| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
PSSRC files are now in 'gs/psi'.
GLSRC files are now in 'gs/base'.
This is to facilitate build modularization and merging in the ghostpdl
tree.
NOTE: msvc32.mak is now in psi, not src.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@9048 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Bug 689995 "segfault with 09-34.PS"
1. (the bug fix) A recent patch doesn't account patterns with no tile,
which represent an empty pattern.
This change inserts checks for empty tiles
and process them with no optimization (redundantly),
because they are small and are not cached.
See comments in code.
2. pcls->band_complexity.uses_color accummulated a pattern address bits instead
a real data. This bug was introduced when serializing colors to clist.
nevertheless this bug has no visible effect because 'pure' is overlayed with
a tile address, which can't be equal to 0 or 0xffffff.
So formally this part of the change is algorithmically equivalent,
but it is important for redability.
Note that when serializing a pattern to clist, 'uses_color' is not properly
computed and needs further development. see new bug 689997.
Minor change : removed the second call to gs_dc_get_pattern_id for a faster code.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8952 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Bug 689865 "clist stores redundant patterns".
The change is pretty simple because the clist reader already implememnts
a pattern cache. We just add pattern id to gx_clist_state
and provide a short format (just the id) when writing and reading
a pattern color to/from clist at second time. A new function gs_dc_get_pattern_id
allows to access the pattern id through closures of color elements.
Note it remembers up to 2 pattern ids for a band
(one for the specific band, and another for "all bands").
A better way would be to remember all patterns cached for a band,
but it would need to save all cached ids individually for each band while clist writing.
We don't want to enlarge the clist writer storage now.
Minor change : Some dependencies of gxclpath.c were missed in makefile.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8940 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Bug 689911 Shifting raster data, Windows MSVC 8 non-debug build.
In function cmd_put_path the variable 'prev' was wrongly localized
in the loop body when need to save/pass data from the previous cycle.
It worked fine with debug build, because the compiler
always commits variables to RAM for debug purpose.
With release build the variable 'prev' never changes
because formally speaking the program never uses its value
after it is initialized.
The bug persists since the repository was created on March 9 2000.
The old behavior may be indeterministic.
With MSVC8 the effect is occasionaly visible when a degenerate curve
{x y x y x y x y} is written into clist.
See the simplified test case attached to the bug report.
EXPECTED DIFFERENCES :
None, but the old behavior may be indeterminiztic.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8815 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
interpretation (continued 4).
DETAILS :
The old code assumes that a pattern always fits into the clist writer buffer.
However ppmraw -r400 176-01.ps violates this assumption.
This patch implements a subdivision of a pattern into slices,
each of which is smaller than the buffer size.
1. The new function cmd_get_buffer_space retrives maximal available space
in the clist writer buffer (gxclutil.c).
2. cmd_put_drawing_color now implements a cycle for slicing big patterns.
Each slice writes to clist as a separate clist instruction (gxclpath.c).
3. gx_dc_pattern_write_raster drops the constraint that the pattern
fits into a single buffer (gsptype1.c).
4. The clist interpreter is enhanced with composing a pattern from slices
(gxclrast.c).
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8770 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
interpretation.
DETAILS :
Bug 689579 "Ghostscript generates a very, very large temporary file".
The old code converts pattern fills into lots of constant color rectangles while writing to clist.
The new code writes pattern data to clist and apply pattern during the clist playback.
So here we complete the change started in revision 8655.
New functions gx_dc_pattern_write_raster, gx_dc_pattern_read_raster
implement color serialization for raster patterns.
The old function gx_dc_pattern_write got minor changes about buffer localization.
Minor chabge (gxclimag.c, gxclpath.c, gxclrect.c) : Propagated error codes.
Also fixed a bug occasionally found in gxclrast.c .
It caused a crash in some unusual conditions.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8712 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
methods for big color data.
DETAILS :
This is a preparation for fixing
the bug 689440 "PostScript file generates infinitely large temp file".
The old serialization methods assume that color data can fit into a small buffer.
Now we want to drop this constraint to allow to write big patterns into clist.
See comment in code for details.
This change is algorithmically equivalent.
This patch adds a new argument to color serialization methods.
