zsh-workers
 help / color / mirror / code / Atom feed
From: Peter Stephenson <p.w.stephenson@ntlworld.com>
To: zsh-workers@zsh.org
Subject: Re: Any way to allow clobbering empty files when noclobber is set?
Date: Mon, 8 Jun 2020 10:30:32 +0100 (BST)	[thread overview]
Message-ID: <1300207061.4910.1591608632997@mail2.virginmedia.com> (raw)
In-Reply-To: <20200608032730.614e197c@tarpaulin.shahaf.local2>

Here's an updated patch, which I think incorporates the
current thinking, which seems pretty much to have iterated
to a conclusion.  The effective change is in fact now
trivial --- an "if" with a return.

Webmail again, as it seemed pretty non-invasive last time...

pws

diff --git a/Doc/Zsh/options.yo b/Doc/Zsh/options.yo
index 2b7637ff4..6da68308f 100644
--- a/Doc/Zsh/options.yo
+++ b/Doc/Zsh/options.yo
@@ -1168,6 +1168,22 @@ If the option is not set, and the option tt(APPEND_CREATE) is also
 not set, `tt(>>!)' or `tt(>>|)' must be used to create a file.
 If either option is set, `tt(>>)' may be used.
 )
+pindex(CLOBBER_EMPTY)
+pindex(NO_CLOBBER_EMPTY)
+pindex(CLOBBEREMPTY)
+pindex(NOCLOBBEREMPTY)
+cindex(clobbering, of empty files)
+cindex(file clobbering, of empty files)
+item(tt(CLOBBER_EMPTY))(
+This option is only used if the option tt(CLOBBER) is not set: note that
+it is set by default.
+
+If this option is set, then regular files of zero length may be
+ovewritten (`clobbered').  Note that it is possible another process
+has written to the file between this test and use of the file by
+the current process.  This option should therefore not be used in
+cases where files to be clobbered may be written to asynchronously.
+)
 pindex(CORRECT)
 pindex(NO_CORRECT)
 pindex(NOCORRECT)
diff --git a/Src/exec.c b/Src/exec.c
index 29f4fc5ca..7b087f3b0 100644
--- a/Src/exec.c
+++ b/Src/exec.c
@@ -2143,14 +2143,15 @@ clobber_open(struct redir *f)
 {
     struct stat buf;
     int fd, oerrno;
+    char *ufname = unmeta(f->name);
 
     /* If clobbering, just open. */
     if (isset(CLOBBER) || IS_CLOBBER_REDIR(f->type))
-	return open(unmeta(f->name),
+	return open(ufname,
 		O_WRONLY | O_CREAT | O_TRUNC | O_NOCTTY, 0666);
 
     /* If not clobbering, attempt to create file exclusively. */
-    if ((fd = open(unmeta(f->name),
+    if ((fd = open(ufname,
 		   O_WRONLY | O_CREAT | O_EXCL | O_NOCTTY, 0666)) >= 0)
 	return fd;
 
@@ -2158,11 +2159,27 @@ clobber_open(struct redir *f)
      * Try opening, and if it's a regular file then close it again    *
      * because we weren't supposed to open it.                        */
     oerrno = errno;
-    if ((fd = open(unmeta(f->name), O_WRONLY | O_NOCTTY)) != -1) {
-	if(!fstat(fd, &buf) && !S_ISREG(buf.st_mode))
-	    return fd;
+    if ((fd = open(ufname, O_WRONLY | O_NOCTTY)) != -1) {
+	if(!fstat(fd, &buf)) {
+	    if (!S_ISREG(buf.st_mode))
+		return fd;
+	    /*
+	     * If CLOBBER_EMPTY is in effect and the file is empty,
+	     * we are allowed to re-use it.
+	     *
+	     * Note: there is an intrinsic race here because another
+	     * process can write to this file at any time.  The only fix
+	     * would be file locking, which we wish to avoid in basic
+	     * file operations at this level.  This would not be
+	     * fixed. just additionally complicated, by re-opening the
+	     * file and truncating.
+	     */
+	    if (isset(CLOBBEREMPTY) && buf.st_size == 0)
+		return fd;
+	}
 	close(fd);
     }
+
     errno = oerrno;
     return -1;
 }
diff --git a/Src/options.c b/Src/options.c
index 7586d21d2..fba021e7d 100644
--- a/Src/options.c
+++ b/Src/options.c
@@ -114,6 +114,7 @@ static struct optname optns[] = {
 {{NULL, "checkjobs",	      OPT_EMULATE|OPT_ZSH},	 CHECKJOBS},
 {{NULL, "checkrunningjobs",   OPT_EMULATE|OPT_ZSH},	 CHECKRUNNINGJOBS},
 {{NULL, "clobber",	      OPT_EMULATE|OPT_ALL},	 CLOBBER},
+{{NULL, "clobberempty",	      0},			 CLOBBEREMPTY},
 {{NULL, "combiningchars",     0},			 COMBININGCHARS},
 {{NULL, "completealiases",    0},			 COMPLETEALIASES},
 {{NULL, "completeinword",     0},			 COMPLETEINWORD},
diff --git a/Src/zsh.h b/Src/zsh.h
index 1f2d774a1..ed123f2b9 100644
--- a/Src/zsh.h
+++ b/Src/zsh.h
@@ -2378,6 +2378,7 @@ enum {
     CHECKJOBS,
     CHECKRUNNINGJOBS,
     CLOBBER,
+    CLOBBEREMPTY,
     APPENDCREATE,
     COMBININGCHARS,
     COMPLETEALIASES,
diff --git a/Test/A04redirect.ztst b/Test/A04redirect.ztst
index d60519064..993138e7d 100644
--- a/Test/A04redirect.ztst
+++ b/Test/A04redirect.ztst
@@ -708,3 +708,17 @@
   cat <&$testfd
 0:Regression test for here document with fd declarator
 >  This is, in some sense, a here document.
+
+  (setopt noclobber clobberempty
+  rm -f foo
+  touch foo
+  print Works >foo
+  cat foo
+  print Works not >foo
+  # Make sure the file was not harmed
+  cat foo
+  )
+0:CLOBBER_EMPTY
+>Works
+>Works
+?(eval):6: file exists: foo

