Accepting request 487561 from home:jsmeix:branches:Printing

Ghostscript version upgrade to 9.21

OBS-URL: https://build.opensuse.org/request/show/487561
OBS-URL: https://build.opensuse.org/package/show/Printing/ghostscript?expand=0&rev=76
This commit is contained in:
Johannes Meixner 2017-04-12 12:35:15 +00:00 committed by Git OBS Bridge
parent 7fb4de27c7
commit 1ec563fa1e
14 changed files with 238 additions and 556 deletions

View File

@ -1,49 +0,0 @@
From 495b79d8e9a44ad0ada965add3a046120646e7e0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Stefan=20Br=C3=83=C2=BCns?= <stefan.bruens@rwth-aachen.de>
Date: Thu, 12 Jan 2017 18:04:57 +0100
Subject: [PATCH 1/2] mkromfs: make build reproducible, use buildtime from
SOURCE_DATE_EPOCH
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
The environment variable SOURCE_DATE_EPOCH is the common approach for
getting reproducible timestamps and thus builds. In case the variable
is not set, keep using the current time of the mkromfs run.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
---
base/mkromfs.c | 10 +++++++++-
1 file changed, 9 insertions(+), 1 deletion(-)
diff --git a/base/mkromfs.c b/base/mkromfs.c
index b7bc6bf..9b81c46 100644
--- a/base/mkromfs.c
+++ b/base/mkromfs.c
@@ -2280,6 +2280,8 @@ main(int argc, char *argv[])
int compaction = 0;
Xlist_element *Xlist_scan = NULL, *Xlist_head = NULL;
char pa[PATH_STR_LEN];
+ time_t buildtime = 0;
+ char* env_source_date_epoch;
memset(pa, 0x00, PATH_STR_LEN);
@@ -2336,7 +2338,13 @@ main(int argc, char *argv[])
#endif
fprintf(out,"\n#include \"stdint_.h\"\n");
fprintf(out,"\n#include \"time_.h\"\n\n");
- fprintf(out," time_t gs_romfs_buildtime = %ld;\n\n", time(NULL));
+
+ if ((env_source_date_epoch = getenv("SOURCE_DATE_EPOCH"))) {
+ buildtime = strtoul(env_source_date_epoch, NULL, 10);
+ }
+ if (!buildtime)
+ buildtime = time(NULL);
+ fprintf(out," time_t gs_romfs_buildtime = %ld;\n\n", buildtime);
/* process the remaining arguments (options interspersed with paths) */
for (; atarg < argc; atarg++) {
--
2.11.0

View File

@ -1,75 +0,0 @@
From ec602a6eadfe7680e0a1008a67afa18903a07ad9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Stefan=20Br=C3=BCns?= <stefan.bruens@rwth-aachen.de>
Date: Thu, 12 Jan 2017 20:55:11 +0100
Subject: [PATCH 2/2] mkromfs: sort gp_enumerate_files output for deterministic
ROM contents
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
gp_enumerate_files_next returns dir entries in the same order as returned
by readdir. Sort by name to generate deterministic output.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
---
base/mkromfs.c | 23 +++++++++++++++++++++++
1 file changed, 23 insertions(+)
diff --git a/base/mkromfs.c b/base/mkromfs.c
index 9b81c46..b7799ae 100644
--- a/base/mkromfs.c
+++ b/base/mkromfs.c
@@ -344,6 +344,7 @@ const gs_malloc_memory_t minimal_memory = {
0 /* max used */
};
+int cmpstringp(const void *p1, const void *p2);
void put_uint32(FILE *out, const unsigned int q);
void put_bytes_padded(FILE *out, unsigned char *p, unsigned int len);
void inode_clear(romfs_inode* node);
@@ -1542,6 +1543,15 @@ static unsigned long pscompact_getcompactedblock(pscompstate *psc, unsigned char
return out-ubuf;
}
+int cmpstringp(const void *p1, const void *p2)
+{
+ /* The actual arguments to this function are "pointers to
+ pointers to char", but strcmp(3) arguments are "pointers
+ to char", hence the following cast plus dereference */
+
+ return strcmp(* (char * const *) p1, * (char * const *) p2);
+}
+
/* This relies on the gp_enumerate_* which should not return directories, nor */
/* should it recurse into directories (unlike Adobe's implementation) */
/* paths are checked to see if they are an ordinary file or a path */
@@ -1561,6 +1571,8 @@ void process_path(char *path, const char *os_prefix, const char *rom_prefix,
FILE *in;
unsigned long psc_len;
pscompstate psc = { 0 };
+ unsigned long numfiles = 0;
+ char **foundfiles = NULL;
prefixed_path = malloc(PATH_STR_LEN);
found_path = malloc(PATH_STR_LEN);
@@ -1598,6 +1610,17 @@ void process_path(char *path, const char *os_prefix, const char *rom_prefix,
if (excluded)
continue;
+ numfiles++;
+ foundfiles = realloc(foundfiles, sizeof(char *) * numfiles);
+ foundfiles[numfiles - 1] = strdup(found_path);
+ }
+
+ qsort(foundfiles, numfiles, sizeof(char *), cmpstringp);
+
+ while (numfiles--) {
+ found_path = *foundfiles;
+ foundfiles++;
+
/* process a file */
node = calloc(1, sizeof(romfs_inode));
/* get info for this file */
--
2.11.0

