This commit fixes a locale problem with locales that use comma as a mantissa
separator. Example: 12.34 en_US = 12,34 pt_BR. Since strtod() is a non-safe
function, virStrToDouble() will have problems to parse double numbers from
kernel settings and other double numbers from static files (XMLs, JSONs, etc).

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457634
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457481

Signed-off-by: Julio Faracco <jcfaracco@gmail.com>
---
 src/util/virstring.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/src/util/virstring.c b/src/util/virstring.c
index 089b539..6dd5138 100644
--- a/src/util/virstring.c
+++ b/src/util/virstring.c
@@ -28,6 +28,7 @@
 #include "base64.h"
 #include "c-ctype.h"
 #include "virstring.h"
+#include "virthread.h"
 #include "viralloc.h"
 #include "virbuffer.h"
 #include "virerror.h"
@@ -516,6 +517,22 @@ virStrToLong_ullp(char const *s, char **end_ptr, int base,
     return 0;
 }
 
+#if HAVE_NEWLOCALE
+
+static locale_t virLocale;
+
+static int
+virLocaleOnceInit(void)
+{
+    virLocale = newlocale(LC_ALL_MASK, "C", (locale_t)0);
+    if (!virLocale)
+        return -1;
+    return 0;
+}
+
+VIR_ONCE_GLOBAL_INIT(virLocale);
+#endif
+
 int
 virStrToDouble(char const *s,
                char **end_ptr,
@@ -526,7 +543,17 @@ virStrToDouble(char const *s,
     int err;
 
     errno = 0;
+#if HAVE_NEWLOCALE
+    locale_t old_loc;
+    if (virLocaleInitialize() < 0)
+        return -1;
+
+    old_loc = uselocale(virLocale);
+#endif
     val = strtod(s, &p); /* exempt from syntax-check */
+#if HAVE_NEWLOCALE
+    uselocale(old_loc);
+#endif
     err = (errno || (!end_ptr && *p) || p == s);
     if (end_ptr)
         *end_ptr = p;
-- 
2.7.4

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On Sun, Jun 18, 2017 at 03:20:11PM -0300, Julio Faracco wrote:
>This commit fixes a locale problem with locales that use comma as a mantissa
>separator. Example: 12.34 en_US = 12,34 pt_BR. Since strtod() is a non-safe
>function, virStrToDouble() will have problems to parse double numbers from
>kernel settings and other double numbers from static files (XMLs, JSONs, etc).
>
>Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457634
>Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457481
>
>Signed-off-by: Julio Faracco <jcfaracco@gmail.com>
>---
> src/util/virstring.c | 27 +++++++++++++++++++++++++++
> 1 file changed, 27 insertions(+)
>

I don't really like the duplication of the data and code.  I would
rather see virDoubleToStr move to virstring (it can be called
virStrFromDouble if some don't like the move), it is called from one
place anyway and that way it can share some of the code and data at
least.

I apologize for asking you for yet another version (and possibly
splitting it into two patches -- the move and the fix), but I haven't
notice in the previous submission.

Have a nice day,
Martin
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

Hi Martin,

2017-06-20 11:42 GMT-03:00 Martin Kletzander <mkletzan@redhat.com>:
> On Sun, Jun 18, 2017 at 03:20:11PM -0300, Julio Faracco wrote:
>>
>> This commit fixes a locale problem with locales that use comma as a
>> mantissa
>> separator. Example: 12.34 en_US = 12,34 pt_BR. Since strtod() is a
>> non-safe
>> function, virStrToDouble() will have problems to parse double numbers from
>> kernel settings and other double numbers from static files (XMLs, JSONs,
>> etc).
>>
>> Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457634
>> Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1457481
>>
>> Signed-off-by: Julio Faracco <jcfaracco@gmail.com>
>> ---
>> src/util/virstring.c | 27 +++++++++++++++++++++++++++
>> 1 file changed, 27 insertions(+)
>>
>
> I don't really like the duplication of the data and code.  I would

I totally agree. I needed to duplicate the code because virDoubleToStr()
is inside virutil as you mentioned.

> rather see virDoubleToStr move to virstring (it can be called
> virStrFromDouble if some don't like the move), it is called from one
> place anyway and that way it can share some of the code and data at
> least.
>

I will join the methods inside virstring and submit a v3.

> I apologize for asking you for yet another version (and possibly
> splitting it into two patches -- the move and the fix), but I haven't
> notice in the previous submission.
>
> Have a nice day,
> Martin

Thanks!

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list