git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Taylor Blau <me@ttaylorr.com>
To: Patrick Steinhardt <ps@pks.im>
Cc: Chandra Pratap via GitGitGadget <gitgitgadget@gmail.com>,
	git@vger.kernel.org, Chandra Pratap <chandrapratap376@gmail.com>,
	Chandra Pratap <chandrapratap3519@gmail.com>
Subject: Re: [PATCH v2] write-or-die: make GIT_FLUSH a Boolean environment variable
Date: Wed, 3 Jan 2024 12:15:26 -0500	[thread overview]
Message-ID: <ZZWWLkY+ixg+OMM4@nand.local> (raw)
In-Reply-To: <ZZUZNQqDTx3bnveJ@tanuki>

On Wed, Jan 03, 2024 at 09:22:13AM +0100, Patrick Steinhardt wrote:
> On Wed, Jan 03, 2024 at 07:58:28AM +0000, Chandra Pratap via GitGitGadget wrote:
> [snip]
> > diff --git a/write-or-die.c b/write-or-die.c
> > index 42a2dc73cd3..a6acabd329f 100644
> > --- a/write-or-die.c
> > +++ b/write-or-die.c
> > @@ -20,15 +20,12 @@ void maybe_flush_or_die(FILE *f, const char *desc)
> >  {
> >  	static int skip_stdout_flush = -1;
> >  	struct stat st;
> > -	char *cp;
> >
> >  	if (f == stdout) {
> >  		if (skip_stdout_flush < 0) {
> > -			/* NEEDSWORK: make this a normal Boolean */
> > -			cp = getenv("GIT_FLUSH");
> > -			if (cp)
> > -				skip_stdout_flush = (atoi(cp) == 0);
> > -			else if ((fstat(fileno(stdout), &st) == 0) &&
> > +			if (!git_env_bool("GIT_FLUSH", -1))
> > +				skip_stdout_flush = 1;
>
> It's a bit surprising to pass `-1` as default value to `git_env_bool()`
> here, as this value would hint that the caller wants to explicitly
> handle the case where the "GIT_FLUSH" envvar is not set at all. We don't
> though, and essentially fall back to "GIT_FLUSH=1", so passing `1` as
> the fallback value would be less confusing.
>
> Anyway, the resulting behaviour is the same regardless of whether we
> pass `1` or `-1`, so I'm not sure whether this is worth a reroll.

Hmm. If we pass -1 as the default value in the call to git_env_bool(),
the only time we'll end up in the else branch is if the environment is
set to some false-y value.

I don't think that matches the existing behavior, since right now we'll
infer skip_stdout_flush based on whether or not stdout is a regular file
or something else.

I think you'd probably want something closer to:

--- 8< ---
diff --git a/write-or-die.c b/write-or-die.c
index 42a2dc73cd..f12e111688 100644
--- a/write-or-die.c
+++ b/write-or-die.c
@@ -19,20 +19,17 @@
 void maybe_flush_or_die(FILE *f, const char *desc)
 {
 	static int skip_stdout_flush = -1;
-	struct stat st;
-	char *cp;

 	if (f == stdout) {
 		if (skip_stdout_flush < 0) {
-			/* NEEDSWORK: make this a normal Boolean */
-			cp = getenv("GIT_FLUSH");
-			if (cp)
-				skip_stdout_flush = (atoi(cp) == 0);
-			else if ((fstat(fileno(stdout), &st) == 0) &&
-				 S_ISREG(st.st_mode))
-				skip_stdout_flush = 1;
-			else
-				skip_stdout_flush = 0;
+			skip_stdout_flush = git_env_bool("GIT_FLUSH", -1);
+			if (skip_stdout_flush < 0) {
+				struct stat st;
+				if (fstat(fileno(f), &st))
+					skip_stdout_flush = 0;
+				else
+					skip_stdout_flush = S_ISREG(st.st_mode);
+			}
 		}
 		if (skip_stdout_flush && !ferror(f))
 			return;
--- >8 ---

You could lose one level of indentation, but it costs an extra fstat()
call in the case where GIT_FLUSH is set to some explicit value. Doing
that would look like this ugly thing instead ;-):

--- 8< ---
diff --git a/write-or-die.c b/write-or-die.c
index 42a2dc73cd..b3275d7577 100644
--- a/write-or-die.c
+++ b/write-or-die.c
@@ -19,20 +19,11 @@
 void maybe_flush_or_die(FILE *f, const char *desc)
 {
 	static int skip_stdout_flush = -1;
-	struct stat st;
-	char *cp;

 	if (f == stdout) {
 		if (skip_stdout_flush < 0) {
-			/* NEEDSWORK: make this a normal Boolean */
-			cp = getenv("GIT_FLUSH");
-			if (cp)
-				skip_stdout_flush = (atoi(cp) == 0);
-			else if ((fstat(fileno(stdout), &st) == 0) &&
-				 S_ISREG(st.st_mode))
-				skip_stdout_flush = 1;
-			else
-				skip_stdout_flush = 0;
+			struct stat st;
+			skip_stdout_flush = git_env_bool("GIT_FLUSH", !fstat(fileno(f), &st) && S_ISREG(st.st_mode));
 		}
 		if (skip_stdout_flush && !ferror(f))
 			return;
--- >8 ---

Thanks,
Taylor

  reply	other threads:[~2024-01-03 17:15 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-12-30 16:54 [PATCH] write-or-die: make GIT_FLUSH a Boolean environment variable Chandra Pratap via GitGitGadget
2024-01-01  8:24 ` Torsten Bögershausen
2024-01-02 16:29   ` Junio C Hamano
2024-01-03  7:58 ` [PATCH v2] " Chandra Pratap via GitGitGadget
2024-01-03  8:22   ` Patrick Steinhardt
2024-01-03 17:15     ` Taylor Blau [this message]
2024-01-03 18:42       ` Torsten Bögershausen
2024-01-03 19:18         ` Junio C Hamano
2024-01-04 17:35           ` Taylor Blau
2024-01-03 17:13   ` Junio C Hamano
2024-01-04 10:20   ` [PATCH v3] " Chandra Pratap via GitGitGadget
2024-01-04 17:36     ` Taylor Blau
2024-01-04 18:35     ` Junio C Hamano

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=ZZWWLkY+ixg+OMM4@nand.local \
    --to=me@ttaylorr.com \
    --cc=chandrapratap3519@gmail.com \
    --cc=chandrapratap376@gmail.com \
    --cc=git@vger.kernel.org \
    --cc=gitgitgadget@gmail.com \
    --cc=ps@pks.im \
    /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).