View File

@ -1,171 +0,0 @@
From: Chris Liddell <chris.liddell@artifex.com>
Date: Wed, 5 Oct 2016 08:55:55 +0000 (+0100)
Subject: Bug 697178: Add a file permissions callback
X-Git-Url: http://git.ghostscript.com/?p=user%2Fchrisl%2Fghostpdl.git;a=commitdiff_plain;h=71ac87493b1e445d6c07554d4246cf7d4f44875c;hp=cf046d2f0fa2c6973c6ca8d582a9b185cc4bd280
Bug 697178: Add a file permissions callback
For the rare occasions when the graphics library directly opens a file
(currently for reading), this allows us to apply any restrictions on
file access normally applied in the interpteter.
---
diff --git a/base/gsicc_manage.c b/base/gsicc_manage.c
index 931c2a6..e9c09c3 100644
--- a/base/gsicc_manage.c
+++ b/base/gsicc_manage.c
@@ -1124,10 +1124,12 @@ gsicc_open_search(const char* pname, int namelen, gs_memory_t *mem_gc,
}
/* First just try it like it is */
- str = sfopen(pname, "r", mem_gc);
- if (str != NULL) {
- *strp = str;
- return 0;
+ if (gs_check_file_permission(mem_gc, pname, namelen, "r") >= 0) {
+ str = sfopen(pname, "r", mem_gc);
+ if (str != NULL) {
+ *strp = str;
+ return 0;
+ }
}
/* If that fails, try %rom% */ /* FIXME: Not sure this is needed or correct */
diff --git a/base/gslibctx.c b/base/gslibctx.c
index eaa0458..37ce1ca 100644
--- a/base/gslibctx.c
+++ b/base/gslibctx.c
@@ -189,7 +189,7 @@ Failure:
mem->gs_lib_ctx = NULL;
return -1;
}
-
+ pio->client_check_file_permission = NULL;
gp_get_realtime(pio->real_time_0);
/* Set scanconverter to 1 (default) */
@@ -343,3 +343,13 @@ void errflush(const gs_memory_t *mem)
fflush(mem->gs_lib_ctx->fstderr);
/* else nothing to flush */
}
+
+int
+gs_check_file_permission (gs_memory_t *mem, const char *fname, const int len, const char *permission)
+{
+ int code = 0;
+ if (mem->gs_lib_ctx->client_check_file_permission != NULL) {
+ code = mem->gs_lib_ctx->client_check_file_permission(mem, fname, len, permission);
+ }
+ return code;
+}
diff --git a/base/gslibctx.h b/base/gslibctx.h
index 7a4e110..020e2d9 100644
--- a/base/gslibctx.h
+++ b/base/gslibctx.h
@@ -32,6 +32,9 @@ typedef struct gs_fapi_server_s gs_fapi_server;
# define gs_font_dir_DEFINED
typedef struct gs_font_dir_s gs_font_dir;
#endif
+
+typedef int (*client_check_file_permission_t) (gs_memory_t *mem, const char *fname, const int len, const char *permission);
+
typedef struct gs_lib_ctx_s
{
gs_memory_t *memory; /* mem->gs_lib_ctx->memory == mem */
@@ -61,6 +64,7 @@ typedef struct gs_lib_ctx_s
struct gx_io_device_s **io_device_table;
int io_device_table_count;
int io_device_table_size;
+ client_check_file_permission_t client_check_file_permission;
/* Define the default value of AccurateScreens that affects setscreen
and setcolorscreen. */
bool screen_accurate_screens;
@@ -133,6 +137,9 @@ int
gs_lib_ctx_get_default_device_list(const gs_memory_t *mem, char** dev_list_str,
int *list_str_len);
+int
+gs_check_file_permission (gs_memory_t *mem, const char *fname, const int len, const char *permission);
+
#define IS_LIBCTX_STDOUT(mem, f) (f == mem->gs_lib_ctx->fstdout)
#define IS_LIBCTX_STDERR(mem, f) (f == mem->gs_lib_ctx->fstderr)
diff --git a/psi/imain.c b/psi/imain.c
index 9a9bb5d..6874128 100644
--- a/psi/imain.c
+++ b/psi/imain.c
@@ -57,6 +57,7 @@
#include "ivmspace.h"
#include "idisp.h" /* for setting display device callback */
#include "iplugin.h"
+#include "zfile.h"
#ifdef PACIFY_VALGRIND
#include "valgrind.h"
@@ -212,6 +213,7 @@ gs_main_init1(gs_main_instance * minst)
"the_gs_name_table");
if (code < 0)
return code;
+ mem->gs_lib_ctx->client_check_file_permission = z_check_file_permissions;
}
code = obj_init(&minst->i_ctx_p, &idmem); /* requires name_init */
if (code < 0)
diff --git a/psi/int.mak b/psi/int.mak
index 4654afc..bb30d51 100644
--- a/psi/int.mak
+++ b/psi/int.mak
@@ -2024,7 +2024,7 @@ $(PSOBJ)imain.$(OBJ) : $(PSSRC)imain.c $(GH) $(memory__h) $(string__h)\
$(ialloc_h) $(iconf_h) $(idebug_h) $(idict_h) $(idisp_h) $(iinit_h)\
$(iname_h) $(interp_h) $(iplugin_h) $(isave_h) $(iscan_h) $(ivmspace_h)\
$(iinit_h) $(main_h) $(oper_h) $(ostack_h)\
- $(sfilter_h) $(store_h) $(stream_h) $(strimpl_h)\
+ $(sfilter_h) $(store_h) $(stream_h) $(strimpl_h) $(zfile_h)\
$(INT_MAK) $(MAKEDIRS)
$(PSCC) $(PSO_)imain.$(OBJ) $(C_) $(PSSRC)imain.c
diff --git a/psi/zfile.c b/psi/zfile.c
index 2c6c958..bc6c70f 100644
--- a/psi/zfile.c
+++ b/psi/zfile.c
@@ -197,6 +197,25 @@ check_file_permissions(i_ctx_t *i_ctx_p, const char *fname, int len,
return check_file_permissions_reduced(i_ctx_p, fname_reduced, rlen, permitgroup);
}
+/* z_check_file_permissions: see zfile.h for explanation
+ */
+int
+z_check_file_permissions(gs_memory_t *mem, const char *fname, const int len, const char *permission)
+{
+ i_ctx_t *i_ctx_p = get_minst_from_memory(mem)->i_ctx_p;
+ gs_parsed_file_name_t pname;
+ char *permitgroup = permission[0] == 'r' ? "PermitFileReading" : "PermitFileWriting";
+ int code = gs_parse_file_name(&pname, fname, len, imemory);
+ if (code < 0)
+ return code;
+
+ if (pname.iodev && i_ctx_p->LockFilePermissions && strcmp(pname.iodev->dname, "%pipe%") == 0)
+ return gs_error_invalidfileaccess;
+
+ code = check_file_permissions(i_ctx_p, fname, len, permitgroup);
+ return code;
+}
+
/* <name_string> <access_string> file <file> */
int /* exported for zsysvm.c */
zfile(i_ctx_t *i_ctx_p)
diff --git a/psi/zfile.h b/psi/zfile.h
index fdf1373..a9399c7 100644
--- a/psi/zfile.h
+++ b/psi/zfile.h
@@ -22,4 +22,11 @@
int zopen_file(i_ctx_t *i_ctx_p, const gs_parsed_file_name_t *pfn,
const char *file_access, stream **ps, gs_memory_t *mem);
+/* z_check_file_permissions: a callback (via mem->gs_lib_ctx->client_check_file_permission)
+ * to allow applying the above permissions checks when opening file(s) from
+ * the graphics library
+ */
+int
+z_check_file_permissions(gs_memory_t *mem, const char *fname,
+ const int len, const char *permission);
#endif

