CVS commit: xsrc/external/mit/fontconfig/dist/src

2023-01-22 Thread Christos Zoulas
Module Name:xsrc
Committed By:   christos
Date:   Sun Jan 22 20:47:13 UTC 2023

Modified Files:
xsrc/external/mit/fontconfig/dist/src: fcmatch.c

Log Message:
deal with vax exponent limit


To generate a diff of this commit:
cvs rdiff -u -r1.12 -r1.13 xsrc/external/mit/fontconfig/dist/src/fcmatch.c

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.

Modified files:

Index: xsrc/external/mit/fontconfig/dist/src/fcmatch.c
diff -u xsrc/external/mit/fontconfig/dist/src/fcmatch.c:1.12 xsrc/external/mit/fontconfig/dist/src/fcmatch.c:1.13
--- xsrc/external/mit/fontconfig/dist/src/fcmatch.c:1.12	Thu Jan 19 01:14:18 2023
+++ xsrc/external/mit/fontconfig/dist/src/fcmatch.c	Sun Jan 22 15:47:13 2023
@@ -26,6 +26,12 @@
 #include 
 
 
+#ifdef __vax__
+#define SENTINEL	1e38
+#else
+#define SENTINEL	1e99
+#endif
+
 static double
 FcCompareNumber (const FcValue *value1, const FcValue *value2, FcValue *bestValue)
 {
@@ -536,8 +542,8 @@ FcCompareDataInit (FcPattern *pat,
 if (!FcHashTableFind (table, key, (void **)))
 {
 e = malloc (sizeof (FamilyEntry));
-e->strong_value = 1e99;
-e->weak_value = 1e99;
+e->strong_value = SENTINEL;
+e->weak_value = SENTINEL;
 FcHashTableAdd (table, (void *)key, e);
 }
 if (l->binding == FcValueBindingWeak)
@@ -573,8 +579,8 @@ FcCompareFamilies (FcPattern   *pat,
 
 assert (table != NULL);
 
-strong_value = 1e99;
-weak_value = 1e99;
+strong_value = SENTINEL;
+weak_value = SENTINEL;
 
 for (v2 = v2orig; v2; v2 = FcValueListNext(v2))
 {



CVS commit: xsrc/external/mit/fontconfig/dist/src

2023-01-22 Thread Christos Zoulas
Module Name:xsrc
Committed By:   christos
Date:   Sun Jan 22 20:47:13 UTC 2023

Modified Files:
xsrc/external/mit/fontconfig/dist/src: fcmatch.c

Log Message:
deal with vax exponent limit


To generate a diff of this commit:
cvs rdiff -u -r1.12 -r1.13 xsrc/external/mit/fontconfig/dist/src/fcmatch.c

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.



CVS commit: xsrc/external/mit/fontconfig/dist/src

2019-05-16 Thread Takeshi Nakayama
Module Name:xsrc
Committed By:   nakayama
Date:   Thu May 16 09:59:37 UTC 2019

Modified Files:
xsrc/external/mit/fontconfig/dist/src: fcdir.c

Log Message:
Do not remove UUID file when a scanned directory is empty.

pulled from the following upstream change:
https://gitlab.freedesktop.org/fontconfig/fontconfig/commit/5f12f564f8748deaa603adb7a4b8f616b6390ad4

It fixes that a cache file for an empty font directory is created
every time fc-cache is executed.


To generate a diff of this commit:
cvs rdiff -u -r1.1.1.8 -r1.2 xsrc/external/mit/fontconfig/dist/src/fcdir.c

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.

