Thread: plperl / locale / win32
I was reminded today of the outstanding issue with plperl setting the
locale on Windows, and our environment workaround not working there.
There has been NO response to the bug I filed (#38193) at rt.perl.org
about this issue.
The attached patch adapts one I previously tested as working, but
instead of calling setlocale() directly it gets perl to do it so that
perl and postgres have the same idea of what the locale should be, which
should meet Greg's and Tom's objection to the previous patch. My Windows
machine is currently doing other work, so I can't test right now - if
someone else could that would be nice.
cheers
andrew
*** plperl.c 2006-01-08 17:27:52.000000000 -0500
--- plperl.c.locfix 2006-01-20 19:50:04.000000000 -0500
***************
*** 45,50 ****
--- 45,51 ----
#include <ctype.h>
#include <fcntl.h>
#include <unistd.h>
+ #include <locale.h>
/* postgreSQL stuff */
#include "commands/trigger.h"
***************
*** 252,257 ****
--- 253,297 ----
"", "-e", PERLBOOT
};
+ #ifdef WIN32
+
+ /*
+ * The perl library on startup does horrible things like call
+ * setlocale(LC_ALL,""). We have protected against that on most
+ * platforms by setting the environment appropriately. However, on
+ * Windows, setlocale() does not consult the environment, so we need
+ * to save the excisting locale settings before perl has a chance to
+ * mangle them and restore them after its dirty deeds are done.
+ *
+ * MSDN ref:
+ * http://msdn.microsoft.com/library/en-us/vclib/html/_crt_locale.asp
+ *
+ * It appaers that we only need to do this on interpreter startup, and
+ * subsequent calls to the interpreter don't mess with the locale
+ * settings.
+ *
+ * We restore them using Perl's POSIX::setlocale() function so that
+ * Perl doesn't have a different idea of the locale from Postgres.
+ *
+ */
+
+ char *loc;
+ char *save_collate, *save_ctype, *save_monetary, *save_numeric, *save_time;
+ char buf[1024];
+
+ loc = setlocale(LC_COLLATE,NULL);
+ save_collate = loc ? pstrdup(loc) : NULL;
+ loc = setlocale(LC_CTYPE,NULL);
+ save_ctype = loc ? pstrdup(loc) : NULL;
+ loc = setlocale(LC_MONETARY,NULL);
+ save_monetary = loc ? pstrdup(loc) : NULL;
+ loc = setlocale(LC_NUMERIC,NULL);
+ save_numeric = loc ? pstrdup(loc) : NULL;
+ loc = setlocale(LC_TIME,NULL);
+ save_time = loc ? pstrdup(loc) : NULL;
+
+ #endif
+
plperl_interp = perl_alloc();
if (!plperl_interp)
elog(ERROR, "could not allocate Perl interpreter");
***************
*** 261,266 ****
--- 301,349 ----
perl_run(plperl_interp);
plperl_proc_hash = newHV();
+
+ #ifdef WIN32
+
+ eval_pv("use POSIX qw(locale_h);", TRUE); /* croak on failure */
+
+ if (save_collate != NULL)
+ {
+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');",
+ "LC_COLLATE",save_collate);
+ eval_pv(buf,TRUE);
+ pfree(save_collate);
+ }
+ if (save_ctype != NULL)
+ {
+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');",
+ "LC_CTYPE",save_ctype);
+ eval_pv(buf,TRUE);
+ pfree(save_ctype);
+ }
+ if (save_monetary != NULL)
+ {
+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');",
+ "LC_MONETARY",save_monetary);
+ eval_pv(buf,TRUE);
+ pfree(save_monetary);
+ }
+ if (save_numeric != NULL)
+ {
+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');",
+ "LC_NUMERIC",save_numeric);
+ eval_pv(buf,TRUE);
+ pfree(save_numeric);
+ }
+ if (save_time != NULL)
+ {
+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');",
+ "LC_TIME",save_time);
+ eval_pv(buf,TRUE);
+ pfree(save_time);
+ }
+
+ #endif
+
}
I have now tested the patch. It passes regression and the test case Tom previously posted. Unless there's an objection I will apply it and backport it in the next few days. cheers andrew I wrote: > > I was reminded today of the outstanding issue with plperl setting the > locale on Windows, and our environment workaround not working there. > > There has been NO response to the bug I filed (#38193) at rt.perl.org > about this issue. > > The attached patch adapts one I previously tested as working, but > instead of calling setlocale() directly it gets perl to do it so that > perl and postgres have the same idea of what the locale should be, > which should meet Greg's and Tom's objection to the previous patch. My > Windows machine is currently