For a while the new actual parameter is always 0,
and the (untrivial) methods check for it.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8632 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This really crops transparency compositor commands with bands
that are covered by the transparency bbox.
Actually the statement above has one exception. When a transparency bbox
covers more that 2/3 of all bands, we still write the commands to
the "all bands" list for a shorter clist file. We noticed that
the clist playback time strongly depends on the file size.
1. The new method gs_composite_type_procs_t::get_cropping
replies what bands are covered by the compositor.
When it gives a non-trivial result, clist_create_compositor
writes the compositor command to those bands,
which are covered with the cropping (Exception :
if too many bands are covered, it writes the command to
the "for all bands" list).
2. Since some bands skip some compositor commands,
while clist reading the association of masks to groups
becomes more complex. Before this patch each group
was associates with the mask, which was created immediately
before it at same transparency stack level.
After this patch some groups may be skipped, so that
clist reader may recieve extra masks, because masks are
still use the container's group bbox. For skipping such extra masks
we provide a numeric identifier of a mask, which is named mask_id.
Here is how it works :
2.1. All masks recieve serial numbers (mask_id) while clist writing.
2.2. Each mask command is written with its mask_id.
2.3. Each group command is writen with mask_id,
which belongs to the mask that is assocuated with the group.
2.4. While clist playback pdf14_pop_transparency_group checks whether
the mask amd the group have same mask_id. If not, the mask is extra
and it must be droped.
2.5. Note the old assiciation logic is still working,
and it may skip some masks as well.
2.6. If the rasterization happens with no clist, mask_id is always zero.
In this case the old logics is only working.
Maybe we'll provide non-zero values someday.
2.7. See comments in code for more details.
3. The method gs_composite_type_procs_t::clist_compositor_write_update
is now some generalized. Before this patch it only updates the clist writer state
with the compositor features. With this patch it also adds
some information from the clist writer state to the compositor parameters.
Particularly it provides mask_id to the compositor command.
Thus now it updates *both* clist writer and the compositor parameters
to comply each another. For a while, in order to simplify the patch,
the compositor argument is still 'const'. Will need to fix someday.
4. Now we need to save mask_id in a stack that reflects the
transparency nesting. The cropping stack, which was introduced
in the last patch, now works for that. To provide that
we add mask_id to clist_writer_cropping_buffer_s,
and use clist_writer_push_no_cropping to save amsk_id
over inner groups.
5. state_update(ctm) in c_pdf14trans_clist_write_update is replaced
with code, which updates CTM from the transparency compositor command.
The old code is not fully correct, because it takes CTM from the imager state,
rather we havn't got prcatical cases when it gives a wrong result.
6. drop_compositor_queue now calls adjust_ctm for compositors being dropped.
The old code is incorrect because it can miss a CTM update when
a compositor sets some CTM and later a non-compositor object
sets same CTM. In this case the clist writer skips the second CTM on writing,
and skipping the first in the clist reader will miss the CTM completely.
We haven't got practical cases for that.
7. Improved some debug printing.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8585 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This is a preparation to the next step.
The patch introduces a transparency stack for clist writer,
which saves cropping when entering a transparency group,
and restores when exiting it. The cropping is being narrowed
with transparency group's bounding box.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8584 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This is a preparation to the next step.
In general, the behavior doesn't change,
but implementation details differ.
The old code maitains a flag cropping_by_y in the clist writer,
which triggers a cropping of objects by the Y axis.
The old code isn't perfect, because
it first crops by [cdev->cropping_min, cdev->cropping_max],
and then by [0, cdev->height].
In the new code we unite both croppings into a single one.
For doing that we remove the flag, and provide
the right interval [cdev->cropping_min, cdev->cropping_max]
in any case. Initially it is [0, cdev->height},
and it is narrowed when a shading bbox is set.
To restore the initial cropping after the shading
is complete, we maintain 2 new fields
save_croping_min and save_cropping_max.
So now the cropping mechanizm does not depend
on shading, and clients may set a croping when they need.
We'll use this feature in the next patch.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8581 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
The clist writer writes the 'create compositor' operation to all bands,
including ones that are not covered by a transparency.
It does so because this operation changes the number of color components.
