* Start extracting Logical Excel functions into separate groups in dedicated classes
* Extracting remaining Logical Excel functions into separate groups in dedicated classes
* Support 'Forms' for ROMAN Function
This seems like an exceptionally silly thing for MS to have implemented
(Wikipedia on Roman Numerals: "There is no indication this is anything
other than an invention by the programmer").
Nevertheless, we can, and therefore probably should, implement it.
Not that I can implement it by an algorithm - Excel describes the various extra
styles as "more concise", "more concise", "more concise", and "simplified".
Nevertheless, since the universe of potential calls is relatively small,
it can be implemented as a table of values where the new forms would return
a different value than "classic". This table is relatively large, so I have
put it its own member to avoid overhead when the function is needed.
* Move ROMAN To Its Own Class
See discussion in PR #1837
* PHP 8.1 Deprecations
PHP8.1 Unit tests failed. 1 line fixes are available for
- Shared/Font
- Shared/XMLWriter
- Style/Color
- Writer/HTML
The problem is that an error is also reported for a strcmp at
line 272 of Cell/Cell. Not only does that line not invoke strcmp,
there is no strcmp in all of Cell/Cell, so I don't know what to make
of the error message. Oh well, let's fix what can be fixed.
Still dealing with the mysterious PHP8.1 unit test failure in Cell\Cell,
which seems to have something to do with strcmp. The only uses of
strcmp that I can find in src/ are in Calculation. I can't find any
use of it in test/ or samples/. So, if this doesn't fix the problem,
I may have to give up.
* ROUND Accepts null, false, and true as First Parameter
Issue #1789 was addressed by PR #1799. In a follow-up discussion,
it came to light that ROUND was not handling the unexpected case where the
first parameter is an empty cell in the same manner that Excel does.
Subsequent investigation showed that a boolean first parameter is permitted.
I broadened my investigation to include the following related functions.
- ROUNDUP
- ROUNDDOWN
- MROUND
- TRUNC
- INT
- FLOOR
- FLOOR.MATH
- FLOOR.PRECISE
- CEILING
- CEILING.MATH
- CEILING.PRECISE
All of these allow a NULL first parameter, and all except MROUND allow boolean.
For completeness, I will note that all treat null string as invalid.
I suspect there are other functions which permit
similarly unexpected parameters, but I consider them out of scope for this PR.
CEILING.MATH and CEILING.PRECISE were unimplemented, and are now supported
as part of this PR.
The tests for each of these functions have been re-coded, though all the original
test data is still included in the test cases, plus several new cases for each.
The new tests now take place as a user would invoke the functions,
through a spreadsheet cell rather than a
direct call to the appropriate function within Calculation/MathTrig.
Aside from being more realistic, the new tests are also more complete.
For example, FLOOR.MATH can take from 1-3 arguments, and the existing tests
confirmed that the function in Calculation could handle a single argument.
However, the function list in Calculation.php erroneously set the number of
arguments for FLOOR.MATH to exactly 3, so, if a user tried to get the calculated
result of a cell containing FLOOR.MATH(1.2), the result would be an Exception.
Aside from the parameter support, there are a few minor code changes.
Ods, as well as Gnumeric, allows the omission of the second parameter for
FLOAT and CEILING; Excel does not. A potential divide-by-zero error is
avoided in CEILING, FLOOR, and FLOORMATH.
I will note that it would probably be beneficial in terms of maintainability
to break MathTrig up into many individual modules. The same would hold for the
other Calculation modules. I would be willing to look into this if you agree
that it would be worthwhile.
* Extract DELTA() and GESTEP() functions from the Engineering class into a dedicated Comparison classes
Retain the original methods in the Engineering class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
Note that unit tests still point to the Engineering class stubs; these should be modified to use the Erf and ErfC classes directly when the stubs are removed
* Extract Permutation functions from the Statistical class into a dedicated Permutations class
Retain the original methods in the Statistical class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
Note that unit tests still point to the Statistical class stubs; these should be modified to use the Permutations class directly when the stubs are removed
Also provided a basic implementationof the PERMUTATIONA() Function
* Extract all Error functions from the Engineering class into a dedicated Erf and ErfC classes
Retain the original methods in the Engineering class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
Note that unit tests still point to the Engineering class stubs; these should be modified to use the Erf and ErfC classes directly when the stubs are removed
* Reminder that ERF is used (either directly or Indirectly) in some of the statistical functions as well
* Extract all Base Conversion functions from the Engineering class into a dedicated Convert<Base> classes extending from a common ConvertBase class
Retain the original methods in the Engineering class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
Note that unit tests still point to the Engineering class stubs; these should be modified to use the Convert<Base> classes directly when the stubs are removed
* Split out into separate base conversion classes, with a ConvertBase class for common methods
* Extract all BitWise functions from the Engineering class into a dedicated BitWise class
Retain the original methods in the Engineering class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
Note that unit tests still point to the Engineering class stubs; these should be modified to use the bessel classes directly when the stubs are removed
* Move Bessel function calculations from the Engineering class to a dedicated Engineering\Bessel class
Retain the original methods in the Engineering class as stubs for BC, but deprecate them. They will be removed for PHPSpreadsheet v2
* Some refactoring of the Bessel calculation logic
* Fix callable for ConvertUOM()
* Attempt to provide allow failure for PHP8.1 unit tests
PHP8.1 Tests show as passed despite the errors, and it requires checking the actual output from the run to see what the rea result is; but I can live with that until github provides functionality for a proper allow_failure option
* Stacked Alignment - Use Class Constant Rather than Literal
PR #1580 defined constants for "stacked" alignment in cells.