Modified files:

Index: xsrc/external/mit/fontconfig/dist/src/fcdir.c
diff -u xsrc/external/mit/fontconfig/dist/src/fcdir.c:1.1.1.8 xsrc/external/mit/fontconfig/dist/src/fcdir.c:1.2
--- xsrc/external/mit/fontconfig/dist/src/fcdir.c:1.1.1.8	Fri Mar  8 09:40:41 2019
+++ xsrc/external/mit/fontconfig/dist/src/fcdir.c	Thu May 16 09:59:37 2019
@@ -421,13 +421,6 @@ FcDirCacheRead (const FcChar8 *dir, FcBo
 /* Not using existing cache file, construct new cache */
 if (!cache)
 	cache = FcDirCacheScan (dir, config);
-if (cache)
-{
-	FcFontSet *fs = FcCacheSet (cache);
-
-	if (cache->dirs_count == 0 && (!fs || fs->nfont == 0))
-	FcDirCacheDeleteUUID (dir, config);
-}
 
 return cache;
 }



CVS commit: xsrc/external/mit/fontconfig/dist/src

2019-05-16 Thread Takeshi Nakayama
Module Name:xsrc
Committed By:   nakayama
Date:   Thu May 16 09:59:37 UTC 2019

Modified Files:
xsrc/external/mit/fontconfig/dist/src: fcdir.c

Log Message:
Do not remove UUID file when a scanned directory is empty.

pulled from the following upstream change:
https://gitlab.freedesktop.org/fontconfig/fontconfig/commit/5f12f564f8748deaa603adb7a4b8f616b6390ad4

It fixes that a cache file for an empty font directory is created
every time fc-cache is executed.


To generate a diff of this commit:
cvs rdiff -u -r1.1.1.8 -r1.2 xsrc/external/mit/fontconfig/dist/src/fcdir.c

Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.



Re: CVS commit: xsrc/external/mit/fontconfig/dist/src

2013-06-27 Thread Valery Ushakov
On Thu, Jun 27, 2013 at 12:44:11 +, Thomas Klausner wrote:

 Modified Files:
   xsrc/external/mit/fontconfig/dist/src: fcname.c
 
 Log Message:
 Fix a comparison of constant warning with clang
 
 From upstream:
 commit 9acc14c34a372b54f9075ec3611588298fb2a501
 Author: Akira TAGOH ak...@tagoh.org
 Date:   Wed Jun 26 12:03:38 2013 +0900

Their fix is incorrect.

-   if (t-type == (unsigned int) -1 || type == t-type)
+   if ((unsigned int) t-type == (unsigned int) -1 || type == t-type)

it does wrong thing with short enums, and non-eabi arm ports still use
it, as far as I understand.

The enum in question contains only small values, so with -fshort-enums
(default for older arms ABIs) gcc will use the narrowest integral type
that can hold all the values, in this case single byte is enough, and
(though I don't remember chapter and verse) it's likely to be
unsigned.  Thus

  t-type = -1;

will be converted to uint8_t and stored as 0xff, and when cast to
unsigned int it will be 0xff, so it will NOT compare equal with
(unsigned int)-1

Consider something like:

#include stdio.h

enum e { E0, E1 };
volatile enum e e;

int
main(void)
{
printf(sizeof enum = %zu\n, sizeof(e));

e = -1;
if ((unsigned int)e == (unsigned int)-1) {
printf(equal\n);
}
else {
printf(NOT equal\n);
}
return 0;
}


$ cc -Wall -Wextra enum.c  ./a.out 
sizeof enum = 4
equal

$ cc -Wall -Wextra -fshort-enums enum.c  ./a.out 
enum.c: In function 'main':
enum.c:12:5: warning: comparison is always false due to limited range of data 
type
sizeof enum = 1
NOT equal

-1 should really be in the enum as one of the enumeration constants.

-uwe


Re: CVS commit: xsrc/external/mit/fontconfig/dist/src

2013-06-27 Thread Martin Husemann
On Thu, Jun 27, 2013 at 07:40:32PM +0400, Valery Ushakov wrote:
 -1 should really be in the enum as one of the enumeration constants.

Indeed, and then the enumerator name should be used instead of the
stupid -1 constant.

Martin