When rendering a specific band, it frequently happens that a compositor
is created and then immediately destroyed. Such thing happens outside
the transparency bbox, and within the transparency bbox if
the band has no transparent objects. When compositor is created,
it allocates a big raster buffer and cleans it.
Due to that we observe a significant CPU time expence
when running the test case of the bug 689155.
This patch is a preparation for further improvements.
This patch includes CTM into pdf14 compositor params
(see changes to c_pdf14trans_read, c_pdf14trans_write, clist_create_compositor).
The result should be same as the old code, but the algorithm is different.
The old code wrote CTM as a separate command before cmd_opv_ext_create_compositor.
The new code does not, and in many cases compositor commads
immediately follow each another.
The last fact changes the behavior of the compositor queue :
before this patch it consisted of 1 element maximum,
but after it the queue becomes longer (up to 10 elements
with SoftMaskGroup.pdf). A bug is fixed in the queue logic
in gxclrast.c ln 1355.
The queue is still immediately executed
when a non-compositor command appears in the input stream.
An annihilation of neighbour idle compositors will be a next step.
Minor changes:
- added a new method adjust_ctm to gs_composite_type_procs_t.
- added type checks with composite_*_proc macros to all compositor types.
- dependencies were broken for gdevp14.c in lib.mak .
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8426 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Debugged with the test case of the bug 689155.
A long ago the graphics library converted the sfill clipping into a path.
It was improved with revision 8017 with introducing
a device response for pattern_manage__shfill_doesnt_need_path.
However the revision 8017 appears incomplete :
it missed the related change to the clist writer device
and to pdf14 device. This patch fixes it.
This patch causes a progression for the problem
known as bug 689338 "Raster depends on bands".
With banded rendering some shadings are now painted some wider,
and the raster appears equal to one from unbanded rendering.
It happens because the old code doesn't convert shfill clipping
into a path when no banding, and does convert when banding.
The new code does not convert in both cases.
EXPECTED DIFFERENCES :
Progression : a banded raster is closer to unbanded :
300dpi banded :
gradmesh.ai
S2_Digitalproof-Forum_x3k.pdf
SmoothShading.pdf
72dpi banded :
shading_prob_800.ps
Bug688631.pdf
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8416 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS:
The compilers we support have supported static for some time now, so
idiomatic, and removes one of the obstacles to compiling Ghostscript
as C++.
The patch also removes support for compilation with -dNOPRIVATE, which
is not longer necessary with modern debuggers.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8250 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
linear color triangles (continued 8).
DETAILS :
Bug 689436 "pattern rendering regression with revision 8095".
Do not write tile phase to clist before trying to write the color.
If the latter fails, the phase may become wrong due to
overloading cmd_opv_set_tile_phase with tile phase and device halftone phase.
With the test case the old code, when writing a trapezoid to clist,
first writes halftone phase, then fails to write the pattern color
and falls back to gx_default_fill_trapezoid.
The latter writes strip_copy_rop with it own phases.
But the clist reader expects strip_copy_rop phase to be written
*after* the tile header, so the reader misses the phase sychronizm.
The fix simply delays writing the phase after the color is written.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8240 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
in revision 8055 (see 8055 log for details).
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8182 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
linear color triangles (continued 8).
DETAILS :
This fixes (some) gcc warnings in the recently added code (part 2).
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8134 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
and linear color triangles.
DETAILS :
This fixes the clist expansion problem for shadings.
1. New functions clist_fill_linear_color_trapezoid, clist_fill_linear_color_triangle
provide the writing of shading parts into clist
without a further decomposition into rectangles.
2. gxclrast.c is enhanced with reading the data from 1.
3. gxfill.c needs a special interaction with the clipper device
for passing the clipping path as a high level object.
The new request pattern_manage__handles_clip_path
checks whether the target device can handle a clipping path.
The clist writer can, so the graphics library first sends the
clipping path, and then decomposes the shading into
trapezoids without installing a clipper device.
Doing so because clipper device can't handle trapezoids.
See comments in gxclpath.c, gxfill.c .
4. The clist reader installs a clipper device when processes
trapezoids and linear color triangles.
The new flag clipper_dev_open controls that.