Using those constants outside of Style/Alignment was beyond the
scope of the original PR, but I said I would get to it.
This PR replaces all uses of literal -165, and appropriate uses of
literal 255, with the named constants, and adds tests to make sure
that the changed code is covered in the test suite.
* use axPos value to determine whether an axis title is mapped to the XaxisLabel or YaxisLabel
* update changelog
* Fix php-cs-fixer violations
Co-authored-by: Darren Maczka <dkm@utk.edu>
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
* Add support for Google Sheets Exported XLSX Charts
Google Sheets XLSX charts use oneCellAnchor positioning and the data series
do not have the *Cache elements with cached values.
* update CHANGELOG
* Add support for Google Sheets Exported XLSX Charts
Google Sheets XLSX charts use oneCellAnchor positioning and the data series
do not have the *Cache elements with cached values. Because the reader had been
assuming *Cache elements existed as children of strRef and numRef, errors about
the node being deleted were thrown when reading Xlsx exported from Google Sheets.
Co-authored-by: Darren Maczka <dkm@utk.edu>
The attribute `$alignmentXml` given to the private method `readAlignmentStyle` is the alignment XML tag that contains the alignment data itself. But inside that method all data are read from another `alignment` XML tag within that given tag. This redundant child-node access resulted in the following error- / notice-message: `PHP Notice: Trying to access array offset on value of type null in /foo/bar/baz/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Reader/Xlsx/Styles.php on line 146`
These changes remove the redundant child-node access in the method `readAlignmentStyle`.
* Treat inline strings like strings in Open Document because it has no specific inline-string format
* implement data-type error
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
* Make DefinedNames Samples Consistent With Other Samples (#1707)
All other Samples write to temporary directory. DefinedNames samples
write to main directory, which (a) means they aren't stored with others,
and (b) they aren't ignored by git so look like changed files.
The tests are also simplified by requiring Header rather than Bootstrap,
making use of Helper.
* Resolve XSS Vulnerability in the HTML Writer (#1719)
Resolve XSS Vulnerability in the HTML Writer
* Drop Travis
* Automatic GitHub releases from git tags
* Improve Coverage in src/PhpSpreadsheet
There are no changes to code. Additional tests are added,
so that the following 6 items now have 100% test coverage:
- Comment
- DefinedName
- DocumentGenerator
- IOFactory
- NamedFormula
- NamedRange
* Changes for Scrutinizer
Two changes to fix minor problems reported by Scrutinizer.
* Spelling: Tou -> You
* Fix for 1735 (Incorrect activeSheetIndex after RemoveSheetByIndex) (#1743)
This is a fix for issue #1735.
It adds tests for this situation, and similar situations involving
adding new sheets and accessing existing ones.
Coverage for Spreadsheet.php increases from 69% to 75% as a result.
* Update change log
* Fix for 3 Issues Involving ReadXlsx and NamedRange (#1742)
* Fix for 3 Issues Involving ReadXlsx and NamedRange
Issues #1686 and #1723, which provide sample spreadsheets, are probably
solved by this ticket. Issue #1730 is also probably solved, but I have
no way to verify.
There are two problems with how PhpSpreadsheet is handling things now.
Although the first problem is much less severe, and isn't really a factor
in the issues named above, it is helpful to get it out of the way first.
If you define a named range in Excel, and then delete the sheet where
the range exists, Excel saves the range as #REF!. If there is a cell which
references the range, it will similarly have the value #REF! when you open
the Excel file.
Currently, PhpSpreadsheet discards the #REF! definition, so a cell which
references the range will appear as #NAME? rather than #REF!.
This PR changes the behavior so that PhpSpreadsheet retains the #REF!
definition, and cells which reference it will appear as #REF!.
The second problem is the more severe, and is, I believe, responsible
for the 3 issues identified above.
If you define a named range and the sheet on which the range is defined
does not exist at the time, Excel will save the range as something like:
'[1]Unknown Sheet'!$A$1
If a cell references such a range, Excel will again display #REF!.
PhpSpreadsheet currently throws an Exception when it encounters
such a definition while reading the file. This PR changes
the behavior so that PhpSpreadsheet saves the definition as #REF!,
and cells which reference it will behave similarly.
For the record, I will note that Excel does not magically recalculate when a
missing sheet is subsequently added, despite the fact that the reference
might now become resolvable. PhpSpreadsheet behaves likewise.
* Remove Dead Code in Test
Identified it after push but before merge.
* Update change log
* Apply Column and Row Styles to Existing Cells (#1721)
* Apply Column and Row Styles to Existing Cells
This is a fix for issue #1712.
When a style is applied to an entire row or column, it is currently
only effective for cells which don't already contain a value.
The code needs to iterate through existing cells in the row/column
in order to apply the style to them.
This could be considered a breaking change, however, I believe that
the change makes things operate as users would expect, and that the
existing implementation is incomplete.
The change also removes protected element conditionalStyles from
the Style class. That element is an unused remnant, and can no longer be
set or retrieved - methods getConditionalStyles and setConditionalStyles
actually act on an element in the Worksheet class.
Finally, additional tests are added so that Style, and in fact the
entire Style directory, now has 100% test coverage.
* Scrutinizer Changes
Scrutinizer flagged 6 statements. 5 can be easily corrected.