View File

@ -1,25 +0,0 @@
From: Chris Liddell <chris.liddell@artifex.com>
Date: Mon, 3 Oct 2016 00:46:28 +0000 (+0100)
Subject: Bug 697169: Be rigorous with SAFER permissions
X-Git-Url: http://git.ghostscript.com/?p=user%2Fchrisl%2Fghostpdl.git;a=commitdiff_plain;h=cf046d2f0fa2c6973c6ca8d582a9b185cc4bd280;hp=3826c0c7a4fc781c8222ef458b706360600f1711
Bug 697169: Be rigorous with SAFER permissions
Once we've opened our input file from the command line, enforce the SAFER
rules.
---
diff --git a/psi/zfile.c b/psi/zfile.c
index b6caea2..2c6c958 100644
--- a/psi/zfile.c
+++ b/psi/zfile.c
@@ -1081,6 +1081,9 @@ lib_file_open(gs_file_path_ptr lib_path, const gs_memory_t *mem, i_ctx_t *i_ctx
gs_main_instance *minst = get_minst_from_memory(mem);
int code;
+ if (i_ctx_p && starting_arg_file)
+ i_ctx_p->starting_arg_file = false;
+
/* when starting arg files (@ files) iodev_default is not yet set */
if (iodev == 0)
iodev = (gx_io_device *)gx_io_device_table[0];

View File

@ -1,22 +0,0 @@
From: Chris Liddell <chris.liddell@artifex.com>
Date: Wed, 5 Oct 2016 08:59:25 +0000 (+0100)
Subject: Bug 697179: Reference count device icc profile
X-Git-Url: http://git.ghostscript.com/?p=user%2Fchrisl%2Fghostpdl.git;a=commitdiff_plain;h=d5ad1e0298e1c193087c824eb4f79628b182e28b;hp=71ac87493b1e445d6c07554d4246cf7d4f44875c
Bug 697179: Reference count device icc profile
when copying a device
---
diff --git a/base/gsdevice.c b/base/gsdevice.c
index 778106f..aea986a 100644
--- a/base/gsdevice.c
+++ b/base/gsdevice.c
@@ -614,6 +614,7 @@ gx_device_init(gx_device * dev, const gx_device * proto, gs_memory_t * mem,
dev->memory = mem;
dev->retained = !internal;
rc_init(dev, mem, (internal ? 0 : 1));
+ rc_increment(dev->icc_struct);
}
void

View File

@ -1,40 +0,0 @@
From: Ken Sharp <ken.sharp@artifex.com>
Date: Wed, 5 Oct 2016 09:10:58 +0000 (+0100)
Subject: DSC parser - validate parameters
X-Git-Url: http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff_plain;h=875a0095f37626a721c7ff57d606a0f95af03913;hp=aef8f6cfbff9080e4b9c54053bc909875c09a2be
DSC parser - validate parameters
Bug #697190 ".initialize_dsc_parser doesn't validate the parameter is a dict type before using it."
Regardless of any security implications, its simply wrong for a PostScript
operator not to validate its parameter(s).
No differences expected.
---
diff --git a/psi/zdscpars.c b/psi/zdscpars.c
index c05e154..9b4b605 100644
--- a/psi/zdscpars.c
+++ b/psi/zdscpars.c
@@ -150,11 +150,16 @@ zinitialize_dsc_parser(i_ctx_t *i_ctx_p)
ref local_ref;
int code;
os_ptr const op = osp;
- dict * const pdict = op->value.pdict;
- gs_memory_t * const mem = (gs_memory_t *)dict_memory(pdict);
- dsc_data_t * const data =
- gs_alloc_struct(mem, dsc_data_t, &st_dsc_data_t, "DSC parser init");
+ dict *pdict;
+ gs_memory_t *mem;
+ dsc_data_t *data;
+ check_read_type(*op, t_dictionary);
+
+ pdict = op->value.pdict;
+ mem = (gs_memory_t *)dict_memory(pdict);
+
+ data = gs_alloc_struct(mem, dsc_data_t, &st_dsc_data_t, "DSC parser init");
if (!data)
return_error(gs_error_VMerror);
data->document_level = 0;