5. A new flag gx_device_clist_writer_s::cropping_by_path
controls an additional cropping of shading components,
which fall outside the shading path's bands.
See comment in gxclpath.c .
6. R_fill_rect_with_const_color is rewritten with linear color functions,
because the old implementation creates a rectangle, which is not clipped with
4,5. It gives an invisible slowdown for unbanded rasters.
There are few things, which need further improvements :
1. A better compression for frac31 values in clist.
Such values usually have many zeros in ending bits.
2. Crop bands with ybot, ytop, fa->ystart, fa->yend in
clist_write_fill_trapezoid.
3. The clist reader must handle cases when linear color functions return 0.
It must perform a decomposition of areas with calling
appropriate parts of gxshade6.c . The current code returns error,
which does not happen in practice.
EXPECTED DIFFERENCES :
There is a problem with pkmraw -r300 Bug689189.pdf
which has to be fixed separately.
This patch causes a minor raster difference with shadings :
with a banded rendering the shading area becomes some smaller
than the old banded rendering. It needs a further analyzis.
Here is a list of files with this problem :
"442-01.ps"
"446-01-fixed.ps"
"464-01-fixed.ps"
"470-01.ps"
"478-01.ps"
"483-01.ps"
"483-05-fixed.ps"
"Altona-Testsuite_p2_S_x3.pdf"
"Altona_Visual_bb_1v1_x3.pdf"
"Altona_Visual_sb_1v1_x3.pdf"
"BEST8-99-Path.fh7.pdf"
"Bug689189.pdf"
"chilis_black.pdf"
"chilis_red.pdf"
"Clarke Tate Manns Chinese.ai"
"gradmesh.ai"
"Openhuis_pdf_zw.pdf"
"S2_Digitalproof-Forum_x3k.pdf"
"self-intersect2.ps"
"shading_prob_800.ps"
"SmoothShading.pdf"
"STEUER-RollingMesh 1(linear).ai"
"STEUER-RollingMesh 2(radial).ai"
"STEUER-RollingMesh 3(Final).ai"
"Testform.v1.0.2.pdf"
This list may be related to bug 689338.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8118 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Bug 689317 "A wrong rendering of a path", Comment #3.
The revision 2948 change to gxclpath.c is not properly documented.
This patch stores assumptions, which are necessary for the
change to be correct.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8109 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This change is algorithmically equivalent.
The purpose is to eliminate gxcldev.h macros
to simplify the debugging of the clist writer methods.
This 6th step simplifies the flow control
with removing unuseful goto and labels.
Also it updates the documentation
in gxcldev.h .
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8090 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This change is algorithmically equivalent.
The purpose is to eliminate gxcldev.h macros
to simplify the debugging of the clist writer methods.
This 5th step eliminates FOR_RECTS and END_RECTS macros.
So now all control flow statements are outside macros.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8087 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This change is algorithmically equivalent.
The purpose is to eliminate gxcldev.h macros
to simplify the debugging of the clist writer methods.
This 3nd step collects local variables
of a rectangle enumerator in a single structure
cmd_rects_enum_s.
In the old code the macro FOR_RECTS modifies
function arguments 'y' and 'height'.
It doesn't comply with C-style.htm .
For fixing that this patch renames
function arguments x,y,width,height with
rx,ry,rwidth,rheight.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8085 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
This change is algorithmically equivalent.
The purpose is to eliminate gxcldev.h macros
to simplify the debugging of the clist writer methods.
This first step removes simple macros and optimized macro duplicates.
Removing others to be done in the next step.
Some compilers may claim unreferenced label error_in_rect.
To be fixed in the next step.
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8078 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
to Artifex Software, Inc.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@8022 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS:
-r600 -sDEVICE=wtsimdi does a performance optimized clist reader that either directly halftones
or uses an intermediate rgb buffer to do non-trivial ROPS correctly then post halftones
on a per band basis.
Uses icc color profiles to convert to device color.
Uses wts halftone algorithm.
The device reads in a link icc profile from "link.icc", and WTS
halftone files from wts_plane_[0123]. It generates a ppm file that
can be used to visualize the halftoned file (it is prepared from the
CMYK bits in the procedure write_pkmraw_row in gdevwts.c - if an
actual device is available, that routine should be modified to produce
data for the device).