One is absolutely wrong (it thinks iterating through cells in column
can return null). Let's see if we can satisfy it.
* Remove Exception For CellIterator on Empty Row/Column
For my first attempt at this change, which corrects a bug by updating styles
for non-empty cells when a style is set on a row or column, I wished to make things
more efficient by using setIterateOnlyExistingCells, something which the
existing documentation recommends. This caused an exception to be generated
when the row or column is empty. So I removed that part of the change while I
researched what was going on.
I have completed that research. The existing code does throw an exception
when the row/column is empty and iterateOnlyExistingCells is true. However,
that does not seem like a reasonable action. This situation is analagous to
iterating over an empty array, and that action is legal and does not throw.
The same should apply here. There were no tests for this situation,
and now there are.
I have added additional tests, and coverage for all of RowCellIterator,
ColumnCellIterator, and CellIterator are all now 100%. Some of my new tests
were added in new members, because the existing tests all relied on mocking,
which was not the best choice for the new tests. One of the existing tests
for RowCellIteratorTest (testSeekOutOfRange) was wrong; it issued the expected
exception, but for the wrong reason. I have added an additional test to
ensure that it fails "correctly".
The existing documentation says that the default value for
IterateOnlyExistingCells is true. In fact, the default value is false.
I have corrected the documentation.
* More Scrutinizer
I believe its analysis is incorrect, but this should silence it.
* DocBlock Correction
ColumnCellIterator DocBlock for current indicated it could return null
or Cell, but it can really return only Cell. This had caused Scrutinizer
to complain earlier.
* PHP8 Environment Appears to be Fixed
Cosmetic change to Doc member. I suspect there is a way to rerun all
the tests without another push, but I have been unable to figure out how.
* Update change log
* TextData Coverage and Minor Bug Fixes (#1744)
This had been intended to get 100% coverage for TextData functions, and it does that.
However, some minor bugs requiring source changes arose during testing.
- the Excel CHAR function restricts its argument to 1-255. PhpSpreadsheet CHARACTER
had been allowing 0+. Also, there is no need to test if iconv exists,
since it is part of Composer requirements.
- The DOLLAR function had been returning NUM for invalid arguments. Excel returns VALUE.
Also, negative amounts were not being handled correctly.
- The FIXEDFORMAT function had been returning NUM for invalid arguments. Excel FIXED returns VALUE.
* Replace anti-xss with html purifier (#1751)
* Replace voku/anti-xss with ezyang/htmlpurifier. Despite anti-xss being a smaller footprint dependency, an a better license fit with our MIT license, there are issues with it's automatic it sanitisation of global variables causing side effects
* Additional unit tests for xss in html writer cell comments
* Fix bug #1626 where values of 0 were "rounded" up/down as if they were not 0 (#1627)
* Fix bug where values of 0 were "rounded" up/down as if they were not 0
* Update change log
* Fix for #1612 - SLK Long File Name (#1706)
Issue has been marked stale, but ...
Sylk read sets worksheet title to filename (minus .slk).
If that is >31 characters, PhpSpreadsheet throws Exception.
This change truncates sheet title, as Excel does, to 31 characters.
* Update change log
* worksheet: fix if cellValue does not exist (#1727)
The condition is FALSE if the cell does not exist in the flipped table,
but anyway, it is sent in to a method requiring 'string' type, causing
it to fail.
* fixes#1655 issue (#1656)
Resolve problem with incorrectly defined hyperlinks
* Add 'ps' suffix to printer settings resources IDs (#1690)
* Add 'ps' suffix to printer settings resources IDs
* Update change log
* Fix pixelsToPoints conversion (for HTML col width) (#1733)
* DocBlock Change in Styles/Conditional (#1697)
Scrutinizer reported a minor error in a test involving a module
which I was not changing.
Styles/Conditional function setConditions can take a scalar or an
array as a parameter, but DocBlock says it only expects array.
I did not wish to add the extra module to my PR, but made a note to
self to fix that after PR was installed.
That has now happened, and it makes for a good case for me
to see all the PHP8/Composer2/etc. changes that have happened recently.
* Merge pull request #1698
* Merge pull request #4 from PHPOffice/master
* Restore Omitted Read XML Test
* Fix for bug #1592 (UPDATED) (#1623)
* Fix for Xls when BIFF8 SST (FCh) has bad Shared string length
* Update change log
* Add nightly PHP 8.1 dev to github actions (#1763)
* Fix compatibility with ext-gd on php 8 (#1762)
* CSV - Guess Encoding, Handle Null-string Escape (#1717)
* CSV - Guess Encoding, Handle Null-string Escape
This is in response to issue #1647 (detect CSV character encoding).
First, my tests with mb_detect_encoding indicate that it doesn't work
well enough; regardless, users can always do that on their own
if they deem it useful.
Rolling my own is also troublesome, but I can at least:
a. Check for BOM (UTF-8, UTF-16BE, UTF-16LE, UTF-32BE, UTF-32LE).
b. Do some heuristic tests for each of the above encodings.
c. Fallback to a user-specified encoding (default CP1252)
if a and b don't yield result.
I think this is probably useful enough to include, and relatively
easy to expand if other potential encodings should be considered.
Starting with PHP7.4, fgetcsv allows specification of null string as
escape character in fgetcsv. This is a much better choice than the PHP
(and PhpSpreadsheet) default of backslash in that it handles the file
in the same manner as Excel does. There is one statement in Reader/CSV
which would be adversely affected if the caller so specified (building
a regular expression under the assumption that escape character is
a single character). Fix that statement appropriately and add tests.
