From 5dcad1f707bfd949bcddc4ac9f3b774184bd36ff Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 8 May 2012 08:33:16 +0000 Subject: [PATCH] Re: [PATCH v2 2/2] new: Centralize file type stat-ing logic --- 84/585bad83fdf29e047afd4615714cf6a8ab556a | 279 ++++++++++++++++++++++ 1 file changed, 279 insertions(+) create mode 100644 84/585bad83fdf29e047afd4615714cf6a8ab556a diff --git a/84/585bad83fdf29e047afd4615714cf6a8ab556a b/84/585bad83fdf29e047afd4615714cf6a8ab556a new file mode 100644 index 000000000..b018d14cb --- /dev/null +++ b/84/585bad83fdf29e047afd4615714cf6a8ab556a @@ -0,0 +1,279 @@ +Return-Path: +X-Original-To: notmuch@notmuchmail.org +Delivered-To: notmuch@notmuchmail.org +Received: from localhost (localhost [127.0.0.1]) + by olra.theworths.org (Postfix) with ESMTP id 635A1431FAF + for ; Tue, 8 May 2012 01:33:23 -0700 (PDT) +X-Virus-Scanned: Debian amavisd-new at olra.theworths.org +X-Spam-Flag: NO +X-Spam-Score: -0.7 +X-Spam-Level: +X-Spam-Status: No, score=-0.7 tagged_above=-999 required=5 + tests=[RCVD_IN_DNSWL_LOW=-0.7] autolearn=disabled +Received: from olra.theworths.org ([127.0.0.1]) + by localhost (olra.theworths.org [127.0.0.1]) (amavisd-new, port 10024) + with ESMTP id cVoVnBSJkCCE for ; + Tue, 8 May 2012 01:33:22 -0700 (PDT) +Received: from mail-qc0-f181.google.com (mail-qc0-f181.google.com + [209.85.216.181]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) + (No client certificate requested) + by olra.theworths.org (Postfix) with ESMTPS id 5AFAA431FAE + for ; Tue, 8 May 2012 01:33:22 -0700 (PDT) +Received: by qcsk26 with SMTP id k26so167726qcs.26 + for ; Tue, 08 May 2012 01:33:21 -0700 (PDT) +X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; + d=google.com; s=20120113; + h=from:to:cc:subject:in-reply-to:references:user-agent:date + :message-id:mime-version:content-type:x-gm-message-state; + bh=rSjJAz3w8woYPeiPZ1rldRZ3luikkCzldglqMO0OrUo=; + b=Z4WfWnzhKHBjy87pLghqOYk+P79MrvcwRYakFmEnwSQ1iaeEobXLWe7nriXrrYcugn + +knmWo9C8roiH4kBuvkBk7M1KygftPQKE2lY5QVU3EELm69c2yuBjXRlZnVuu6wqFuma + Pu6iew3Ms0VfqR+OOaZRAdgBZXYiSDvl/ft+pkYiF7GtTY2LQ+qgvwt1/UPnlQTKMY71 + dLOwEMf9u0XXx5NTwQpZMiJ2o1LQighRjeuBtPRE1pMNkN4qoO6Kx9rrP/sg5noXe1aq + Aek1c6SpffBN2TNSXKECly+YafSzZ9OqaRuprV7Iat0S91lbftIHw4k2EKWTkAj2xM1+ + w/Sg== +Received: by 10.224.194.1 with SMTP id dw1mr14002459qab.87.1336466001685; + Tue, 08 May 2012 01:33:21 -0700 (PDT) +Received: from localhost (nikula.org. [92.243.24.172]) + by mx.google.com with ESMTPS id hv18sm2806450qab.19.2012.05.08.01.33.18 + (version=SSLv3 cipher=OTHER); Tue, 08 May 2012 01:33:20 -0700 (PDT) +From: Jani Nikula +To: Austin Clements , notmuch@notmuchmail.org +Subject: Re: [PATCH v2 2/2] new: Centralize file type stat-ing logic +In-Reply-To: <87r4uvdryz.fsf@nikula.org> +References: <1336414186-15293-1-git-send-email-amdragon@mit.edu> + <1336429240-1114-1-git-send-email-amdragon@mit.edu> + <1336429240-1114-3-git-send-email-amdragon@mit.edu> + <87r4uvdryz.fsf@nikula.org> +User-Agent: Notmuch/0.11.1+222~ga47a98c (http://notmuchmail.org) Emacs/23.1.1 + (i686-pc-linux-gnu) +Date: Tue, 08 May 2012 08:33:16 +0000 +Message-ID: <87obpzdqcz.fsf@nikula.org> +MIME-Version: 1.0 +Content-Type: text/plain; charset=us-ascii +X-Gm-Message-State: + ALoCoQk2wZuM1mSaEBs4An72CY6ht4d4zqMAYpZDDf4mcqmw3Uncze4/YaXBItm9HgE9Qc8X4cXs +Cc: Vladimir Marek +X-BeenThere: notmuch@notmuchmail.org +X-Mailman-Version: 2.1.13 +Precedence: list +List-Id: "Use and development of the notmuch mail system." + +List-Unsubscribe: , + +List-Archive: +List-Post: +List-Help: +List-Subscribe: , + +X-List-Received-Date: Tue, 08 May 2012 08:33:23 -0000 + +On Tue, 08 May 2012 07:58:28 +0000, Jani Nikula wrote: +> On Mon, 7 May 2012 18:20:40 -0400, Austin Clements wrote: +> > This moves our logic to get a file's type into one function. This has +> > several benefits: we can