View File

@ -1,39 +0,0 @@
From: Chris Liddell <chris.liddell@artifex.com>
Date: Sat, 8 Oct 2016 15:10:27 +0000 (+0100)
Subject: Bug 697203: check for sufficient params in .sethalftone5
X-Git-Url: http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff_plain;h=f5c7555c30393e64ec1f5ab0dfae5b55b3b3fc78;hp=a5360401495654e89301b2516703c1d2877fc5ba
Bug 697203: check for sufficient params in .sethalftone5
and param types
---
diff --git a/psi/zht2.c b/psi/zht2.c
index fb4a264..dfa27a4 100644
--- a/psi/zht2.c
+++ b/psi/zht2.c
@@ -82,14 +82,22 @@ zsethalftone5(i_ctx_t *i_ctx_p)
gs_memory_t *mem;
uint edepth = ref_stack_count(&e_stack);
int npop = 2;
- int dict_enum = dict_first(op);
+ int dict_enum;
ref rvalue[2];
int cname, colorant_number;
byte * pname;
uint name_size;
int halftonetype, type = 0;
gs_gstate *pgs = igs;
- int space_index = r_space_index(op - 1);
+ int space_index;
+
+ if (ref_stack_count(&o_stack) < 2)
+ return_error(gs_error_stackunderflow);
+ check_type(*op, t_dictionary);
+ check_type(*(op - 1), t_dictionary);
+
+ dict_enum = dict_first(op);
+ space_index = r_space_index(op - 1);
mem = (gs_memory_t *) idmemory->spaces_indexed[space_index];

19
CVE-2017-5951.patch Normal file
View File

@ -0,0 +1,19 @@
--- psi/iparam.c.orig 2017-03-16 11:12:02.000000000 +0100
+++ psi/iparam.c 2017-04-12 11:42:57.000000000 +0200
@@ -770,12 +770,13 @@ ref_param_read_typed(gs_param_list * pli
gs_param_enumerator_t enumr;
gs_param_key_t key;
ref_type keytype;
+ dict_param_list *dlist = (dict_param_list *) pvalue->value.d.list;
param_init_enumerator(&enumr);
- if (!(*((iparam_list *) plist)->enumerate)
- ((iparam_list *) pvalue->value.d.list, &enumr, &key, &keytype)
+ if (!(*(dlist->enumerate))
+ ((iparam_list *) dlist, &enumr, &key, &keytype)
&& keytype == t_integer) {
- ((dict_param_list *) pvalue->value.d.list)->int_keys = 1;
+ dlist->int_keys = 1;
pvalue->type = gs_param_type_dict_int_keys;
}
}

View File

@ -1,3 +0,0 @@
version https://git-lfs.github.com/spec/v1
oid sha256:949b64b46ecf8906db54a94ecf83ab97534ebf946f770d3c3f283cb469cb6e14
size 36080631

3
ghostscript-9.21.tar.gz Normal file
View File

@ -0,0 +1,3 @@
version https://git-lfs.github.com/spec/v1
oid sha256:02bceadbc4dddeb6f2eec9c8b1623d945d355ca11b8b4df035332b217d58ce85
size 38398778

