Releases: Touffy/client-zip
Releases · Touffy/client-zip
chmod
2.5.0:
- new:
modefield in input objects can override the default POSIX file mode.
2.4.6:
- bugfix: the
filenameattribute in a Content-Disposition header is now parsed correctly- when it is not the last attribute in the header
- and when it is the extended
filename*attribute
2.4.5:
- bugfix: the output stream can now be transferred without breaking the central repository.
Error propagation
2.4.4:
- options argument was missing in
makeZiptype declaration
2.4.3:
- Cancelling the output stream will really cause an error in the source iterator
Unicode Filename Flag
2.4.2:
- Filenames can be flagged as UTF-8 using the EFS bit in the file header flags.
- This is configurable with the new
buffersAreUTF8option.
- This is configurable with the new
- And it should help the Windows Zip utility read non-ASCII filenames correctly.
2.4.1:
- Cancelling the output stream will now cause an error in the source iterator.
Unicode Filename Flag (v1 branch)
1.6.2:
- Filenames can be flagged as UTF-8 using the EFS bit in the file header flags.
- This is configurable with the new
buffersAreUTF8option.
- This is configurable with the new
- And it should help the Windows Zip utility read non-ASCII filenames correctly.
1.6.1:
- Cancelling the output stream will now cause an error in the source iterator.
No more WebAssembly (for now…)
2.4.0
- No more WASM for CRC32, JavaScript is faster now anyway.
Backporting recent improvements from v2
1.6.0
- No more WASM for CRC32, JavaScript is faster now anyway.
- Added type and exports to package.json so client-zip can run in Node.js.
- Remove trailing slashes from file names.
- Fixed filename extraction from a Response when its URL has a trailing slash.
- Fixed content-length prediction when some files have a length of zero (previously, this threw an error).
Empty folders, filename heuristics
2.3.0
- Fixed filename extraction from a Response when its URL has a trailing slash.
- Empty folders may be added to the archive with
{ name, lastModified? }.- Remove trailing slashes from file names and add one to folder names if needed.
- Also export
makeZipwhich returns a ReadableStream without its Response wrapper.
Content-Length prediction
2.2.0 (with typing and empty file size bugs fixed in 2.2.2 and 2.2.1)
- Zip file size can now be predicted by the
predictLengthfunction before creating it. - Added an options parameter to
downloadZip.- options.length can be set to include a "Content-Length" header in the Response.
- options.metadata can instead be given the same argument as
predictLength, as a shortcut to compute and set the Content-Length.
Content-Length prediction (without Zip64)
1.5.0 (and a typing fix in 1.5.1)
- Zip file size can now be predicted by the
predictLengthfunction before creating it. - Added an options parameter to
downloadZip.- options.length can be set to include a "Content-Length" header in the Response.
- options.metadata can instead be given the same argument as
predictLength, as a shortcut to compute and set the Content-Length.
- The worker script now supports all the input types supported in the ES module, not only Responses.
Small files in big archives
2.0.1
- Fixed a bug where small files requiring a 64-bit offset (but not size) were corrupted.
- Never use 64-bit sizes in data descriptor for small files, even at large offsets.