  reply	other threads:[~2020-06-08  9:31 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20200603020919eucas1p13e26ebcbb335784d14bfb97b137f385a@eucas1p1.samsung.com>
2020-06-03  2:08 ` Martin Tournoij
2020-06-03 12:04   ` Peter Stephenson
2020-06-04  1:48     ` Daniel Shahaf
2020-06-04  2:43       ` Bart Schaefer
2020-06-04  4:06         ` Daniel Shahaf
2020-06-04  5:00           ` Bart Schaefer
2020-06-05  3:10             ` Daniel Shahaf
2020-06-05  3:18               ` Daniel Shahaf
2020-06-06  1:07               ` Bart Schaefer
2020-06-06  4:48                 ` Daniel Shahaf
2020-06-06  7:04                   ` Bart Schaefer
2020-06-04  6:31       ` Martin Tournoij
2020-06-05  2:22         ` Daniel Shahaf
2020-06-04  2:13     ` Vin Shelton
2020-06-04  2:35       ` Bart Schaefer
2020-06-04  2:36       ` Daniel Shahaf
2020-06-04 11:57         ` Vin Shelton
2020-06-04  5:06     ` Bart Schaefer
2020-06-04  5:41       ` Roman Perepelitsa
2020-06-05  2:07         ` Daniel Shahaf
2020-06-05  4:38           ` Roman Perepelitsa
2020-06-06  1:41             ` Bart Schaefer
2020-06-06  4:55               ` Daniel Shahaf
2020-06-06  6:25                 ` Roman Perepelitsa
2020-06-06  7:08                 ` Bart Schaefer
2020-06-06  8:03                   ` Daniel Shahaf
     [not found]           ` <1941572212.466119.1591360860372@mail2.virginmedia.com>
     [not found]             ` <e7f7dfe2-eb4a-457b-85fb-091935a74c0e@www.fastmail.com>
2020-06-06 11:57               ` Peter Stephenson
2020-06-06 12:48                 ` Roman Perepelitsa
2020-06-06 15:24                   ` Bart Schaefer
2020-06-06 16:24                     ` Peter Stephenson
2020-06-07 11:55                       ` Daniel Shahaf
2020-06-07 17:00                         ` Peter Stephenson
2020-06-08  3:27                           ` Daniel Shahaf
2020-06-08  9:30                             ` Peter Stephenson [this message]
2020-06-07 17:20                         ` Bart Schaefer
2020-06-06 15:09                 ` Bart Schaefer
2020-06-04  6:47       ` Martin Tournoij
2020-06-04  9:42       ` Peter Stephenson
2020-06-04 12:20         ` Roman Perepelitsa
2020-06-04 12:26           ` Peter Stephenson
2020-06-04 12:15     ` Peter Stephenson
2020-06-04 20:35       ` Bart Schaefer
2020-06-05  1:59         ` Daniel Shahaf

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=1300207061.4910.1591608632997@mail2.virginmedia.com \
    --to=p.w.stephenson@ntlworld.com \
    --cc=zsh-workers@zsh.org \
    /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.
Code repositories for project(s) associated with this public inbox

	https://git.vuxu.org/mirror/zsh/

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).