+ POD for the main module complete
+ Improved POD for the worksheet module
v0.5_1 2014-11-07 08:23:08-08:00 America/Los_Angeles
+ Initial near CPAN ready release with a full test suit for cross platform testing
+ Changed the Cell class to be just an encapulated data holder
! Fixed the resources links in the META files
+ Used type support from Type::Coercions to allow for data transformation definitions
! The documentation is all wrong. Next release I promise!
! Reader parsing support only
- DOM parsing contract dependant or some (possibly distant) time in the future
v0.1_1 2014-06-06 10:06:00-07:00 America/Los_Angeles
+ Initial (non-functional) github release