Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

xar: fix Linux build on staging-next #352507

Merged
merged 1 commit into from
Nov 10, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 5 additions & 1 deletion pkgs/by-name/xa/xar/package.nix
Original file line number Diff line number Diff line change
Expand Up @@ -48,7 +48,11 @@ stdenv.mkDerivation (finalAttrs: {
# # …
# rm -r ../pkgs/by-name/xa/xar/patches
# git format-patch --zero-commit --output-directory ../pkgs/by-name/xa/xar/patches main
patches = lib.filesystem.listFilesRecursive ./patches;
patches =
# Avoid Darwin rebuilds on staging-next
lib.filter (
p: stdenv.hostPlatform.isDarwin -> baseNameOf p != "0020-Fall-back-to-readlink-on-Linux.patch"
) (lib.filesystem.listFilesRecursive ./patches);

# We do not use or modify files outside of the xar subdirectory.
patchFlags = [ "-p2" ];
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,48 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Randy Eckenrode <randy@largeandhighquality.com>
Date: Wed, 30 Oct 2024 20:19:20 -0400
Subject: [PATCH 20/20] Fall back to readlink on Linux

---
xar/lib/archive.c | 21 +++++++++++++++++++++
1 file changed, 21 insertions(+)

diff --git a/xar/lib/archive.c b/xar/lib/archive.c
index b7f9cbf..3a79c74 100644
--- a/xar/lib/archive.c
+++ b/xar/lib/archive.c
@@ -507,10 +507,31 @@ xar_t xar_fdopen_digest_verify(int fd, int32_t flags, void *expected_toc_digest,
// If there are hardlinks, the path we pick is the most recently opened by
// the filesystem; which is effectively random.
char path_buff[PATH_MAX];
+#if defined(__APPLE__)
if (fcntl(fd, F_GETPATH, path_buff) < 0) {
close(fd);
return NULL;
}
+#elif defined(__linux__)
+ // Linux has to get the path to the file via `/proc`.
+ char proc_buff[PATH_MAX];
+ if (snprintf(proc_buff, PATH_MAX, "/proc/self/fd/%i", fd) < 0) {
+ close(fd);
+ return NULL;
+ }
+ if (readlink(proc_buff, &path_buff, PATH_MAX) < 0) {
+ close(fd);
+ return NULL;
+ }
+ // The filename is the file’s when the fd was created. Check to make sure it still exists.
Copy link
Member

@tie tie Oct 31, 2024

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we really have to check that the file exists? How does F_GETPATH on Darwin handle this case (e.g. fcntl after unlink)?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure. The documentation just says it gets the path. I’m going to write some test programs later today to probe the behavior to see what Darwin does.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Darwin tracks name changes for renames. When the file is unlinked, F_GETPATH appears to return the last name it had. It may be possible to track renames in Linux (not sure), but it’s probably not worth the effort. Bailing if the underlying file was removed or renamed seems like the safest/easiest option.

+ struct stat stat_buff;
+ if (stat(path_buff, &stat_buff) < 0) {
+ close(fd);
+ return NULL;
+ }
+#else
+#error "Unknown platform. Please update with an implementation of `F_GETPATH`."
+#endif

// Don't trust the position of the descriptor we were given, reset it back to 0
if (lseek(fd, 0, SEEK_SET) != 0) {
--
2.47.0