support OSes and file systems that do not +> > provide dirent.d_type or always return DT_UNKNOWN, complex +> > symlink-handling logic has been replaced by a simple stat fall-through +> > in one place, and the error message for un-stat-able file is more +> > accurate (previously, the error always mentioned directories, even +> > though a broken symlink is not a directory). +> +> LGTM. + +Okay, it's good, but I think you can make it even better: + +add_files_recursive() has check for "! S_ISDIR (st.st_mode)" in the +beginning, returning silently in case it recursed based on a symlink to +regular file. IIUC, this will no longer happen with your patch, as +symlinks are resolved and stat'ed before recursing. + +add_files() exists to fail loudly in the same situation, and has +otherwise the same checks in the beginning. I think you could now use +the checks from add_files() to replace the ones in +add_files_recursive(), and get rid of add_files() altogether. + +Please double check my thinking. Also this should probably be a separate +patch, no need to change the current one. + +Jani. + + +> +> > --- +> > notmuch-new.c | 103 +++++++++++++++++++++++++++++++++++---------------------- +> > test/new | 2 +- +> > 2 files changed, 64 insertions(+), 41 deletions(-) +> > +> > diff --git a/notmuch-new.c b/notmuch-new.c +> > index cb720cc..8955677 100644 +> > --- a/notmuch-new.c +> > +++ b/notmuch-new.c +> > @@ -154,6 +154,48 @@ dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b) +> > return strcmp ((*a)->d_name, (*b)->d_name); +> > } +> > +> > +/* Return the type of a directory entry relative to path as a stat(2) +> > + * mode. Like stat, this follows symlinks. Returns -1 and sets errno +> > + * if the file's type cannot be determined (which includes dangling +> > + * symlinks). +> > + */ +> > +static int +> > +dirent_type (const char *path, const struct dirent *entry) +> > +{ +> > + struct stat statbuf; +> > + char *abspath; +> > + int err, saved_errno; +> > + +> > +#ifdef _DIRENT_HAVE_D_TYPE +> > + /* Mapping from d_type to stat mode_t. We omit DT_LNK so that +> > + * we'll fall through to stat and get the real file type. */ +> > + static const mode_t modes[] = { +> > + [DT_BLK] = S_IFBLK, +> > + [DT_CHR] = S_IFCHR, +> > + [DT_DIR] = S_IFDIR, +> > + [DT_FIFO] = S_IFIFO, +> > + [DT_REG] = S_IFREG, +> > + [DT_SOCK] = S_IFSOCK +> > + }; +> > + if (entry->d_type < ARRAY_SIZE(modes) && modes[entry->d_type]) +> > + return modes[entry->d_type]; +> > +#endif +> > + +> > + abspath = talloc_asprintf (NULL, "%s/%s", path, entry->d_name); +> > + if (!abspath) { +> > + errno = ENOMEM; +> > + return -1; +> > + } +> > + err = stat(abspath, &statbuf); +> > + saved_errno = errno; +> > + talloc_free (abspath); +> > + if (err < 0) { +> > + errno = saved_errno; +> > + return -1; +> > + } +> > + return statbuf.st_mode & S_IFMT; +> > +} +> > + +> > /* Test if the directory looks like a Maildir directory. +> > * +> > * Search through the array of directory entries to see if we can find all +> > @@ -162,12 +204,12 @@ dirent_sort_strcmp_name (const struct dirent **a, const struct dirent **b) +> > * Return 1 if the directory looks like a Maildir and 0 otherwise. +> > */ +> > static int +> > -_entries_resemble_maildir (struct dirent **entries, int count) +> > +_entries_resemble_maildir (const char *path, struct dirent **entries, int count) +> > { +> > int i, found = 0; +> > +> > for (i = 0; i < count; i++) { +> > - if (entries[i]->d_type != DT_DIR && entries[i]->d_type != DT_UNKNOWN) +> > + if (dirent_type (path, entries[i]) != S_IFDIR) +> > continue; +> > +> > if (strcmp(entries[i]->d_name, "new") == 0 || +> > @@ -250,7 +292,7 @@ add_files_recursive (notmuch_database_t *notmuch, +> > notmuch_status_t status, ret = NOTMUCH_STATUS_SUCCESS; +> > notmuch_message_t *message = NULL; +> > struct dirent **fs_entries = NULL; +> > - int i, num_fs_entries; +> > + int i, num_fs_entries, entry_type; +> > notmuch_directory_t *directory; +> > notmuch_filenames_t *db_files = NULL; +> > notmuch_filenames_t *db_subdirs = NULL; +> > @@ -317,7 +359,7 @@ add_files_recursive (notmuch_database_t *notmuch, +> > } +> > +> > /* Pass 1: Recurse into all sub-directories. */ +> > - is_maildir = _entries_resemble_maildir (fs_entries, num_fs_entries); +> > + is_maildir = _entries_resemble_maildir (path, fs_entries, num_fs_entries); +> > +> > for (i = 0; i < num_fs_entries; i++) { +> > if (interrupted) +> > @@ -325,17 +367,16 @@ add_files_recursive (notmuch_database_t *notmuch, +> > +> > entry = fs_entries[i]; +> > +> > - /* We only want to descend into directories. +> > - * But symlinks can be to directories too, of course. +> > - * +> > - * And if the filesystem doesn't tell us the file type in the +> > - * scandir results, then it might be a directory (and if not, +> > - * then we'll stat and return immediately in the next level of +> > - * recursion). */ +> > - if (entry->d_type != DT_DIR && +> > - entry->d_type != DT_LNK && +> > - entry->d_type != DT_UNKNOWN) +> > - { +> > + /* We only want to descend into directories (and symlinks to +> > + * directories). */ +> > + entry_type = dirent_type (path, entry); +> > + if (entry_type == -1) { +> > + /* Be pessimistic, e.g. so we don't lose lots of mail just +> > + * because a user broke a symlink. */ +> > + fprintf (stderr, "Error reading file %s/%s: %s\n", +> > + path, entry->d_name, strerror (errno)); +> > + return NOTMUCH_STATUS_FILE_ERROR; +> > + } else if (entry_type != S_IFDIR) { +> > continue; +> > } +> > +> > @@ -425,31 +466,13 @@ add_files_recursive (notmuch_database_t *notmuch, +> > notmuch_filenames_move_to_next (db_subdirs); +> > } +> > +> > - /* If we're looking at a symlink, we only want to add it if it +> > - * links to a regular file, (and not to a directory, say). +> > - * +> > - * Similarly, if the file is of unknown type (due to filesystem +> > - * limitations), then we also need to look closer. +> > - * +> > - * In either case, a stat does the trick. +> > - */ +> > - if (entry->d_type == DT_LNK || entry->d_type == DT_UNKNOWN) { +> > - int err; +> > - +> > - next = talloc_asprintf (notmuch, "%s/%s", path, entry->d_name); +> > - err = stat (next, &st); +> > - talloc_free (next); +> > - next = NULL; +> > - +> > - /* Don't emit an error for a link pointing nowhere, since +> > - * the directory-traversal pass will have already done +> > - * that. */ +> > - if (err) +> > - continue; +> > - +> > - if (! S_ISREG (st.st_mode)) +> > - continue; +> > - } else if (entry->d_type != DT_REG) { +> > + /* Only add regular files (and symlinks to regular files). */ +> > + entry_type = dirent_type (path, entry); +> > + if (entry_type == -1) { +> > + fprintf (stderr, "Error reading file %s/%s: %s\n", +> > + path, entry->d_name, strerror (errno)); +> > + return NOTMUCH_STATUS_FILE_ERROR; +> > + } else if (entry_type != S_IFREG) { +> > continue; +> > } +> > +> > diff --git a/test/new b/test/new +> > index 26253db..e3900f5 100755 +> > --- a/test/new +> > +++ b/test/new +> > @@ -140,7 +140,7 @@ test_begin_subtest "Broken symlink aborts" +> > ln -s does-not-exist "${MAIL_DIR}/broken" +> > output=$(NOTMUCH_NEW 2>&1) +> > test_expect_equal "$output" \ +> > -"Error reading directory /run/shm/nm/tmp.new/mail/broken: No such file or directory +> > +"Error reading file /run/shm/nm/tmp.new/mail/broken: No such file or directory +> > Note: A fatal error was encountered: Something went wrong trying to read or write a file +> > No new mail." +> > rm "${MAIL_DIR}/broken" +> > -- +> > 1.7.10 +> > +> > _______________________________________________ +> > notmuch mailing list +> > notmuch@notmuchmail.org +> > http://notmuchmail.org/mailman/listinfo/notmuch -- 2.26.2