This device _does_ switch on a per-band basis between continuous
tone pathways (true branch of the "if" statement in
wtsimdi_create_buf_device) and halftone, based on whether there are
rops. That is the major point of the work.
This is not yet production ready, but is useful for performance and integration work.
Known issues:
1. Banding only, low res non-banding will crash,
the device is intended for printer resolution halftoning, at 600dpi
The halftones are designed for 600dpi.
2. Color profiles and halftone screens are not in romfs,
These need to be copied into the same object/binary directory
that the application is in for now.
./pspcl6 and ./wts_* and ./*.icc in the same directory.
Failure to find these files will crash.
3. Halftone phase not set correctly in halftone band.
4. copy_mono doesn't optimize all-0 and all-1 colors as much as it
should - it can avoid processing the screen at that point.
5. halftone mode implemented only for rgb color model, not CMYK.
6. size of band_colors_used[] array is statically determined
EXPECTED DIFFERENCES :
None, unless -sDEVICE=wtsimdi is used.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@7819 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
discipline for color spaces. Previously, color spaces were allocated on
stack, on the heap, as sub-structures within other color space structures,
and were very commonly copied. Reference counting was used for the "params"
block of some, but not all, color spaces, but there were correctness
problems with that logic.
In addition, a number of "const" qualifiers have been removed, because
the action of bumping the reference count on an object is not consistent
with the C semantics of const.
This change is required to resolve problems with enhancement 688924. It
should also help with memory leaks.
Other clients, such as the PCL interpreter, will require changes to be
compatible. For simple device colorspaces, the change to gs_setcmykcolor()
(gscolor1.c) is representative of the code change required.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@7765 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@6651 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
memory.
Bug 688572
DETAILS:
The following code doesn't initialize all the fields in the stream structure.
stream s; sread_string(&s, ...); but it initializes enough fields to be mostly
working. The bug has been detected by Valgrind.
DIFFERENCES:
None
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@6643 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
DETAILS :
Bug 688255 "ai7 pdf fails on 7.03, runs for ten + minutes on 8.51".
The old code always allocates a transparency buffers for entire band.
The new code accounts group bbox to minimize buffers.
Due to that buffers appear empty for many of bands.
The time consumption for the test case of the bug 688255 is dropped in about 100 times
(from 8000 seconds to 71 seconds on a 3.07GHz machine, measured with debug build).
1. The transparency bbox computes in pdf14_begin_transparency_group from
the group bbox and the CTM (gdevp14.c).
2. Handle an empty buffer pdf14_buf_new, pdf14_pop_transparency_group (gdevp14.c).
3. Fixed a bug in the rectangle clipping in
pdf14_mark_fill_rectangle, pdf14_mark_fill_rectangle_ko_simple.
The old code didn't sense it because bbox always covered entire band (gdevp14.c).
4. Write the bbox to clist in c_pdf14trans_write and read it in c_pdf14trans_read.
5. The pdf14 compositor needs CTM to transform the group bbox to the device space.
Forced the writing of CTM to clist before writing the compositor in clist_create_compositor.
(Sorry, it appears some ugly due to pcte->type->procs.write creates a body
of a command, but we need to create a set of two commands;
Another minor optimization - a narrowing the set of bands - is delayed,
see comments in code in clist_create_compositor) (gxclimag.c).
6. New functions cmd_write_ctm_return_length, cmd_write_ctm are factored out for (5)
(gxclpath.c, gxclpath.h). This part of the change is algorithmically eqiuivalent.
7. Minor change : fixed coding style of "} else {" in gdevp14.c .
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@6147 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
Bug 687709 that could result in a Segmentation Fault (SEGV)
DETAILS:
The normal s_alloc() streams included s_init() on the allocated stream,
but C stack streams could have uninitialized elements in the stream struct.
Since this was UMR related, and affected the clist mode, the jpeg device
and the pswrite device (and possibly some pdfwrite data), this would be
different on different platforms.
The premature 'CloseTarget' action when the stream->close_strm was not
initialized as 0 (false) would show up as a SEGV when the -Z@ was used.