* Update changelog
* Update Units of Measure supported by the CONVERT() function (#1768)
Now supports all current UoM in all categories, with both 1- and 2-character multiplier prefixes, and binary multiplier prefixes, including the new Temperature scales
* Changelog for 1.16.0 release
* Fix date tests withut specified year for current year 2021 (#1774)
* Mrand of zero to any multiple should return 0 (#1773)
* Inherited Scrutinizer Recommendations - 3 of 3
I tried to sync my fork with the main project, as I have done several times before.
However, the GitHub interface to do this had changed, and it appears that I
did not make the optimal selection when I had a choice. Consequently, all
the merges that happened to base between the last time I synchronized and this time
appear to be part of any PR that I push. "Files changed" remains correct for
my new PRs, but there appear to be many more commits involved than is actually the
case. I will, at some point, delete and re-create my fork, and pay much closer
attention in future when I want to sync my fork with the main project.
Because of this set-up, Scrutinizer reports flaws in code that I haven't
actually changed in my PRs #1799 and #1800.
It still passes them, but, as long as I'm aware of the problems,
I may as well attempt to correct them. The following are not part of those PRs:
- 5 problems spread over 4 different members
- 12 problems in Calculation/Engineering
- 15 problems in Reader/XML
I shall attempt to resolve these via 3 separate PRs,
of which this is the third.
* The Usual
Fixed some Scrutinizer problems, some new ones popped up.
* More Scrutinizer
I think it's wrong in a lot of these cases. Although I am working
around all of them, I intend to file a bug report with them.
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
Co-authored-by: Adrien Crivelli <adrien.crivelli@gmail.com>
Co-authored-by: Ryan McAllen <rhyno@rhynodesigns.com>
Co-authored-by: Flinsch <220455+Flinsch@users.noreply.github.com>
Co-authored-by: Jan Sverre Riksfjord <jasverix@gmail.com>
Co-authored-by: Max Kalyabin <maksim@kalyabin.ru>
Co-authored-by: Sébastien Despont <sebastien.despont@gmail.com>
Co-authored-by: Guilliam Xavier <guilliamxavier@users.noreply.github.com>
Co-authored-by: Gianluca Giovinazzo <g.giovinazzo@gmail.com>
Co-authored-by: Alexander M. Turek <me@derrabus.de>
Co-authored-by: Martins Sipenko <martins.sipenko@gmail.com>
* Make DefinedNames Samples Consistent With Other Samples (#1707)
All other Samples write to temporary directory. DefinedNames samples
write to main directory, which (a) means they aren't stored with others,
and (b) they aren't ignored by git so look like changed files.
The tests are also simplified by requiring Header rather than Bootstrap,
making use of Helper.
* Resolve XSS Vulnerability in the HTML Writer (#1719)
Resolve XSS Vulnerability in the HTML Writer
* Drop Travis
* Automatic GitHub releases from git tags
* Improve Coverage in src/PhpSpreadsheet
There are no changes to code. Additional tests are added,
so that the following 6 items now have 100% test coverage:
- Comment
- DefinedName
- DocumentGenerator
- IOFactory
- NamedFormula
- NamedRange
* Changes for Scrutinizer
Two changes to fix minor problems reported by Scrutinizer.
* Spelling: Tou -> You
* Fix for 1735 (Incorrect activeSheetIndex after RemoveSheetByIndex) (#1743)
This is a fix for issue #1735.
It adds tests for this situation, and similar situations involving
adding new sheets and accessing existing ones.
Coverage for Spreadsheet.php increases from 69% to 75% as a result.
* Update change log
* Fix for 3 Issues Involving ReadXlsx and NamedRange (#1742)
* Fix for 3 Issues Involving ReadXlsx and NamedRange
Issues #1686 and #1723, which provide sample spreadsheets, are probably
solved by this ticket. Issue #1730 is also probably solved, but I have
no way to verify.
There are two problems with how PhpSpreadsheet is handling things now.
Although the first problem is much less severe, and isn't really a factor
in the issues named above, it is helpful to get it out of the way first.
If you define a named range in Excel, and then delete the sheet where
the range exists, Excel saves the range as #REF!. If there is a cell which
references the range, it will similarly have the value #REF! when you open
the Excel file.
Currently, PhpSpreadsheet discards the #REF! definition, so a cell which
references the range will appear as #NAME? rather than #REF!.
This PR changes the behavior so that PhpSpreadsheet retains the #REF!
definition, and cells which reference it will appear as #REF!.
The second problem is the more severe, and is, I believe, responsible
for the 3 issues identified above.
If you define a named range and the sheet on which the range is defined
does not exist at the time, Excel will save the range as something like:
'[1]Unknown Sheet'!$A$1
If a cell references such a range, Excel will again display #REF!.
PhpSpreadsheet currently throws an Exception when it encounters
such a definition while reading the file. This PR changes
the behavior so that PhpSpreadsheet saves the definition as #REF!,
and cells which reference it will behave similarly.
For the record, I will note that Excel does not magically recalculate when a
missing sheet is subsequently added, despite the fact that the reference
might now become resolvable. PhpSpreadsheet behaves likewise.
* Remove Dead Code in Test
Identified it after push but before merge.
* Update change log
* Apply Column and Row Styles to Existing Cells (#1721)
* Apply Column and Row Styles to Existing Cells
This is a fix for issue #1712.
