* R1C1 Format and Internationalization, plus Relative Offsets
Fix#1704, albeit imperfectly. Excel's implementation of this feature makes it impossible to fix perfectly. I don't know why it was necessary to internationalize R1C1 in the first place - the benefits are so minimal,and the result is worksheets that break when opened in different locales. Ugh. I can't even find complete documentation about the format in different languages; I am using https://answers.microsoft.com/en-us/officeinsider/forum/all/indirect-function-is-broken-at-least-for-excel-in/1fcbcf20-a103-4172-abf1-2c0dfe848e60 as my definitive reference.
This fix concentrates on the original report, using the INDIRECT function; there may be other areas similarly affected. As with ambiguous date formats, PhpSpreadsheet will do a little better than Excel itself when reading spreadsheets with internationalized R1C1 by trying all possibilities before giving up. When it does give up, it will now return `#REF!`, as Excel does, rather than throwing an exception, which is certainly friendlier. Although read now works better, when writing it will use whatever the user specified, so spreadsheets breaking in the wrong locale will still happen.
There were some bugs that turned up as I added test cases, all of them concerning relative addressing in R1C1 format, e.g. `R[+1]C[-1]`. The regexp for validating the format allowed for minus signs, but not plus signs. Also, the relevant functions did not allow for passing the current cell address, which made relative addressing impossible. The code now allows these, and suitable test cases are added.
* Use Locale for Formats, but Not for XML
Implementing a suggestion from @MarkBaker to use the system locale for determining R1C1 format rather than looping through a set of regexes and accepting any that work. This is closer to how Excel itself operates. The assumption we are making is to use the first character of the translated ROW and COLUMN functions. This will not work for Russian or Bulgarian, where each starts with the same letter, but it appears that Russian, at least, still uses R1C1. So our algorithm will not use non-ASCII characters, nor characters where ROW and COLUMN start with the same letter, falling back to R/C in those cases. Turkish falls into that category. Czech uses an accented character for one of the functions, and I'm guessing to use the unaccented character in that case. Polish COLUMN function is NR.KOLUMNY, and I'm guessing to use K in that case.
The function that converts R1C1 references is also used by the XML reader *where the format is always R1C1*, not locale-based (confirmed by successfully opening in Excel an XML spreadsheet when my language is set to French). The conversion code now handles that distinction through the use of an extra parameter. Xml Reader Load Test is duplicated to confirm that spreadsheet is loaded properly whether the locale is English or French. (No, I did not add an INDIRECT function to the Xml spreadsheet.)
Tests CsvIssue2232Test and TranslationTest both changed locale without resetting it when done. That omission was exposed by the new code, and both are now corrected.
* OpenOffice and Gnumeric
OpenOffice and Gnumeric make it much easier to test with other languages - they can be handled with an environment variable. Sensibly, they require R and C as the characters for R1C1 notation regardless of the language. Change code to recognize this difference from Excel.
* Handle Output of ADDRESS Function
One other function has to deal with R1C1 format as a string. Unlike INDIRECT, which receives the string on input, ADDRESS generates the string on output. Ensure that the ADDRESS output is consistent with the INDIRECT input.
ADDRESS expects its 4th arg to be bool, but it can also accept int, and many examples on the net supply it as an int. This had not been handled properly, but is now corrected.
* More Structured Test
I earlier introduced a new test for relative R1C1 addressing. Rewrite it to be clearer.
* Add Row for This to Locale Spreadsheet
It took a while for me to figure out how it all works. I have added a new row (with English value `*RC`) to Translations.xlsx, in the "Lookup and Reference" section of sheet "Excel Functions". By starting the "function name" with an asterisk, it will not be confused with a "real" function (confirmed by a new test). This approach also gives us the flexibility to do something similar if another surprise case occurs in future; in particular, I think this is more flexible than adding this as another option on the "Excel Localisation" sheet. It also means that any errors or omissions in the list below will be handled as with any other translation problem, by updating the spreadsheet without needing to touch any code.
The spreadsheet has the following entries in the *RC row:
- first letter of ROW/COLUMN functions for da, de, es, fi, fr, hu, nl, nb, pt, pt_br, sv
- no value for locales where ROW/COLUMN functions start with same letter - bg, ru, tr
- no value for locales with a multi-part name for ROW and/or COLUMN - it, pl (I had not previously noted Italian as an exception)
- no value for locales where ROW and/or COLUMN starts with a non-ASCII character - cs (this would also apply to bg and ru which are already included under "same letter")
- it does nothing for locales which are defined on the "Excel Localisation" sheet but have no entries yet on the "Excel Functions" sheet (e.g. eu)
Note that all but the first bullet item will continue to use R/C, which leaves them no worse off than they were before this change.