Other problems that this patch fixes are not linked to a test file at
the current time.
This change also corrects a trivial typo and a NUL character found in
src/gxtype1.c as a header file related warning in src/imain.c
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@5408 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
EXPECTED DIFFERENCES :
None.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@4406 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
Old code skipped segments above the page boundary regardless of the line width.
Fixes bug 686788 from customer 710.
DETAILS:
The area is calculated now as [y - expansion.y, y + height + expansion.y]
for every band. Beware of FOR_RECTS_NO_ERROR loop silently updating y and
height.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@4225 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
Update some Windows color mapping functions for DeviceN.
Change order of arguments for the gs_pdf14_device_filter_push
function and the calls to the device filter push procedure to
match the prototype.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@3122 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
1. Most of the code changes are required to turn off overprint mode when
processing images, smooth shading fills (shfill operator), or
when using PatternType 2 (smooth shading) patterns.
2. A few other changes were required to correct improper updating of the
overprint compositor information.
3. The code has been modified to avoid the crash that was observed for
the file 86554321.pdf on MSWindows XP systems.
4. The command list playback code has been modified to handle filling of
thin lines in a manner consistent with the noband case. This allows
files such as mspro.pdf to be rendered consistently in the banded and
noband case.
Jan
Modified Files:
lib/gs_img.ps
Added code to turn off overprint mode for images.
lib/gs_ll3.ps
Added code to turn off overprint mode for shfill.
src/gscolor2.c
1. Modified gs_setcolorspace to call gs_do_set_overprint rather than
invoking the color space's set_overprint routine directly. This
allows pattern color spaces to be handled (somewhat more) correctly.
2. Modified gs_setcolor to update overprint information. This is
necessary for pattern color spaces, where the color space to be used
for overprint purposes may be part of the color.
src/gscspace.c
Modified gx_set_overprint_DeviceCMYK to use the effective_overprint_mode
parameter in the image state (rather than the overprint_mode parameter
itself). This is part of the change necessary to allow overprint mode to
be turned off when a PatternType2 pattern is used as the current color.
src/gsdevice.c
Modified gs_set_device_no_init to call gs_do_set_overprint rather than
invoking the color space's set_overprint routine directly.
src/gsdps1.c
Modified rectfill to use the effective_overprint_mode parameter in the
imager state (rather than the overprint_mode parameter).
src/gsovrc.c
1. Fixed a bug in overprint_draw_thin_line: the third and fourth
operands in the call to gx_default_draw_thin_line (the y coordinate
of the first point and the x-coordinate of the second) were
in reverse order. This caused random lines to appear on various
overprint mode tests (e.g.: BEST99-00-Path.fh7.pdf,
Testform.v1.0.2.pdf).
2. Fixed typo in comment.
3. Modified the fill_in_procs procedure to initialize the device color
model information for the temporary device used in this routine.
Though the color model information does not affect the output of
the routine, reasonable values must be provided to avoid problems
when gx_device_forward_fill_in_procs calls gx_device_fill_in_procs.
This would cause certain systems to crash with files that involved
overprint/overprint mode (e.g.: 86554321.pdf on MSWinXP/MSVC systems).
4. Modified several checks that treated the separable_and_linear flag
in the device color model information as a boolean to handle this
parameter as an enumeration instead (since it is an enumeration).
src/gsptype1.c
Added gs_pattern1_set_color, the set_color method appropriate for
PatternType 1 patterns. This method performs overprint updating at
setcolor time, a requirement for patterns as the "colorspace" to be
used for overprint purposes is determined (in part) by the current
"color" (pattern).
src/gsptype2.c
Added gs_pattern2_set_color, the set_color method appropriate for
PatternType 2 patterns.
src/gsstate.c
1. Changed the criterion used in gs_grestore_only to determine if the
overprint compositor must be updated. Restoring the graphic state
may change the current color, which in the case of patterns may
change the "colorspace" to be used for overprint purposes. The code
now properly handles this possibility.
2. Added the gs_do_set_overprint routine, and modified various other
routines that directly call the current color space's
set_overprint method to call this procedure instead. The procedure
implements the special handling required for pattern color spaces,
where the current color may need to be considered when determining
which color model components must be retained.