When a style is applied to an entire row or column, it is currently
only effective for cells which don't already contain a value.
The code needs to iterate through existing cells in the row/column
in order to apply the style to them.
This could be considered a breaking change, however, I believe that
the change makes things operate as users would expect, and that the
existing implementation is incomplete.
The change also removes protected element conditionalStyles from
the Style class. That element is an unused remnant, and can no longer be
set or retrieved - methods getConditionalStyles and setConditionalStyles
actually act on an element in the Worksheet class.
Finally, additional tests are added so that Style, and in fact the
entire Style directory, now has 100% test coverage.
* Scrutinizer Changes
Scrutinizer flagged 6 statements. 5 can be easily corrected.
One is absolutely wrong (it thinks iterating through cells in column
can return null). Let's see if we can satisfy it.
* Remove Exception For CellIterator on Empty Row/Column
For my first attempt at this change, which corrects a bug by updating styles
for non-empty cells when a style is set on a row or column, I wished to make things
more efficient by using setIterateOnlyExistingCells, something which the
existing documentation recommends. This caused an exception to be generated
when the row or column is empty. So I removed that part of the change while I
researched what was going on.
I have completed that research. The existing code does throw an exception
when the row/column is empty and iterateOnlyExistingCells is true. However,
that does not seem like a reasonable action. This situation is analagous to
iterating over an empty array, and that action is legal and does not throw.
The same should apply here. There were no tests for this situation,
and now there are.
I have added additional tests, and coverage for all of RowCellIterator,
ColumnCellIterator, and CellIterator are all now 100%. Some of my new tests
were added in new members, because the existing tests all relied on mocking,
which was not the best choice for the new tests. One of the existing tests
for RowCellIteratorTest (testSeekOutOfRange) was wrong; it issued the expected
exception, but for the wrong reason. I have added an additional test to
ensure that it fails "correctly".
The existing documentation says that the default value for
IterateOnlyExistingCells is true. In fact, the default value is false.
I have corrected the documentation.
* More Scrutinizer
I believe its analysis is incorrect, but this should silence it.
* DocBlock Correction
ColumnCellIterator DocBlock for current indicated it could return null
or Cell, but it can really return only Cell. This had caused Scrutinizer
to complain earlier.
* PHP8 Environment Appears to be Fixed
Cosmetic change to Doc member. I suspect there is a way to rerun all
the tests without another push, but I have been unable to figure out how.
* Update change log
* TextData Coverage and Minor Bug Fixes (#1744)
This had been intended to get 100% coverage for TextData functions, and it does that.
However, some minor bugs requiring source changes arose during testing.
- the Excel CHAR function restricts its argument to 1-255. PhpSpreadsheet CHARACTER
had been allowing 0+. Also, there is no need to test if iconv exists,
since it is part of Composer requirements.
- The DOLLAR function had been returning NUM for invalid arguments. Excel returns VALUE.
Also, negative amounts were not being handled correctly.
- The FIXEDFORMAT function had been returning NUM for invalid arguments. Excel FIXED returns VALUE.
* Replace anti-xss with html purifier (#1751)
* Replace voku/anti-xss with ezyang/htmlpurifier. Despite anti-xss being a smaller footprint dependency, an a better license fit with our MIT license, there are issues with it's automatic it sanitisation of global variables causing side effects
* Additional unit tests for xss in html writer cell comments
* Fix bug #1626 where values of 0 were "rounded" up/down as if they were not 0 (#1627)
* Fix bug where values of 0 were "rounded" up/down as if they were not 0
* Update change log
* Fix for #1612 - SLK Long File Name (#1706)
Issue has been marked stale, but ...
Sylk read sets worksheet title to filename (minus .slk).
If that is >31 characters, PhpSpreadsheet throws Exception.
This change truncates sheet title, as Excel does, to 31 characters.
* Update change log
* worksheet: fix if cellValue does not exist (#1727)
The condition is FALSE if the cell does not exist in the flipped table,
but anyway, it is sent in to a method requiring 'string' type, causing
it to fail.
* fixes#1655 issue (#1656)
Resolve problem with incorrectly defined hyperlinks
* Add 'ps' suffix to printer settings resources IDs (#1690)
* Add 'ps' suffix to printer settings resources IDs
* Update change log
* Fix pixelsToPoints conversion (for HTML col width) (#1733)
* DocBlock Change in Styles/Conditional (#1697)
Scrutinizer reported a minor error in a test involving a module
which I was not changing.
Styles/Conditional function setConditions can take a scalar or an
array as a parameter, but DocBlock says it only expects array.
I did not wish to add the extra module to my PR, but made a note to
self to fix that after PR was installed.
That has now happened, and it makes for a good case for me
to see all the PHP8/Composer2/etc. changes that have happened recently.
* Merge pull request #1698
* Merge pull request #4 from PHPOffice/master
* Restore Omitted Read XML Test
* Fix for bug #1592 (UPDATED) (#1623)
* Fix for Xls when BIFF8 SST (FCh) has bad Shared string length
* Update change log
* Add nightly PHP 8.1 dev to github actions (#1763)
* Fix compatibility with ext-gd on php 8 (#1762)
* CSV - Guess Encoding, Handle Null-string Escape (#1717)
* CSV - Guess Encoding, Handle Null-string Escape
This is in response to issue #1647 (detect CSV character encoding).
First, my tests with mb_detect_encoding indicate that it doesn't work
well enough; regardless, users can always do that on their own
if they deem it useful.