View File

@ -1,3 +1,90 @@
-------------------------------------------------------------------
Wed Apr 12 11:12:27 CEST 2017 - jsmeix@suse.de
- CVE-2016-10317 (bsc#1032230)
heap buffer overflow in fill_threshhold_buffer()
is not yet fixed because there is no fix available at
https://bugs.ghostscript.com/show_bug.cgi?id=697459
- CVE-2016-10219 (bsc#1032138)
divide by zero in intersect()
https://bugs.ghostscript.com/show_bug.cgi?id=697453
is fixed in the version 9.21 upstream sources
- CVE-2016-10218 (bsc#1032135)
null pointer dereference in pdf14_pop_transparency_group()
https://bugs.ghostscript.com/show_bug.cgi?id=697444
is fixed in the version 9.21 upstream sources.
- CVE-2016-10217 (bsc#1032130)
use-after-free in pdf14_cleanup_parent_color_profiles()
that is related to pdf14_open() in base/gdevp14.c
https://bugs.ghostscript.com/show_bug.cgi?id=697456
is fixed in the version 9.21 upstream sources.
- CVE-2016-10220 (bsc#1032120)
null pointer dereference in gx_device_finalize() that is
related to gs_makewordimagedevice() in base/gsdevmem.c
https://bugs.ghostscript.com/show_bug.cgi?id=697450
is fixed in the version 9.21 upstream sources.
- CVE-2017-5951.patch fixes
null pointer dereference in ref_stack_index() that is
related to mem_get_bits_rectangle() in base/gdevmem.c
https://bugs.ghostscript.com/show_bug.cgi?id=697548
(bsc#1032114)
-------------------------------------------------------------------
Mon Apr 10 14:06:09 CEST 2017 - jsmeix@suse.de
- Version upgrade to 9.21.
For details see the News.htm and History9.htm files.
Highlights in this release include:
* pdfwrite now preserves annotations from
input PDFs (where possible).
* The GhostXPS interpreter now provides the pdfwrite device
with the data it requires to emit a ToUnicode CMap: thus
allowing fully searchable PDFs to be created from XPS
input (in the vast majority of cases).
* Ghostscript now allows the default color space
for PDF transparency blends.
* The Ghostscript/GhostPDL configure script now has much
better/fuller support for cross compiling.
* The tiffscaled and tiffscaled4 devices can now
use ETS (Even Tone Screening)
* The toolbin/pdf_info.ps utility can now emit
the PDF XML metadata.
* Ghostscript has a new scan converter available
(currently optional, but will become the default in a near
future release). It can be enabled by using the command line
option: '-dSCANCONVERTERTYPE=2'. This new implementation
provides vastly improved performance with large and complex
paths.
* The usual round of bug fixes, compatibility changes,
and incremental improvements.
Incompatible changes:
* The planned device API tidy (still!) did not happen for
this release, due to time pressures, but we still intend
to undertake the following: We plan to somewhat tidy up
the device API. We intend to remove deprecated device
procs (methods/function pointers) and change the device API
so every device proc takes a graphics state parameter
(rather than the current scheme where only a very few procs
take an imager state parameter). This should serve as notice
to anyone maintaining a Ghostscript device outside the
canonical source tree that you may (probably will) need to
update your device(s) when these changes happen. Devices using
only the non-deprecated procs should be trivial to update.
- CVE-2016-7976.patch and CVE-2016-7977.patch and
CVE-2016-7978.patch and CVE-2016-7979.patch and
CVE-2016-8602.patch are no longer needed because
those issues are fixed in the upstream sources.
- 0001-mkromfs-make-build-reproducible-use-buildtime-from-S.patch
and
0002-mkromfs-sort-gp_enumerate_files-output-for-determini.patch
are no longer needed because both are included
in the upstream sources, see the upstream issue
https://bugs.ghostscript.com/show_bug.cgi?id=697484
- Again use the zlib sources from Ghostscript upstream
and disable remove-zlib-h-dependency.patch because
Ghostscript 9.21 does no longer build this way,
cf. the entry below dated "Wed Nov 18 11:46:58 UTC 2015"
-------------------------------------------------------------------
Thu Jan 12 17:13:58 UTC 2017 - stefan.bruens@rwth-aachen.de