These changes have already been implemented twice, and been regressed twice. I'll try once more (with a different approach), then give up ...
As configured, Phpstan running under Php7 reports no errors. However, running under Php8, it reports 100 (!) errors. The vast majority of these are due to two reasons:
- renaming parameters in Php builtin functions in preparation for named parameters.
- using the new class GdImage rather than type resource as the argument type for many image-based functions.
Regardless of the cause, this will be a problem sooner or later. This PR is an attempt to get ahead of that problem. For source members, it mostly adds annotations or updates doc-blocks. Only 2 members have changes to executable code, and these are very minor - BitWise and Writer/Xlsx. For test members, all baseline errors are deleted and the code is fixed. Php7 and Php8 both report no errors with this configuration.
* Document Properties - Coverage and 32-bit-safe Timestamps
While researching an issue, I noticed that coverage of Document/Properties was poor. Further, the use of int timestamps will eventually lead to problems for 32-bit PHP (see issue #1826).
Coverage Changes:
- Many property types with no special handling are enumerated but not tested. These are removed, but will continue to function as before.
- Existing code theoretically allows property to be set to an object, but there is no means to read or write such a property, and, even if there were, I don't believe Excel supports it. Setting a property to an object will now be changed to a no-op (can throw an exception if preferred).
- Since the Properties object now has no members which are themselves objects, there is no need for a deep clone. The untested __clone method is removed.
- Large switch statements are replaced with associative arrays. Scrutinizer will like that.
- Coverage is now 100%.
<!-- end of coverage changes list -->
Timestamp Changes:
- Timestamps will be stored as int if possible, or float if not. This is, or will soon be, needed for 32-bit systems. Tests have been added for beyond-epoch dates, and run successfully with 32-bit.
- LibreOffice doesn't quite get the Created/Modified properties correct. These are written to the file as a string which includes offset from UTC, but LibreOffice ignores the offset portion when displaying them. Code had been generating these in UTC, but now generates them in default timezone, which should meet user's expectations.
<!-- end of timestamp changes list -->
Other Changes:
- Custom properties added to ODS Writer.
- Samples had not been generating any ODS files. One is now generated.
- Ods uses a single 'keywords' property rather than multiple 'keyword' properties.
- Breaking change - default company is changed to null string from Microsoft Corporation.
- Breaking change of sorts - PropertiesTest incorrectly tested a custom date property against a string, Reader/XlsxTest correctly tested against a timestamp converted to a string. PropertiesTest was defective, and will no longer work as coded; anyone using it as a model will likewise have a problem.
- PHP8.1 has been complaining for weeks about a time zone conversion test. I have now downloaded a version, and changed the code so that it will work in 8.1 as well as prior releases. (It is still likely that the existing code should work in 8.1, but I haven't yet figured out how to file a bug report.) In the course of testing, 3 additional 8.1 problems were reported (all along the lines of "can't pass null to strpos"), and are fixed with null coercion.
- Two Calculation tests failed because of large results on 32-bit system. These are corrected by allowing the functions involved to return float|int rather than int. I suspect that there are other functions with this problem, and will investigate as a follow-up activity.
- See issue #2090. I believe that changes between 17.1 and master will merely cause the problematic spreadsheet to fail in a different way. I believe that enclosing in quotes some variables passed to Document/Properties by Reader/Xlsx will eliminate the problem, but, in the absence of an example file, cannot say for sure.
- Properties tests are now separated out from Reader/XlsxTest and Reader/OdsTest, and now test both Read and Write (via reload).
<!-- end of other changes list -->
Miscellaneous Notes:
- There remains no support for Custom Properties in Xls Reader or Writer.
- We now have default timezones for all of PHP itself, Shared/Date, and Shared/Timezone. That is least one too many. I was unable to disentangle the latter two for this change, but will look into deprecating one or the other in future.
* Phpstan
6 baseline deletions, 2 docblock changes
* Scrutinizer's Turn
3 minor errors that hadn't blocked the request.
* Reader XML Properties - Eliminate strtotime
Piggyback on top of prior changes to eliminate 32-bit-unsafe call.
