GitHub: fix GH-225
With Ruby 3.0.2 and csv 3.2.1, the file
```ruby
require "csv"
File.open("example.tsv", "w") { |f| f.puts("foo\t\tbar") }
CSV.read("example.tsv", col_sep: "\t", strip: true)
```
produces the error
```
lib/csv/parser.rb:935:in `parse_quotable_robust': TODO: Meaningful
message in line 1. (CSV::MalformedCSVError)
```
However, the CSV in this example is not malformed; instead, ambiguous
options were provided to the parser. It is not obvious (to me) whether
the string should be parsed as
- `["foo\t\tbar"]`,
- `["foo", "bar"]`,
- `["foo", "", "bar"]`, or
- `["foo", nil, "bar"]`.
This commit adds code that raises an exception when this situation is
encountered. Specifically, it checks if the column separator either ends
with or starts with the characters that would be stripped away.
This commit also adds unit tests and updates the documentation.
https://github.com/ruby/csv/commit/cc317dd42d
The implementation of the `CSV` shortcut method is broken in Ruby 3
for calls that look like this:
```ruby
CSV(write_stream, col_sep: "|", headers: headers, write_headers: true) do |csv|
...
end
```
The above will result in the following error when the `CSV` method attempts to pass
on arguments to `CSV#instance`:
```
ArgumentError: wrong number of arguments (given 2, expected 0..1)
```
The issue is due to the changes in Ruby 3 relating to positional & keyword arguments.
This commit updates the `CSV()` shortcut implementation to work with Ruby 3, and also
updates the documentation for the shortcut method.
https://github.com/ruby/csv/commit/310dee45fa
because the name "MJIT" is an internal code name, it's inconsistent with
--jit while they are related to each other, and I want to discourage future
JIT implementation-specific (e.g. MJIT-specific) APIs by this rename.
[Feature #17490]
This PR adds `undef: :replace` option for `CSV.open`.
`File.open` has `undef: :replace` option, but `CSV.open` does not.
It would be convenient if `CSV.open` could have a shortcut by having
`undef: :replace` option.
https://github.com/ruby/csv/commit/cff8b18480
* Do not loop forever when skip_lines regexp matches zero length with anchors
* Remove needless white spaces
* Add missing eos check in skip_needless_lines
* Simplify test
https://github.com/ruby/csv/commit/3b15d4a3e8
* Make CSV::Row#dup return a usable Row
Previously, calling `dup` on a `CSV::Row` object yielded an object whose
copy was too shallow. Changing the clone's fields would also change the
fields on the source. This change makes the clone more distinct from the
source, so that changes can be made to its fields without affecting the
source.
* Simplify
https://github.com/ruby/csv/commit/64a1ea06fc
I noticed that some files in rubygems were executable, and I could think
of no reason why they should be.
In general, I think ruby files should never have the executable bit set
unless they include a shebang, so I run the following command over the
whole repo:
```bash
find . -name '*.rb' -type f -executable -exec bash -c 'grep -L "^#!" $1 || chmod -x $1' _ {} \;
```
This includes performance improvements and backward incompatibility
fixes.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@67560 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This includes performance improvement especially writing. Writing is
about 2 times faster.
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@66507 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
Hiroshi SHIBATA
Takashi Kokubun
Sutou Kouhei
adamroyjones
Nobuyoshi Nakada
rm155
Anthony Hernandez
Joakim Antman
Burdette Lamar
Koichi ITO
Seiei Miyagi
Mike MacDonald
Jim Kane
David Rodríguez
Yusuke Endoh
Sutou Kouhei
Jeremy Evans
kou
nobu
hsbt