View File

@ -38,7 +38,7 @@ Url: http://www.ghostscript.com/
# so that we keep additionally the previous version number to upgrade from the previous version:
#Version: 9.19pre20rc1
# Normal version for Ghostscript releases is the upstream version:
Version: 9.20
Version: 9.21
Release: 0
# tarball_version is used below to specify the directory via "setup -n":
# Special tarball_version needed for Ghostscript release candidates e.g. "define tarball_version 9.15rc1".
@ -49,7 +49,7 @@ Release: 0
# Separated built_version needed in case of Ghostscript release candidates e.g. "define built_version 9.15".
# For Ghostscript releases built_version and version are the same (i.e. the upstream version):
%define built_version %{version}
#%define built_version 9.20
#define built_version 9.20
# Source0...Source9 is for sources from upstream:
# Special URLs for Ghostscript release candidates:
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920rc1/ghostscript-9.20rc1.tar.gz
@ -57,37 +57,13 @@ Release: 0
# wget -O ghostscript-9.20rc1.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920rc1/ghostscript-9.20rc1.tar.gz
#Source0: ghostscript-%{tarball_version}.tar.gz
# Normal URLs for Ghostscript releases:
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/ghostscript-9.20.tar.gz
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/ghostscript-9.21.tar.gz
# How to download it:
# wget -O ghostscript-9.20.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/ghostscript-9.20.tar.gz
# URL for MD5 checksums: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/MD5SUMS
# MD5 checksum for Source0: 93c5987cd3ab341108be1ebbaadc24fe
# wget -O ghostscript-9.21.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/ghostscript-9.21.tar.gz
# URL for MD5 checksums: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/MD5SUMS
# MD5 checksum for Source0: 5f213281761d2750fcf27476c404d17f
Source0: ghostscript-%{version}.tar.gz
# Patch0...Patch9 is for patches from upstream:
# Patch1 CVE-2016-7976.patch fixes that
# various userparams allow %pipe% in paths, allowing remote shell command execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697178
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch1: CVE-2016-7976.patch
# Patch2 CVE-2016-7977.patch fixes that
# .libfile doesn't check PermitFileReading array, allowing remote file disclosure
# see http://bugs.ghostscript.com/show_bug.cgi?id=697169
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch2: CVE-2016-7977.patch
# Patch3 CVE-2016-7978.patch fixes that
# reference leak in .setdevice allows use-after-free and remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697179
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch3: CVE-2016-7978.patch
# Patch4 CVE-2016-7979.patch fixes that
# type confusion in .initialize_dsc_parser allows remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697190
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch4: CVE-2016-7979.patch
# Patch5 CVE-2016-8602.patch fixes a NULL dereference in .sethalftone5
# see http://bugs.ghostscript.com/show_bug.cgi?id=697203
# and https://bugzilla.suse.com/show_bug.cgi?id=1004237
Patch5: CVE-2016-8602.patch
# Source10...Source99 is for sources from SUSE which are intended for upstream:
# Patch10...Patch99 is for patches from SUSE which are intended for upstream:
# Patch11 ppc64le-support.patch is a remainder of the previous patch
@ -96,16 +72,17 @@ Patch5: CVE-2016-8602.patch
# but the hunk for LCMS2 (lcms2/include/lcms2.h) is still needed
# see http://bugs.ghostscript.com/show_bug.cgi?id=695544
Patch11: ppc64le-support.patch
# Patch12 adds a reproducible timestamp to the mkromfs output, using the
# SOURCE_DATE_EPOCH environment variable
Patch12: 0001-mkromfs-make-build-reproducible-use-buildtime-from-S.patch
# Patch13 sorts the ROM contents by name for deterministic contents
Patch13: 0002-mkromfs-sort-gp_enumerate_files-output-for-determini.patch
# Source100...Source999 is for sources from SUSE which are not intended for upstream:
# Patch100...Patch999 is for patches from SUSE which are not intended for upstream:
# Patch100 remove-zlib-h-dependency.patch removes dependency on zlib/zlib.h
# in makefiles as we do not use the zlib sources from the Ghostscript upstream tarball:
Patch100: remove-zlib-h-dependency.patch
# Patch101 CVE-2017-5951.patch fixes
# null pointer dereference in ref_stack_index() that is
# related to mem_get_bits_rectangle() in base/gdevmem.c
# https://bugs.ghostscript.com/show_bug.cgi?id=697548
# (bsc#1032114)
Patch101: CVE-2017-5951.patch
# RPM dependencies:
Conflicts: ghostscript
Conflicts: ghostscript-x11
@ -174,50 +151,28 @@ This package contains the development files for Minimal Ghostscript.
# Be quiet when unpacking and
# use a directory name matching Source0 to make it work also for ghostscript-mini:
%setup -q -n ghostscript-%{tarball_version}
# Patch1 CVE-2016-7976.patch fixes that
# various userparams allow %pipe% in paths, allowing remote shell command execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697178
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch1 -p1 -b CVE-2016-7976.orig
# Patch2 CVE-2016-7977.patch fixes that
# .libfile doesn't check PermitFileReading array, allowing remote file disclosure
# see http://bugs.ghostscript.com/show_bug.cgi?id=697169
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch2 -p1 -b CVE-2016-7977.orig
# Patch3 CVE-2016-7978.patch fixes that
# reference leak in .setdevice allows use-after-free and remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697179
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch3 -p1 -b CVE-2016-7978.orig
# Patch4 CVE-2016-7979.patch fixes that
# type confusion in .initialize_dsc_parser allows remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697190
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch4 -p1 -b CVE-2016-7979.orig
# Patch5 CVE-2016-8602.patch fixes a NULL dereference in .sethalftone5
# see http://bugs.ghostscript.com/show_bug.cgi?id=697203
# and https://bugzilla.suse.com/show_bug.cgi?id=1004237
%patch5 -p1 -b CVE-2016-8602.orig
# Patch11 ppc64le-support.patch is a remainder of the previous patch
# now the hunk for LCMS (lcms/include/lcms.h) is removed
# because LCMS 1.x is removed since Ghostscript 9.16
# but the hunk for LCMS2 (lcms2/include/lcms2.h) is still needed
# see http://bugs.ghostscript.com/show_bug.cgi?id=695544
%patch11 -p1 -b ppc64le-support.orig
# Patch12 adds a reproducible timestamp to the mkromfs output, using the
# SOURCE_DATE_EPOCH environment variable
%patch12 -p1 -b mkromfs-buildtime.orig
# Patch13 sorts the ROM contents by name for deterministic contents
%patch13 -p1 -b mkromfs-sort-contents.orig
# Patch100 remove-zlib-h-dependency.patch removes dependency on zlib/zlib.h
# in makefiles as we do not use the zlib sources from the Ghostscript upstream tarball:
%patch100 -p1 -b remove-zlib-h-dependency.orig
#patch100 -p1 -b remove-zlib-h-dependency.orig
# Do not use the freetype jpeg libpng tiff zlib sources from the Ghostscript upstream tarball
# because we prefer to use for long-established standard libraries the ones from SUSE
# in particular to automatically get SUSE security updates for standard libraries.
# In contrast we use e.g. lcms2 from the Ghostscript upstream tarball because this one
# is specially modified to work with Ghostscript so that we cannot use lcms2 from SUSE:
rm -rf freetype jpeg libpng tiff zlib
#rm -rf freetype jpeg libpng tiff zlib
rm -rf freetype jpeg libpng tiff
# Patch101 CVE-2017-5951.patch fixes
# null pointer dereference in ref_stack_index() that is
# related to mem_get_bits_rectangle() in base/gdevmem.c
# https://bugs.ghostscript.com/show_bug.cgi?id=697548
# (bsc#1032114)
%patch101 -b CVE-2017-5951.orig
%build
# Derive build timestamp from latest changelog entry