Add explicit tests for created, modified, and custom date properties.
I ran the test suite using 32-bit PHP. There were 2 places where changes
were needed due to 32-bit timestamps.
Reader\\Xml.php was using strtotime as an intermediate step in converting
a string timestamp to an Excel timestamp. The XML file type stores pure timestamps
(i.e. no date portion) as, e.g., 1899-12-31T02:30:00.000, and that value
causes an error using strtotime on a 32-bit system. However, it is sufficient
to use that value in a DateTime constructor, and that will work for 32- and 64-bit.
There was no test for that particular cell, so I added one to the XML read test.
And that's when I discovered the getFormattedValue bug. The cell's format
is `hh":"mm":"ss`. The quotes around the colons are disrupting the formatting.
PhpSpreadsheet formats the cell by converting the Excel format
to a Php Date format, in this case `H\:m\:s`.
That's a problem,
since Excel thinks 'm' means *minutes*, but PHP thinks it means *months*.
This is not a problem when the colon is not quoted; there are ample tests for that.
I added my best guess as to how to recognize this situation,
changing `\:m` to `:i`. The XML read test
now succeeds, and no other tests were broken by this change.
Test Shared\\DateTest had one test where the expected result of converting to a
Unix timestamp exceeds 2**32. Since a Unix timestamp is strictly an int,
that test fails on a 32-bit system. In the discussion regarding recently merged
PR #1870, it was felt that the user base might still be using the functions
that convert to and from a timestamp. So, we should not drop this test, but,
since it cannot succeed on a 32-bit system, I changed it to be skipped
whenever the expected result exceeded PHP_INT_MAX. There are 3 "toTimestamp"
functions within that test. Only one of these had been affected, but I thought
it was a good idea to add additional tests to the others to demonstrate this
condition.
In the course of testing, I also discovered some 32-bit problems with
bitwise and base-conversion functions. I am preparing separate PRs to
deal with those.
* Improving Coverage for Excel2003 XML Reader
Reader/Xml is now 100% covered.
File templates/Excel2003XMLTest.xml, used in some tests, is *not*
readable by a current version of Excel. I have substituted a new file
excel2003.xml to be used in its place. I have not deleted the original
in case someone in future (possibly me) wants to see what it needs to
make it usable.
There are minimal code changes.
- Unused protected functions pixel2WidthUnits and widthUnits2Pixel
are deleted.
- One regex looking to convert hex characters is changed from a-z to a-f,
and made case insensitive.
- No calculation performed for "error" cell (previously calculation
was attempted and threw exception).
- Empty relative row/cell is now handled correctly.
- Style applied to empty cell when appropriate.
- Support added for textRotation.
- Support added for border styles.
- Support added for diagonal borders.
- Support added for superscript and subscript.
- Support added for fill patterns.
In theory, encodings other than UTF-8 were supported.
In fact, I was unable to get SecurityScanner to pass *any* xml which is
not UTF-8. Eliminating the assumption that strings might not be UTF-8
allowed much of the code to be greatly simplified.
After that, I added some code that would permit the use of
some ASCII-compatible encodings (there is a test of ISO-8859-1).
It would be more difficult to handle other encodings (such as UTF-16).
I am not convinced that even the ISO-8859 effort is worth it,
but am willing to investigate either expanding or eliminating
non-UTF8 support.
I added a number of tests, creating an Xml directory, and moving
XmlTest to that directory.
Pull Request had problems reading old invalid sample in the code
coverage phase, not in any of the other test phases, and not in
the code coverage phase on my local machine.
As it turns out, aside from being invalid, the sample
is much larger than any of the other samples. Tests have been
adjusted accordingly.
* Smaller Test File
Should eliminate need to avoid test during xml coverage.
* Break Up Style Test into Multiple Tests
Per suggestion from Mark Baker.
* Integrate AddressHelper Change
The introduction of AddressHelper introduced a conflict which needed to
be resolved. I wanted to test it locally before resolving. This required
me to add (unchanged) AddressHelper to my local copy. I hope this is
an okay manner of resolving the conflict.
* Weird Travis Error
XmlOddTest works just fine on my local machine, but Travis failed it.
Even worse, the lines which Travis flags don't even make any sense
(one was the empty line between two methods!).
This test is not essential to the rest of the change. I am removing
it from the package, and will attempt to re-add it when I have a chance
to sync up my fork with the main project.
oleibman
Adrien Crivelli
oleibman
Mark Baker