|
From: Yura P. <cha...@us...> - 2005-06-21 18:54:26
|
Changes by: cha0smaster Update of /cvsroot/linux-ntfs/ntfsprogs/ntfsprogs In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv30226/ntfsprogs Modified Files: ntfscp.c Log Message: ntfscp: Fix signal handling. Index: ntfscp.c =================================================================== RCS file: /cvsroot/linux-ntfs/ntfsprogs/ntfsprogs/ntfscp.c,v retrieving revision 1.21 retrieving revision 1.22 diff -u -p -r1.21 -r1.22 --- ntfscp.c 20 Jun 2005 14:31:59 -0000 1.21 +++ ntfscp.c 21 Jun 2005 13:06:53 -0000 1.22 @@ -54,7 +54,7 @@ struct options { static const char *EXEC_NAME = "ntfscp"; static struct options opts; -static int caught_sigint = 0; +static int caught_terminate = 0; GEN_PRINTF (Eprintf, stderr, NULL, FALSE) GEN_PRINTF (Vprintf, stderr, &opts.verbose, TRUE) @@ -235,17 +235,18 @@ static int parse_options (int argc, char } /** - * sigint_handler - Handle SIGINT: abort write, sync and exit. + * signal_handler - Handle SIGINT and SIGTERM: abort write, sync and exit. */ -static void sigint_handler(int arg __attribute__((unused))) +static void signal_handler(int arg) { - caught_sigint++; - if (caught_sigint > 3) { - Eprintf("SIGTERM received more than 3 times. " + caught_terminate++; + if (caught_terminate > 3) { + Eprintf("SIGTERM or SIGINT received more than 3 times. " "Exit immediately.\n"); exit(2); } else - Eprintf("SIGTERM received. Aborting write.\n"); + Eprintf("%s received. Aborting write.\n", + (arg == SIGINT) ? "SIGINT" : "SIGTERM"); } /** @@ -277,10 +278,15 @@ int main (int argc, char *argv[]) utils_set_locale(); /* Set SIGINT handler. */ - if (signal(SIGINT, sigint_handler) == SIG_ERR) { + if (signal(SIGINT, signal_handler) == SIG_ERR) { perror("Failed to set SIGINT handler"); return 1; } + /* Set SIGTERM handler. */ + if (signal(SIGTERM, signal_handler) == SIG_ERR) { + perror("Failed to set SIGTERM handler"); + return 1; + } if (opts.noaction) flags = MS_RDONLY; @@ -401,7 +407,7 @@ int main (int argc, char *argv[]) Vprintf("Starting write.\n"); offset = 0; - while (!feof(in) && !caught_sigint) { + while (!feof(in) && !caught_terminate) { br = fread(buf, 1, NTFS_BUF_SIZE, in); if (!br) { if (!feof(in)) perror("ERROR: fread failed"); |
|
From: Szakacsits S. <sz...@si...> - 2005-06-21 19:13:57
|
Hi, On Tue, 21 Jun 2005, Yura Pakhuchiy wrote: > Log Message: > ntfscp: Fix signal handling. [ ... once earlier .... ] > Update attribute definition handling. Please include in the log message what _exactly_ you fix, update. In itself these messages don't have really useful information. Additionally it's much easier and faster to check changes if we know what one really intend to fix or update. Thanks, Szaka |
|
From: Yura P. <pak...@ip...> - 2005-06-21 19:31:56
|
On Tue, 2005-06-21 at 22:12 +0300, Szakacsits Szabolcs wrote:
> On Tue, 21 Jun 2005, Yura Pakhuchiy wrote:
>
> > Log Message:
> > ntfscp: Fix signal handling.
>
> [ ... once earlier .... ]
>
> > Update attribute definition handling.
>
> Please include in the log message what _exactly_ you fix, update. In itself
> these messages don't have really useful information. Additionally it's much
> easier and faster to check changes if we know what one really intend to fix
> or update.
I thought this patch was quite small and obvious. If no, sorry, anyway I
didn't think that somebody watching for changes to ntfscp. OK, fixed
message will look like: "ntfscp: SIGINT was handled, but printf was
about SIGTERM. Now handle both and print respective message."
About "Update attribute definition handling.": this commit was ported
from kernel driver, and commit message I cut-n-pasted from Anton's
commit message.
--
Best regards,
Yura
|
|
From: Anton A. <ai...@ca...> - 2005-06-21 20:36:15
|
On Tue, 21 Jun 2005, Yura Pakhuchiy wrote: > On Tue, 2005-06-21 at 22:12 +0300, Szakacsits Szabolcs wrote: > > On Tue, 21 Jun 2005, Yura Pakhuchiy wrote: > > > Log Message: > > > ntfscp: Fix signal handling. > > > > [ ... once earlier .... ] > > > > > Update attribute definition handling. > > > > Please include in the log message what _exactly_ you fix, update. In itself > > these messages don't have really useful information. Additionally it's much > > easier and faster to check changes if we know what one really intend to fix > > or update. > > I thought this patch was quite small and obvious. If no, sorry, anyway I > didn't think that somebody watching for changes to ntfscp. OK, fixed > message will look like: "ntfscp: SIGINT was handled, but printf was > about SIGTERM. Now handle both and print respective message." > > About "Update attribute definition handling.": this commit was ported > from kernel driver, and commit message I cut-n-pasted from Anton's > commit message. Yes, my commit messages are not always very good. Sorry about that. )-: Best regards, Anton -- Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @) Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK Linux NTFS maintainer / IRC: #ntfs on irc.freenode.net WWW: http://linux-ntfs.sf.net/ & http://www-stu.christs.cam.ac.uk/~aia21/ |
|
From: Szakacsits S. <sz...@si...> - 2005-06-21 22:18:09
|
On Tue, 21 Jun 2005, Yura Pakhuchiy wrote: > > I thought this patch was quite small and obvious. The length and complexity of the patches are irrelevant from the log message's point of view. Expressive log messages help readers better understand what's going on in the project even without looking or understanding the relevant code and also help peer developers to revise the patches. > OK, fixed message will look like: "ntfscp: SIGINT was handled, > but printf was about SIGTERM. Now handle both and print respective > message." Wonderful! :-) BTW, printf isn't signal-safe, it can't be used in signal handlers. Szaka |
|
From: Yura P. <pak...@ip...> - 2005-06-22 23:38:32
|
On Wed, 2005-06-22 at 01:17 +0300, Szakacsits Szabolcs wrote:
> BTW, printf isn't signal-safe, it can't be used in signal handlers.
I'm sorry, but can you explain this problem more detailer? I read signal
man page and found nothing about it.
--
Best regards,
Yura
|
|
From: Szakacsits S. <sz...@si...> - 2005-06-23 23:07:33
|
On Thu, 23 Jun 2005, Yura Pakhuchiy wrote: > On Wed, 2005-06-22 at 01:17 +0300, Szakacsits Szabolcs wrote: >> BTW, printf isn't signal-safe, it can't be used in signal handlers. > > I'm sorry, but can you explain this problem more detailer? I read signal > man page and found nothing about it. 'man 2 signal' and see the NOTES section. Better, advanced Unix programming books explain signal handling related typical problems but google could help too, e.g. these pages look excellent: http://www.slamb.org/projects/sigsafe/api/patternref.html Szaka |