3. Added code to initialize the effective_overprint_mode parameter in
the imager state, and provided two external routines to manipulate
this parameter. The first routine, gs_disable_effective_overprint_mode,
sets the effective_overprint_mode parameter to false, irrespective
of the setting of the overprint_mode parameter. The second,
gs_reset_effective_overprint_mode, sets the effective_overprint_mode
parameter to have the same value as the overprint_mode parameter.
The effective_overprint_mode parameter is used to turn off
overprint_mode when a PatternType 2 pattern is being used.
src/gsstate.h
Added prototypes for gs_do_set_overprint,
gs_disable_effective_overprint_mode, and
gs_reset_effective_overprint_mode.
src/gxccache.c
Modified gx_image_cached_char to invoke gx_set_dev_color before calling
(non-standard) device fill_mask methods. This handles several
situations in which a device color is not set when a cached character
is to be imaged, in particular for the command list device.
src/gxclpath.c
Added a comment to indicate that, though the overprint and
overprint_mode parameters are passed through the command list, they
usually have no effect, as overprint is controlled by the compositor
information that is (also) passed through the command list.
src/gxclrast.c
1. Added a comment to indicate that the overprint and overprint mode
parameters passed through the command list usually have no effect.
2. Added code to initialize the effective_parameter_mode field in the
imager state (though this field is usually in-effective).
3. Initialized the fill_zero_width field of the fill_params structure
in clist_playback_band to have the same value as in the noband case.
This parameter was previously always set to false, which led to
dropped lines for some files (e.g.: mspro.pdf).
src/gxistate.h
Added the effective_overprint_mode parameter.
src/gxpcolor.h
Added a set_color method to the pattern method vector. This method
performs any PatternType-specific actions necessary at setcolor
time. For now, this just updates the overprint compositor as required.
src/int.mak
src/lib.mak
Updated various dependency lists.
src/zcolor.c
Added code to disable overprint mode for PatternType 2 patterns.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@3020 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
we released on Aug. 26 to two problems:
1. The halftone phase information is not adjusted for bands that
start at locations other than the top of the page.
2. The "moveto closepath" style of degenerate path, which our code
allows to be passed through the band list, is used in the path
encoding code to indicate that a path lies entirely outside the
current band (i.e.: all of the path operators between the initial
moveto and the final closepath have been skipped because they are
outside of the band). This resulted in the code passing isolated
"closepath" segments through the command list, which much confused
the band list renderer.
The attached code release fixes both of these problems, and a couple of
other more minor issues we discovered. With this change, our system has
identical output for the banded and unbanded case for the pkmraw device
at 300 dpi, for all but one FTS file (027-09.ps). In that one case, the
banded and unbanded output are visually identical, so the output difference
may not be significant. We will look into this situation, and test some
of the other devices in the next day or so.
As best we can tell, halftone phase adjustment has never been correctly
implemented for the command list device. The pre-DeviceN code handles
binary halftone colors properly, but does not set the phase for general
colored halftones. General colored halftones are much more common in the
DeviceN code, which probably is the reason Dan first noticed the problem
with FTS file 245-07.ps (and others) with the DeviceN code.
Jan
Modified Files:
src/gsht.c
Modified the code that calculates the lcm_height field of a device
halftone in gx_imager_dev_ht_install to use the full_height rather
than the height of a halftone order. This typo was introduced sometime
during the DeviceN development; the pre-DeviceN code was correct.
src/gsptype1.c
src/gsptype2.c
src/gxwts.c
Added the "get_phase" method to various gx_device_color_type_t
instances.
src/gxcht.c
src/gxht.c
Added the "get_phase" method to a pair of gx_device_color_type_t
instances. Also modified the "write" and "read" methods of these
structures to ignore halftone phase information. This information
must be passed to and be directly handled by the command list
writer code, so that the command list renderer code may correct it
for the band starting coordinates.
src/gxclpath.c
1. Modified cmd_put_drawing_color to directly handle halftone phase
information, using the new "get_phase" method of device colors.
The passing of this information must be perfomed at the command
list level, so that the command list renderer can correct the
phase for the starting point of the band.
