\x output blowing up
Hi,
On CVS tip, whenever I do \x output, it dies with an internal error in
glibc on free. If you run it under valgrind, it complains about these
lines of code:
700 {
701 char *my_cell = pg_local_malloc(cell_w[i] + 1);
702
703 [Inv write 1 byte] strcpy(my_cell, *ptr);
704 if (opt_align[i % col_count] == 'r' && opt_numeric_locale)
705 format_numeric_locale(my_cell);
706 if (opt_border < 2)
707 fprintf(fout, "%s\n", my_cell);
708 else
709 [Inv read 1 byte] fprintf(fout, "%-s%*s |\n", my_cell, dwidth - cell_w[i], "");
710 free(my_cell);
711 }
Now, apart from the fact that the cell width != strlen in multibyte
encodings, there must be something else because this is just "select *
from pg_proc" and there are no multiple characters there AFAIK. I can't
see it though.
Hope this helps,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
Show quoted text
Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.
On Sat, Sep 24, 2005 at 11:45:08PM +0200, Martijn van Oosterhout wrote:
Hi,
On CVS tip, whenever I do \x output, it dies with an internal error in
glibc on free. If you run it under valgrind, it complains about these
lines of code:
<snip>
Ok, I worked out the direct cause, pg_wcswidth only returns the length
upto the first newline and the line it breaks on is the multiline
definition in "_pg_expandarray".
The quick fix should be to only allocate memory if it's going to call
format_numeric_locale(), since then you know it's a number. It makes
the code slightly more convoluated but it should be slightly more
efficient.
I actually have a working psql that handles and displays newlines
properly, but it's too late for 8.1. It fixes all these issues
properly.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
Show quoted text
Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.
Well, it seems we are going to have to fix it somehow for 8.1. It is
not crashing here so I can't work up a patch. Can you submit a minimal
fix for 8.1? Thanks.
---------------------------------------------------------------------------
Martijn van Oosterhout wrote:
-- Start of PGP signed section.
On Sat, Sep 24, 2005 at 11:45:08PM +0200, Martijn van Oosterhout wrote:
Hi,
On CVS tip, whenever I do \x output, it dies with an internal error in
glibc on free. If you run it under valgrind, it complains about these
lines of code:<snip>
Ok, I worked out the direct cause, pg_wcswidth only returns the length
upto the first newline and the line it breaks on is the multiline
definition in "_pg_expandarray".The quick fix should be to only allocate memory if it's going to call
format_numeric_locale(), since then you know it's a number. It makes
the code slightly more convoluated but it should be slightly more
efficient.I actually have a working psql that handles and displays newlines
properly, but it's too late for 8.1. It fixes all these issues
properly.Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.
-- End of PGP section, PGP failed!
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073On Sat, Sep 24, 2005 at 07:18:16PM -0400, Bruce Momjian wrote:
Well, it seems we are going to have to fix it somehow for 8.1. It is
not crashing here so I can't work up a patch. Can you submit a minimal
fix for 8.1? Thanks.
Ah, it would only happen if your encoding was UTF-8 since that's the
only case psql handles differently. I've attached a patch which fixes
it. With a bit more rearrangement you could probably simplify it a bit
but this works.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
Show quoted text
Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.
Attachments:
patchtext/plain; charset=us-asciiDownload+16-3
Martijn van Oosterhout wrote:
-- Start of PGP signed section.
On Sat, Sep 24, 2005 at 07:18:16PM -0400, Bruce Momjian wrote:
Well, it seems we are going to have to fix it somehow for 8.1. It is
not crashing here so I can't work up a patch. Can you submit a minimal
fix for 8.1? Thanks.Ah, it would only happen if your encoding was UTF-8 since that's the
only case psql handles differently. I've attached a patch which fixes
it. With a bit more rearrangement you could probably simplify it a bit
but this works.
Fixed. You were right that the use of cell_w was incorrect for
non-numeric values (UTF8), and in fact was just too fragile to use.
I redesigned format_numeric_locale() to return an allocated result,
which removed this problem and simplified the code too.
Patch attached and applied.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073Attachments:
/rtmp/difftext/plainDownload+84-82
On Tue, Sep 27, 2005 at 12:32:14PM -0400, Bruce Momjian wrote:
I redesigned format_numeric_locale() to return an allocated result,
which removed this problem and simplified the code too.
Much better, I had that in my own tree but figured it wasn't "minimal"
enough for this stage of beta. It is, IMHO, the much better solution.
Thanks for the fix.
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
Show quoted text
Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.