View File

@ -1,3 +1,90 @@
-------------------------------------------------------------------
Wed Apr 12 11:12:27 CEST 2017 - jsmeix@suse.de
- CVE-2016-10317 (bsc#1032230)
heap buffer overflow in fill_threshhold_buffer()
is not yet fixed because there is no fix available at
https://bugs.ghostscript.com/show_bug.cgi?id=697459
- CVE-2016-10219 (bsc#1032138)
divide by zero in intersect()
https://bugs.ghostscript.com/show_bug.cgi?id=697453
is fixed in the version 9.21 upstream sources
- CVE-2016-10218 (bsc#1032135)
null pointer dereference in pdf14_pop_transparency_group()
https://bugs.ghostscript.com/show_bug.cgi?id=697444
is fixed in the version 9.21 upstream sources.
- CVE-2016-10217 (bsc#1032130)
use-after-free in pdf14_cleanup_parent_color_profiles()
that is related to pdf14_open() in base/gdevp14.c
https://bugs.ghostscript.com/show_bug.cgi?id=697456
is fixed in the version 9.21 upstream sources.
- CVE-2016-10220 (bsc#1032120)
null pointer dereference in gx_device_finalize() that is
related to gs_makewordimagedevice() in base/gsdevmem.c
https://bugs.ghostscript.com/show_bug.cgi?id=697450
is fixed in the version 9.21 upstream sources.
- CVE-2017-5951.patch fixes
null pointer dereference in ref_stack_index() that is
related to mem_get_bits_rectangle() in base/gdevmem.c
https://bugs.ghostscript.com/show_bug.cgi?id=697548
(bsc#1032114)
-------------------------------------------------------------------
Mon Apr 10 14:06:09 CEST 2017 - jsmeix@suse.de
- Version upgrade to 9.21.
For details see the News.htm and History9.htm files.
Highlights in this release include:
* pdfwrite now preserves annotations from
input PDFs (where possible).
* The GhostXPS interpreter now provides the pdfwrite device
with the data it requires to emit a ToUnicode CMap: thus
allowing fully searchable PDFs to be created from XPS
input (in the vast majority of cases).
* Ghostscript now allows the default color space
for PDF transparency blends.
* The Ghostscript/GhostPDL configure script now has much
better/fuller support for cross compiling.
* The tiffscaled and tiffscaled4 devices can now
use ETS (Even Tone Screening)
* The toolbin/pdf_info.ps utility can now emit
the PDF XML metadata.
* Ghostscript has a new scan converter available
(currently optional, but will become the default in a near
future release). It can be enabled by using the command line
option: '-dSCANCONVERTERTYPE=2'. This new implementation
provides vastly improved performance with large and complex
paths.
* The usual round of bug fixes, compatibility changes,
and incremental improvements.
Incompatible changes:
* The planned device API tidy (still!) did not happen for
this release, due to time pressures, but we still intend
to undertake the following: We plan to somewhat tidy up
the device API. We intend to remove deprecated device
procs (methods/function pointers) and change the device API
so every device proc takes a graphics state parameter
(rather than the current scheme where only a very few procs
take an imager state parameter). This should serve as notice
to anyone maintaining a Ghostscript device outside the
canonical source tree that you may (probably will) need to
update your device(s) when these changes happen. Devices using
only the non-deprecated procs should be trivial to update.
- CVE-2016-7976.patch and CVE-2016-7977.patch and
CVE-2016-7978.patch and CVE-2016-7979.patch and
CVE-2016-8602.patch are no longer needed because
those issues are fixed in the upstream sources.
- 0001-mkromfs-make-build-reproducible-use-buildtime-from-S.patch
and
0002-mkromfs-sort-gp_enumerate_files-output-for-determini.patch
are no longer needed because both are included
in the upstream sources, see the upstream issue
https://bugs.ghostscript.com/show_bug.cgi?id=697484
- Again use the zlib sources from Ghostscript upstream
and disable remove-zlib-h-dependency.patch because
Ghostscript 9.21 does no longer build this way,
cf. the entry below dated "Wed Nov 18 11:46:58 UTC 2015"
-------------------------------------------------------------------
Thu Jan 12 17:13:58 UTC 2017 - stefan.bruens@rwth-aachen.de