Rolling my own is also troublesome, but I can at least:
a. Check for BOM (UTF-8, UTF-16BE, UTF-16LE, UTF-32BE, UTF-32LE).
b. Do some heuristic tests for each of the above encodings.
c. Fallback to a user-specified encoding (default CP1252)
if a and b don't yield result.
I think this is probably useful enough to include, and relatively
easy to expand if other potential encodings should be considered.
Starting with PHP7.4, fgetcsv allows specification of null string as
escape character in fgetcsv. This is a much better choice than the PHP
(and PhpSpreadsheet) default of backslash in that it handles the file
in the same manner as Excel does. There is one statement in Reader/CSV
which would be adversely affected if the caller so specified (building
a regular expression under the assumption that escape character is
a single character). Fix that statement appropriately and add tests.
* Update changelog
* Update Units of Measure supported by the CONVERT() function (#1768)
Now supports all current UoM in all categories, with both 1- and 2-character multiplier prefixes, and binary multiplier prefixes, including the new Temperature scales
* Changelog for 1.16.0 release
* Fix date tests withut specified year for current year 2021 (#1774)
* Mrand of zero to any multiple should return 0 (#1773)
* Inherited Scrutinizer Recommendations - 2 of 3
I tried to sync my fork with the main project, as I have done several times before.
However, the GitHub interface to do this had changed, and it appears that I
did not make the optimal selection when I had a choice. Consequently, all
the merges that happened to base between the last time I synchronized and this time
appear to be part of any PR that I push. "Files changed" remains correct for
my new PRs, but there appear to be many more commits involved than is actually the
case. I will, at some point, delete and re-create my fork, and pay much closer
attention in future when I want to sync my fork with the main project.
Because of this set-up, Scrutinizer reports flaws in code that I haven't
actually changed in my PRs #1799 and #1800.
It still passes them, but, as long as I'm aware of the problems,
I may as well attempt to correct them. The following are not part of those PRs:
- 5 problems spread over 4 different members
- 12 problems in Calculation/Engineering
- 15 problems in Reader/XML
I shall attempt to resolve these via 3 separate PRs,
of which this is the second.
* Fixed Most of the Problems, But Some New Ones Cropped Up
Trying once more to satisfy Scrutinizer.
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
Co-authored-by: Adrien Crivelli <adrien.crivelli@gmail.com>
Co-authored-by: Ryan McAllen <rhyno@rhynodesigns.com>
Co-authored-by: Flinsch <220455+Flinsch@users.noreply.github.com>
Co-authored-by: Jan Sverre Riksfjord <jasverix@gmail.com>
Co-authored-by: Max Kalyabin <maksim@kalyabin.ru>
Co-authored-by: Sébastien Despont <sebastien.despont@gmail.com>
Co-authored-by: Guilliam Xavier <guilliamxavier@users.noreply.github.com>
Co-authored-by: Gianluca Giovinazzo <g.giovinazzo@gmail.com>
Co-authored-by: Alexander M. Turek <me@derrabus.de>
Co-authored-by: Martins Sipenko <martins.sipenko@gmail.com>
* Make DefinedNames Samples Consistent With Other Samples (#1707)
All other Samples write to temporary directory. DefinedNames samples
write to main directory, which (a) means they aren't stored with others,
and (b) they aren't ignored by git so look like changed files.
The tests are also simplified by requiring Header rather than Bootstrap,
making use of Helper.
* Resolve XSS Vulnerability in the HTML Writer (#1719)
Resolve XSS Vulnerability in the HTML Writer
* Drop Travis
* Automatic GitHub releases from git tags
* Improve Coverage in src/PhpSpreadsheet
There are no changes to code. Additional tests are added,
so that the following 6 items now have 100% test coverage:
- Comment
- DefinedName
- DocumentGenerator
- IOFactory
- NamedFormula
- NamedRange
* Changes for Scrutinizer
Two changes to fix minor problems reported by Scrutinizer.
* Spelling: Tou -> You
* Fix for 1735 (Incorrect activeSheetIndex after RemoveSheetByIndex) (#1743)
This is a fix for issue #1735.
It adds tests for this situation, and similar situations involving
adding new sheets and accessing existing ones.
Coverage for Spreadsheet.php increases from 69% to 75% as a result.
* Update change log
* Fix for 3 Issues Involving ReadXlsx and NamedRange (#1742)
* Fix for 3 Issues Involving ReadXlsx and NamedRange
Issues #1686 and #1723, which provide sample spreadsheets, are probably
solved by this ticket. Issue #1730 is also probably solved, but I have
no way to verify.
There are two problems with how PhpSpreadsheet is handling things now.
Although the first problem is much less severe, and isn't really a factor
in the issues named above, it is helpful to get it out of the way first.
If you define a named range in Excel, and then delete the sheet where
the range exists, Excel saves the range as #REF!. If there is a cell which
references the range, it will similarly have the value #REF! when you open
the Excel file.
Currently, PhpSpreadsheet discards the #REF! definition, so a cell which
references the range will appear as #NAME? rather than #REF!.
This PR changes the behavior so that PhpSpreadsheet retains the #REF!
definition, and cells which reference it will appear as #REF!.
The second problem is the more severe, and is, I believe, responsible
for the 3 issues identified above.
If you define a named range and the sheet on which the range is defined
does not exist at the time, Excel will save the range as something like:
'[1]Unknown Sheet'!$A$1
If a cell references such a range, Excel will again display #REF!.
