Bazel fails to unzip archives containing files with non-latin characters in their name

[rickwebiii]

Hello,
I'm porting our build over to Bazel and am running into a problem that only reproduces on mac. Namely, I created an http_archive repo rule points to a zip file that contains an é in a filename. This results in the following error:

/Users/***/maestro/platform-monolith/BUILD.bazel:3:10: //platform-monolith:artifacts depends on @monolith_osx//:artifacts in repository @monolith_osx which failed to fetch. no such package '@monolith_osx//': java.io.IOException: Error extracting /private/var/tmp/_bazel_rweber/47c212f6310157f06bb1214862485fca/external/monolith_osx/desktop_linux.zip to /private/var/tmp/_bazel_rweber/47c212f6310157f06bb1214862485fca/external/monolith_osx: /private/var/tmp/_bazel_rweber/47c212f6310157f06bb1214862485fca/external/monolith_osx/tableau-1.3/install/defaults/Datasources/fr_FR-EU/Exemple - Hypermarch?.tds (Illegal byte sequence)

The offending filename contains an e with an accent-aigu.

I did a little more investigation and found I can repro this problem by simply zipping an archive containing the one offending file and writing a repository_rule that simple tries to call extract on that file:

rules.bzl:

def _http_archive_workaround(ctx):
  ctx.extract(
    archive = ctx.attr.file,
    output = "./",
  )
 

http_archive_workaround = repository_rule(
  implementation = _http_archive_workaround,
  attrs = {
    "file": attr.string(),
  }
)

WORKSPACE:

load("//:bazel/rules/mac/rules.bzl", "http_archive_workaround")

http_archive_workaround(
  name = "monolith_osx",
  file = "/Users/***/maestro/repro.zip",
)

I'm using bazel 3.3.0 and this doesn't repro on Windows or Linux, only mac.

As a workaround, I created a repoitory rule that downloads the zip file and calls the build-in unzip program to extract, then only specify filegroups that don't contain the offending files (I don't actually need them, but they're stopping the archive extraction).

[aiuto]

Do you have a small github repo which contains a self contained repro?

There could be any of several things going on. The mac difference might have to do with Macos's strange choice of how they encode file names. (Nice summary here: https://superuser.com/questions/999232/unicode-filenames-in-windows-vs-mac-os-x).

[y3llowcake]

Not a particularly small repro, but we just ran into this issue extracting a recent golang release:

If the definition of 'golang_sdk_osx-x86_64' was updated, verify that the hashes were also updated.
ERROR: An error occurred during the fetch of repository 'golang_sdk_osx-x86_64':
  Traceback (most recent call last):
	File "/private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/bazel_tools/tools/build_defs/repo/http.bzl", line 111, column 45, in _http_archive_impl
		download_info = ctx.download_and_extract(
Error in download_and_extract: java.io.IOException: Error extracting /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796/go1.16.12.darwin-amd64.tar.gz to /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796: /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/test/fixedbugs/issue27836.dir/?foo.go (Illegal byte sequence)
ERROR: Error fetching repository: Traceback (most recent call last):
	File "/private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/bazel_tools/tools/build_defs/repo/http.bzl", line 111, column 45, in _http_archive_impl
		download_info = ctx.download_and_extract(
Error in download_and_extract: java.io.IOException: Error extracting /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796/go1.16.12.darwin-amd64.tar.gz to /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796: /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/test/fixedbugs/issue27836.dir/?foo.go (Illegal byte sequence)
ERROR: /Users/snagaraj/dev/goslackgo/BUILD.bazel:8:7: //:go depends on @golang_sdk_osx-x86_64//:toolchain_info in repository @golang_sdk_osx-x86_64 which failed to fetch. no such package '@golang_sdk_osx-x86_64//': java.io.IOException: Error extracting /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796/go1.16.12.darwin-amd64.tar.gz to /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/temp17079543164555012796: /private/var/tmp/_bazel_snagaraj/bf4a8035aa09cf910ceaead73516862a/external/golang_sdk_osx-x86_64/test/fixedbugs/issue27836.dir/?foo.go (Illegal byte sequence)```

[y3llowcake]

Seems like this may be dupe of #12986 ?

[aiuto]

Indeed it is a dup. Closing