From: "René Scharfe" <rene.scharfe@lsrfire.ath.cx>
To: Johannes Sixt <j6t@kdbg.org>
Cc: Jeff King <peff@peff.net>, Erik Faye-Lund <kusmabite@gmail.com>,
Junio C Hamano <gitster@pobox.com>,
git@vger.kernel.org
Subject: Re: [PATCH v2 4/4] upload-archive: use start_command instead of fork
Date: Sat, 06 Aug 2011 11:40:45 +0200 [thread overview]
Message-ID: <4E3D0C1D.9000807@lsrfire.ath.cx> (raw)
In-Reply-To: <4E388A55.6080606@kdbg.org>
Am 03.08.2011 01:37, schrieb Johannes Sixt:
> Am 02.08.2011 20:13, schrieb Jeff King:
>> Hmm. So it's not _just_ the pipe vs file thing. What's different about
>> calling it from the shell, versus the way we call it from git-archive?
>
> When the parent process of an MSYS process is itself an MSYS process,
> such as bash, then the child does not do its own
> binary-mode-vs.-text-mode detection, but just uses whatever it is told
> by the parent. This is achieved by MSYS's fork emulation.
>
> But if the parent is a regular Windows program, such as git(-archive),
> then the autodection happens and file descriptors pointing to files are
> put into text mode.
So here's an ugly patch to implement an internal passthrough filter to
avoid newline conversions. It makes the tar filter command (gzip etc.)
write to a pipe instead of directly to a file.
The patch does this unconditionally, which is a waste on all unaffected
systems, of course. We could #ifdef it out, but is there perhaps a nice
way to integrate this functionality into run_command/finish_command?
René
---
archive-tar.c | 31 +++++++++++++++++++++++++++++++
1 files changed, 31 insertions(+), 0 deletions(-)
diff --git a/archive-tar.c b/archive-tar.c
index 20af005..22d52c5 100644
--- a/archive-tar.c
+++ b/archive-tar.c
@@ -314,6 +314,25 @@ static int write_tar_archive(const struct archiver *ar,
return err;
}
+static int cat_fn(int in, int out, void *data)
+{
+ for (;;) {
+ char buf[16 * 1024];
+ ssize_t got = xread(in, buf, sizeof(buf));
+ if (got == 0)
+ break;
+ if (got < 0) {
+ if (errno == EPIPE)
+ break;
+ die_errno("read errror");
+ }
+ write_or_die(out, buf, got);
+ }
+ close(in);
+ close(out);
+ return 0;
+}
+
static int write_tar_filter_archive(const struct archiver *ar,
struct archiver_args *args)
{
@@ -321,6 +340,14 @@ static int write_tar_filter_archive(const struct archiver *ar,
struct child_process filter;
const char *argv[2];
int r;
+ struct async cat;
+
+ memset(&cat, 0, sizeof(cat));
+ cat.proc = cat_fn;
+ cat.in = -1;
+ cat.out = dup(1);
+ if (start_async(&cat))
+ die("unable to start passthrough filter");
if (!ar->data)
die("BUG: tar-filter archiver called with no filter defined");
@@ -335,6 +362,7 @@ static int write_tar_filter_archive(const struct archiver *ar,
filter.argv = argv;
filter.use_shell = 1;
filter.in = -1;
+ filter.out = cat.in;
if (start_command(&filter) < 0)
die_errno("unable to start '%s' filter", argv[0]);
@@ -349,6 +377,9 @@ static int write_tar_filter_archive(const struct archiver *ar,
if (finish_command(&filter) != 0)
die("'%s' filter reported error", argv[0]);
+ if (finish_async(&cat))
+ die("passthrough filter reported error");
+
strbuf_release(&cmd);
return r;
}
next prev parent reply other threads:[~2011-08-06 9:41 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-07-18 18:08 [PATCH v2 0/4] port upload-archive to Windows Erik Faye-Lund
2011-07-18 18:08 ` [PATCH v2 1/4] compat/win32/sys/poll.c: upgrade from upstream Erik Faye-Lund
2011-07-18 18:08 ` [PATCH v2 2/4] mingw: fix compilation of poll-emulation Erik Faye-Lund
2011-07-18 18:08 ` [PATCH v2 3/4] enter_repo: do not modify input Erik Faye-Lund
2011-07-18 18:08 ` [PATCH v2 4/4] upload-archive: use start_command instead of fork Erik Faye-Lund
2011-07-19 21:09 ` Junio C Hamano
2011-07-28 8:32 ` Erik Faye-Lund
2011-07-28 16:08 ` Jeff King
2011-07-28 16:47 ` Jeff King
2011-07-28 17:02 ` Jeff King
2011-08-01 14:45 ` Erik Faye-Lund
2011-08-01 17:46 ` Jeff King
2011-08-01 18:02 ` Erik Faye-Lund
2011-08-01 18:25 ` Jeff King
2011-08-01 20:48 ` René Scharfe
2011-08-01 21:20 ` Johannes Sixt
2011-08-01 21:42 ` René Scharfe
2011-08-01 21:52 ` René Scharfe
2011-08-02 4:00 ` Jeff King
2011-08-02 16:46 ` René Scharfe
2011-08-02 18:13 ` Jeff King
2011-08-02 23:37 ` Johannes Sixt
2011-08-03 5:49 ` Jeff King
2011-08-06 9:40 ` René Scharfe [this message]
2011-08-07 20:02 ` Johannes Sixt
2011-08-07 21:06 ` Jeff King
2011-08-08 17:10 ` René Scharfe
2011-08-09 5:02 ` Jeff King
2011-08-09 10:25 ` René Scharfe
2011-08-09 20:05 ` Jeff King
2011-09-29 19:54 ` Erik Faye-Lund
2011-09-29 20:18 ` René Scharfe
2011-09-29 20:20 ` Erik Faye-Lund
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4E3D0C1D.9000807@lsrfire.ath.cx \
--to=rene.scharfe@lsrfire.ath.cx \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=j6t@kdbg.org \
--cc=kusmabite@gmail.com \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).