PhpSpreadsheet currently throws an Exception when it encounters
such a definition while reading the file. This PR changes
the behavior so that PhpSpreadsheet saves the definition as #REF!,
and cells which reference it will behave similarly.
For the record, I will note that Excel does not magically recalculate when a
missing sheet is subsequently added, despite the fact that the reference
might now become resolvable. PhpSpreadsheet behaves likewise.
* Remove Dead Code in Test
Identified it after push but before merge.
* Update change log
* Apply Column and Row Styles to Existing Cells (#1721)
* Apply Column and Row Styles to Existing Cells
This is a fix for issue #1712.
When a style is applied to an entire row or column, it is currently
only effective for cells which don't already contain a value.
The code needs to iterate through existing cells in the row/column
in order to apply the style to them.
This could be considered a breaking change, however, I believe that
the change makes things operate as users would expect, and that the
existing implementation is incomplete.
The change also removes protected element conditionalStyles from
the Style class. That element is an unused remnant, and can no longer be
set or retrieved - methods getConditionalStyles and setConditionalStyles
actually act on an element in the Worksheet class.
Finally, additional tests are added so that Style, and in fact the
entire Style directory, now has 100% test coverage.
* Scrutinizer Changes
Scrutinizer flagged 6 statements. 5 can be easily corrected.
One is absolutely wrong (it thinks iterating through cells in column
can return null). Let's see if we can satisfy it.
* Remove Exception For CellIterator on Empty Row/Column
For my first attempt at this change, which corrects a bug by updating styles
for non-empty cells when a style is set on a row or column, I wished to make things
more efficient by using setIterateOnlyExistingCells, something which the
existing documentation recommends. This caused an exception to be generated
when the row or column is empty. So I removed that part of the change while I
researched what was going on.
I have completed that research. The existing code does throw an exception
when the row/column is empty and iterateOnlyExistingCells is true. However,
that does not seem like a reasonable action. This situation is analagous to
iterating over an empty array, and that action is legal and does not throw.
The same should apply here. There were no tests for this situation,
and now there are.
I have added additional tests, and coverage for all of RowCellIterator,
ColumnCellIterator, and CellIterator are all now 100%. Some of my new tests
were added in new members, because the existing tests all relied on mocking,
which was not the best choice for the new tests. One of the existing tests
for RowCellIteratorTest (testSeekOutOfRange) was wrong; it issued the expected
exception, but for the wrong reason. I have added an additional test to
ensure that it fails "correctly".
The existing documentation says that the default value for
IterateOnlyExistingCells is true. In fact, the default value is false.
I have corrected the documentation.
* More Scrutinizer
I believe its analysis is incorrect, but this should silence it.
* DocBlock Correction
ColumnCellIterator DocBlock for current indicated it could return null
or Cell, but it can really return only Cell. This had caused Scrutinizer
to complain earlier.
* PHP8 Environment Appears to be Fixed
Cosmetic change to Doc member. I suspect there is a way to rerun all
the tests without another push, but I have been unable to figure out how.
* Update change log
* TextData Coverage and Minor Bug Fixes (#1744)
This had been intended to get 100% coverage for TextData functions, and it does that.
However, some minor bugs requiring source changes arose during testing.
- the Excel CHAR function restricts its argument to 1-255. PhpSpreadsheet CHARACTER
had been allowing 0+. Also, there is no need to test if iconv exists,
since it is part of Composer requirements.
- The DOLLAR function had been returning NUM for invalid arguments. Excel returns VALUE.
Also, negative amounts were not being handled correctly.
- The FIXEDFORMAT function had been returning NUM for invalid arguments. Excel FIXED returns VALUE.
* Replace anti-xss with html purifier (#1751)
* Replace voku/anti-xss with ezyang/htmlpurifier. Despite anti-xss being a smaller footprint dependency, an a better license fit with our MIT license, there are issues with it's automatic it sanitisation of global variables causing side effects
* Additional unit tests for xss in html writer cell comments
* Fix bug #1626 where values of 0 were "rounded" up/down as if they were not 0 (#1627)
* Fix bug where values of 0 were "rounded" up/down as if they were not 0
* Update change log
* Fix for #1612 - SLK Long File Name (#1706)
Issue has been marked stale, but ...
Sylk read sets worksheet title to filename (minus .slk).
If that is >31 characters, PhpSpreadsheet throws Exception.
This change truncates sheet title, as Excel does, to 31 characters.
* Update change log
* worksheet: fix if cellValue does not exist (#1727)
The condition is FALSE if the cell does not exist in the flipped table,
but anyway, it is sent in to a method requiring 'string' type, causing
it to fail.
* fixes#1655 issue (#1656)
Resolve problem with incorrectly defined hyperlinks
* Add 'ps' suffix to printer settings resources IDs (#1690)
* Add 'ps' suffix to printer settings resources IDs
* Update change log
* Fix pixelsToPoints conversion (for HTML col width) (#1733)
* DocBlock Change in Styles/Conditional (#1697)
Scrutinizer reported a minor error in a test involving a module
which I was not changing.
Styles/Conditional function setConditions can take a scalar or an
array as a parameter, but DocBlock says it only expects array.
I did not wish to add the extra module to my PR, but made a note to
self to fix that after PR was installed.
That has now happened, and it makes for a good case for me
to see all the PHP8/Composer2/etc. changes that have happened recently.