doing other work, so I can't test right > now - if someone else could that would be nice. > >------------------------------------------------------------------------ > >*** plperl.c 2006-01-08 17:27:52.000000000 -0500 >--- plperl.c.locfix 2006-01-20 19:50:04.000000000 -0500 >*************** >*** 45,50 **** >--- 45,51 ---- > #include <ctype.h> > #include <fcntl.h> > #include <unistd.h> >+ #include <locale.h> > > /* postgreSQL stuff */ > #include "commands/trigger.h" >*************** >*** 252,257 **** >--- 253,297 ---- > "", "-e", PERLBOOT > }; > >+ #ifdef WIN32 >+ >+ /* >+ * The perl library on startup does horrible things like call >+ * setlocale(LC_ALL,""). We have protected against that on most >+ * platforms by setting the environment appropriately. However, on >+ * Windows, setlocale() does not consult the environment, so we need >+ * to save the excisting locale settings before perl has a chance to >+ * mangle them and restore them after its dirty deeds are done. >+ * >+ * MSDN ref: >+ * http://msdn.microsoft.com/library/en-us/vclib/html/_crt_locale.asp >+ * >+ * It appaers that we only need to do this on interpreter startup, and >+ * subsequent calls to the interpreter don't mess with the locale >+ * settings. >+ * >+ * We restore them using Perl's POSIX::setlocale() function so that >+ * Perl doesn't have a different idea of the locale from Postgres. >+ * >+ */ >+ >+ char *loc; >+ char *save_collate, *save_ctype, *save_monetary, *save_numeric, *save_time; >+ char buf[1024]; >+ >+ loc = setlocale(LC_COLLATE,NULL); >+ save_collate = loc ? pstrdup(loc) : NULL; >+ loc = setlocale(LC_CTYPE,NULL); >+ save_ctype = loc ? pstrdup(loc) : NULL; >+ loc = setlocale(LC_MONETARY,NULL); >+ save_monetary = loc ? pstrdup(loc) : NULL; >+ loc = setlocale(LC_NUMERIC,NULL); >+ save_numeric = loc ? pstrdup(loc) : NULL; >+ loc = setlocale(LC_TIME,NULL); >+ save_time = loc ? pstrdup(loc) : NULL; >+ >+ #endif >+ > plperl_interp = perl_alloc(); > if (!plperl_interp) > elog(ERROR, "could not allocate Perl interpreter"); >*************** >*** 261,266 **** >--- 301,349 ---- > perl_run(plperl_interp); > > plperl_proc_hash = newHV(); >+ >+ #ifdef WIN32 >+ >+ eval_pv("use POSIX qw(locale_h);", TRUE); /* croak on failure */ >+ >+ if (save_collate != NULL) >+ { >+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');", >+ "LC_COLLATE",save_collate); >+ eval_pv(buf,TRUE); >+ pfree(save_collate); >+ } >+ if (save_ctype != NULL) >+ { >+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');", >+ "LC_CTYPE",save_ctype); >+ eval_pv(buf,TRUE); >+ pfree(save_ctype); >+ } >+ if (save_monetary != NULL) >+ { >+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');", >+ "LC_MONETARY",save_monetary); >+ eval_pv(buf,TRUE); >+ pfree(save_monetary); >+ } >+ if (save_numeric != NULL) >+ { >+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');", >+ "LC_NUMERIC",save_numeric); >+ eval_pv(buf,TRUE); >+ pfree(save_numeric); >+ } >+ if (save_time != NULL) >+ { >+ snprintf(buf, sizeof(buf),"setlocale(%s,'%s');", >+ "LC_TIME",save_time); >+ eval_pv(buf,TRUE); >+ pfree(save_time); >+ } >+ >+ #endif >+ > } > > > > >------------------------------------------------------------------------ > > >---------------------------(end of broadcast)--------------------------- >TIP 2: Don't 'kill -9' the postmaster > >
applied to HEAD, 8.1 and 8.0. cheers andrew I wrote: > > I have now tested the patch. It passes regression and the test case > Tom previously posted. Unless there's an objection I will apply it and > backport it in the next few days. > > cheers > > andrew > > I wrote: > >> >> I was reminded today of the outstanding issue with plperl setting the >> locale on Windows, and our environment workaround not working there. >> >> There has been NO response to the bug I filed (#38193) at rt.perl.org >> about this issue. >> >> The attached patch adapts one I previously tested as working, but >> instead of calling setlocale() directly it gets perl to do it so that >> perl and postgres have the same idea of what the locale should be, >> which should meet Greg's and Tom's objection to the previous patch. >> My Windows machine is currently doing other work, so I can't test >> right now - if someone else could that would be nice. > >