View File

@ -58,7 +58,7 @@ Url: http://www.ghostscript.com/
# so that we keep additionally the previous version number to upgrade from the previous version:
#Version: 9.19pre20rc1
# Normal version for Ghostscript releases is the upstream version:
Version: 9.20
Version: 9.21
Release: 0
# tarball_version is used below to specify the directory via "setup -n":
# Special tarball_version needed for Ghostscript release candidates e.g. "define tarball_version 9.15rc1".
@ -69,7 +69,7 @@ Release: 0
# Separated built_version needed in case of Ghostscript release candidates e.g. "define built_version 9.15".
# For Ghostscript releases built_version and version are the same (i.e. the upstream version):
%define built_version %{version}
#%define built_version 9.20
#define built_version 9.20
# Source0...Source9 is for sources from upstream:
# Special URLs for Ghostscript release candidates:
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920rc1/ghostscript-9.20rc1.tar.gz
@ -77,37 +77,13 @@ Release: 0
# wget -O ghostscript-9.20rc1.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920rc1/ghostscript-9.20rc1.tar.gz
#Source0: ghostscript-%{tarball_version}.tar.gz
# Normal URLs for Ghostscript releases:
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/ghostscript-9.20.tar.gz
# URL for Source0: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/ghostscript-9.21.tar.gz
# How to download it:
# wget -O ghostscript-9.20.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/ghostscript-9.20.tar.gz
# URL for MD5 checksums: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs920/MD5SUMS
# MD5 checksum for Source0: 93c5987cd3ab341108be1ebbaadc24fe
# wget -O ghostscript-9.21.tar.gz https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/ghostscript-9.21.tar.gz
# URL for MD5 checksums: https://github.com/ArtifexSoftware/ghostpdl-downloads/releases/download/gs921/MD5SUMS
# MD5 checksum for Source0: 5f213281761d2750fcf27476c404d17f
Source0: ghostscript-%{version}.tar.gz
# Patch0...Patch9 is for patches from upstream:
# Patch1 CVE-2016-7976.patch fixes that
# various userparams allow %pipe% in paths, allowing remote shell command execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697178
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch1: CVE-2016-7976.patch
# Patch2 CVE-2016-7977.patch fixes that
# .libfile doesn't check PermitFileReading array, allowing remote file disclosure
# see http://bugs.ghostscript.com/show_bug.cgi?id=697169
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch2: CVE-2016-7977.patch
# Patch3 CVE-2016-7978.patch fixes that
# reference leak in .setdevice allows use-after-free and remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697179
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch3: CVE-2016-7978.patch
# Patch4 CVE-2016-7979.patch fixes that
# type confusion in .initialize_dsc_parser allows remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697190
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
Patch4: CVE-2016-7979.patch
# Patch5 CVE-2016-8602.patch fixes a NULL dereference in .sethalftone5
# see http://bugs.ghostscript.com/show_bug.cgi?id=697203
# and https://bugzilla.suse.com/show_bug.cgi?id=1004237
Patch5: CVE-2016-8602.patch
# Source10...Source99 is for sources from SUSE which are intended for upstream:
# Patch10...Patch99 is for patches from SUSE which are intended for upstream:
# Patch11 ppc64le-support.patch is a remainder of the previous patch
@ -116,16 +92,17 @@ Patch5: CVE-2016-8602.patch
# but the hunk for LCMS2 (lcms2/include/lcms2.h) is still needed
# see http://bugs.ghostscript.com/show_bug.cgi?id=695544
Patch11: ppc64le-support.patch
# Patch12 adds a reproducible timestamp to the mkromfs output, using the
# SOURCE_DATE_EPOCH environment variable
Patch12: 0001-mkromfs-make-build-reproducible-use-buildtime-from-S.patch
# Patch13 sorts the ROM contents by name for deterministic contents
Patch13: 0002-mkromfs-sort-gp_enumerate_files-output-for-determini.patch
# Source100...Source999 is for sources from SUSE which are not intended for upstream:
# Patch100...Patch999 is for patches from SUSE which are not intended for upstream:
# Patch100 remove-zlib-h-dependency.patch removes dependency on zlib/zlib.h
# in makefiles as we do not use the zlib sources from the Ghostscript upstream tarball:
Patch100: remove-zlib-h-dependency.patch
# Patch101 CVE-2017-5951.patch fixes
# null pointer dereference in ref_stack_index() that is
# related to mem_get_bits_rectangle() in base/gdevmem.c
# https://bugs.ghostscript.com/show_bug.cgi?id=697548
# (bsc#1032114)
Patch101: CVE-2017-5951.patch
# RPM dependencies:
# Additional RPM Provides of the ghostscript-library packages in openSUSE 11.4 from
# "rpm -q --provides ghostscript-library" and "rpm -q --provides ghostscript-x11":
@ -310,50 +287,28 @@ This package contains the development files for Ghostscript.
# Be quiet when unpacking and
# use a directory name matching Source0 to make it work also for ghostscript-mini:
%setup -q -n ghostscript-%{tarball_version}
# Patch1 CVE-2016-7976.patch fixes that
# various userparams allow %pipe% in paths, allowing remote shell command execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697178
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch1 -p1 -b CVE-2016-7976.orig
# Patch2 CVE-2016-7977.patch fixes that
# .libfile doesn't check PermitFileReading array, allowing remote file disclosure
# see http://bugs.ghostscript.com/show_bug.cgi?id=697169
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch2 -p1 -b CVE-2016-7977.orig
# Patch3 CVE-2016-7978.patch fixes that
# reference leak in .setdevice allows use-after-free and remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697179
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch3 -p1 -b CVE-2016-7978.orig
# Patch4 CVE-2016-7979.patch fixes that
# type confusion in .initialize_dsc_parser allows remote code execution
# see http://bugs.ghostscript.com/show_bug.cgi?id=697190
# and https://bugzilla.suse.com/show_bug.cgi?id=1001951
%patch4 -p1 -b CVE-2016-7979.orig
# Patch5 CVE-2016-8602.patch fixes a NULL dereference in .sethalftone5
# see http://bugs.ghostscript.com/show_bug.cgi?id=697203
# and https://bugzilla.suse.com/show_bug.cgi?id=1004237
%patch5 -p1 -b CVE-2016-8602.orig
# Patch11 ppc64le-support.patch is a remainder of the previous patch
# now the hunk for LCMS (lcms/include/lcms.h) is removed
# because LCMS 1.x is removed since Ghostscript 9.16
# but the hunk for LCMS2 (lcms2/include/lcms2.h) is still needed
# see http://bugs.ghostscript.com/show_bug.cgi?id=695544
%patch11 -p1 -b ppc64le-support.orig
# Patch12 adds a reproducible timestamp to the mkromfs output, using the
# SOURCE_DATE_EPOCH environment variable
%patch12 -p1 -b mkromfs-buildtime.orig
# Patch13 sorts the ROM contents by name for deterministic contents
%patch13 -p1 -b mkromfs-sort-contents.orig
# Patch100 remove-zlib-h-dependency.patch removes dependency on zlib/zlib.h
# in makefiles as we do not use the zlib sources from the Ghostscript upstream tarball:
%patch100 -p1 -b remove-zlib-h-dependency.orig
#patch100 -p1 -b remove-zlib-h-dependency.orig
# Do not use the freetype jpeg libpng tiff zlib sources from the Ghostscript upstream tarball
# because we prefer to use for long-established standard libraries the ones from SUSE
# in particular to automatically get SUSE security updates for standard libraries.
# In contrast we use e.g. lcms2 from the Ghostscript upstream tarball because this one
# is specially modified to work with Ghostscript so that we cannot use lcms2 from SUSE:
rm -rf freetype jpeg libpng tiff zlib
#rm -rf freetype jpeg libpng tiff zlib
rm -rf freetype jpeg libpng tiff
# Patch101 CVE-2017-5951.patch fixes
# null pointer dereference in ref_stack_index() that is
# related to mem_get_bits_rectangle() in base/gdevmem.c
# https://bugs.ghostscript.com/show_bug.cgi?id=697548
# (bsc#1032114)
%patch101 -b CVE-2017-5951.orig
%build
# Derive build timestamp from latest changelog entry