* Merge pull request #1698
* Merge pull request #4 from PHPOffice/master
* Restore Omitted Read XML Test
* Fix for bug #1592 (UPDATED) (#1623)
* Fix for Xls when BIFF8 SST (FCh) has bad Shared string length
* Update change log
* Add nightly PHP 8.1 dev to github actions (#1763)
* Fix compatibility with ext-gd on php 8 (#1762)
* CSV - Guess Encoding, Handle Null-string Escape (#1717)
* CSV - Guess Encoding, Handle Null-string Escape
This is in response to issue #1647 (detect CSV character encoding).
First, my tests with mb_detect_encoding indicate that it doesn't work
well enough; regardless, users can always do that on their own
if they deem it useful.
Rolling my own is also troublesome, but I can at least:
a. Check for BOM (UTF-8, UTF-16BE, UTF-16LE, UTF-32BE, UTF-32LE).
b. Do some heuristic tests for each of the above encodings.
c. Fallback to a user-specified encoding (default CP1252)
if a and b don't yield result.
I think this is probably useful enough to include, and relatively
easy to expand if other potential encodings should be considered.
Starting with PHP7.4, fgetcsv allows specification of null string as
escape character in fgetcsv. This is a much better choice than the PHP
(and PhpSpreadsheet) default of backslash in that it handles the file
in the same manner as Excel does. There is one statement in Reader/CSV
which would be adversely affected if the caller so specified (building
a regular expression under the assumption that escape character is
a single character). Fix that statement appropriately and add tests.
* Update changelog
* Update Units of Measure supported by the CONVERT() function (#1768)
Now supports all current UoM in all categories, with both 1- and 2-character multiplier prefixes, and binary multiplier prefixes, including the new Temperature scales
* Changelog for 1.16.0 release
* Fix date tests withut specified year for current year 2021 (#1774)
* Mrand of zero to any multiple should return 0 (#1773)
* WIP Inherited Scrutinizer Recommendations - 1 of 3
I tried to sync my fork with the main project, as I have done several times before.
However, the GitHub interface to do this had changed, and it appears that I
did not make the optimal selection when I had a choice. Consequently, all
the merges that happened to base between the last time I synchronized and this time
appear to be part of any PR that I push. "Files changed" remains correct for
my new PRs, but there appear to be many more commits involved than is actually the
case. I will, at some point, delete and re-create my fork, and pay much closer
attention in future when I want to sync my fork with the main project.
Because of this set-up, Scrutinizer reports flaws in code that I haven't
actually changed in my PRs #1799 and #1800.
It still passes them, but, as long as I'm aware of the problems,
I may as well attempt to correct them. The following are not part of those PRs:
- 5 problems spread over 4 different members
- 12 problems in Calculation/Engineering
- 15 problems in Reader/XML
I shall attempt to resolve these via 3 separate PRs,
of which this is the first.
* Try Hyperlink Again
Scrutinizer still didn't like it as fixed.
* Another Crack at Hyperlink
This should work.
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
Co-authored-by: Adrien Crivelli <adrien.crivelli@gmail.com>
Co-authored-by: Ryan McAllen <rhyno@rhynodesigns.com>
Co-authored-by: Flinsch <220455+Flinsch@users.noreply.github.com>
Co-authored-by: Jan Sverre Riksfjord <jasverix@gmail.com>
Co-authored-by: Max Kalyabin <maksim@kalyabin.ru>
Co-authored-by: Sébastien Despont <sebastien.despont@gmail.com>
Co-authored-by: Guilliam Xavier <guilliamxavier@users.noreply.github.com>
Co-authored-by: Gianluca Giovinazzo <g.giovinazzo@gmail.com>
Co-authored-by: Alexander M. Turek <me@derrabus.de>
Co-authored-by: Martins Sipenko <martins.sipenko@gmail.com>
* Extract remaining Financial function unit tests into separate test classes for each function
This makes it easier to manage unit tests if they are individual files rather than all in a single file
It also provides a stepping stone toward making it easier to test Excel functions when Excel errors no longer return a string, but an actual Excel exception that can be handled more cleanly
* Additional unit tests for previously untested financial functions, and some additions to follow untested paths
* Start splitting Financial function tests out from the large FinancialTests class into individual test classes for each function
* Revert "Fix cant get right format chinese date format error"
This reverts commit 8c58385d6c.
* formatAsDate strip language metadata (fixes#1616)
Co-authored-by: Mark Baker <mark@lange.demon.co.uk>
Implemented the databar of Conditional Type for XLSX Files.
- DataBar can be read, written, and added for basic use.
- Supports reading, writing and adding using "extLst".
About "extLst"
- https://docs.microsoft.com/en-us/openspecs/office_standards/ms-xlsx/07d607af-5618-4ca2-b683-6a78dc0d9627
The following setting items on the Excel setting screen can be read, written, and added.
- (minimum, maximum)type: Automatic, LowestValue, Number, Percent, Formula, Percentile
- Direction: context, leftToRight, rightToLeft (show data bar only)
- Fills Solid, Gradient
- FillColor: PositiveValues, NegativeValues
- Borders: Solid, None
- BorderColor: PositiveValues, NegativeValues
- Axis position: Automatic, Midpoint, None
- Axis color
The previous exception will be included when loading the content as DOM Document fails. This makes debugging the reason behind the failure much easier.
Mark Baker
oleibman
ElPopcorn
Mats Sibelius
Adrien Crivelli
Roland Eigelsreiter
Darren Maczka
Gerrit Addiks
Alexander Gunkel
SheetJSDev
もりもと たかひろ
Jasper Zonneveld
Richard van Velzen