2. Also in cmd_put_drawing_color, "unset" the saved color information
if the halftone id does not match the most recently set halftone,
rather than setting the saved color pointer to 0.
3. In cmd_put_path, when processing a segment containing a
moveto followed by a closepath, check if the moveto was to a point
outside of the current band, and discard the closepath if this
is the case. This is necessary to avoid inserting isolated
closepath segments into the command list.
src/gxclrast.c
Added code to initialize the dev_color variable to clist_playback_band
to "no color".
src/gxdcolor.c
src/gxdcolor.h
Added the "get_phase" method to the gx_device_color_type_t structure,
and installed this method in various instances of the structure. Also
provided the two canonical implementations of this method.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2948 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
src/gdevdbit.c:
Changed gx_default_fill_mask to use the fill_masked routine of the
device color. This avoids an instance in which a possibly
uninitialized halftone tile is used for drawing. It also makes the
code considerably cleaner.
src/gscdefs.h
Re-installed the definition of extern_gs_find_compositor, which had
in-advertently been lost in a recent code merge.
src/gsdcolor.h
Added the colors.binary.b_level field to the gx_device_color_saved
structure. Also added a comment explaining why this structure is
used rather than the gx_device_color structure itself.
src/gsht.c
This file should be identical to the version released on Aug. 22nd.
We are releasing again here just to guard against this code being
committed without the former change.
src/gsht1.c
Modified the check for "binaryness" of a device color model in
gs_sethalftone_try_wts to handle cases in which not all of the depth
bits are used.
src/gstext.c,
src/gxtext.h
src/zchar.c
These files are identical to the set released on Aug. 23rd (the fix
for color changes in a kshow procedure). We are releasing them again
here to make certain the rest of this code is not committed without
the former change.
src/gxacpath.c
Modified gx_cpath_accum_end to set the new bbox_accurate field rather
than the bbox_set field. This avoids overloading the bbox_set field
in a way that can lead to inconsistent output when the setbbox
operator is used with a path.
src/gxcht.c
src/gxht.c
src/gxdcolor.c
src/gxdcolor.h
Modified the "write" methods for various color spaces so that their
return value can indicate that the operand and last transmitted
colors are the same. Also fixed various bugs uncovered when the
modular device color serialization/de-serialization was first used.
The changes to gxdcolor.h involve only comments.
src/gxcldev.h
1. Added a saved device color (gx_device_color_saved) to the band
state structure.
2. Removed the (long since unnecessary) type operand from the
cmd_put_halftone procedure.
3. Commented out some now obsolete command codes.
src/gxclimag.c
1. Modified clist_fill_mask to work with the new, modular drawing
color serialization/de-serialization routines.
2. Rewrote cmd_put_halftone as required by the new mechanism for
serializing/de-serializing device halftones.
3. Modified various invocations of cmd_put_halftone to discard the
the type operand.
Note: The existing code impose a requirement that image objects that
used a device color (e.g.: image masks) would only be passed
through the command list if that color is pure. We do not
know why this limitation was necessary, and suspect it is no
longer necessary, but we have left it in place for now.
src/gxclist.c
Added gxdcolor.h to the included file list.
src/gxclpath.c
1. Rewrote cmd_put_drawing_color to use the new mechanism for
serialization/de-serialization of device colors.
2. Removed the device color type information from the stroke/
fill/polyfill commands.
3. Modified cmd_put_path so that "moveto closepath" is not
passed through the command list as just "moveto".
src/gxclpath.h
1. Added the cmd_opv_ext_put_ht_seg, to allow device halftones to be
split across multiple command buffers. Also added the macro
cbuf_ht_seg_max_size, to indicate the largest halftone segment to
be put in a single command buffer.
2. Commented out some now obsolete command codes.
src/gxclrast.c
1. Modified the code to read device halftones and device colors to
use the new serialization/de-serialization methods for these
structures.
2. Modified various locations in which partial/temporary device
colors are used to use the current color for the band.
3. Removed code for the various stroke/fill/polyfill commands that
encode device color type as part of the command.
4. Added code to save the current color when a pure color must be
set to generate a clip region using the fill procedure.
5. Modified the clip path reading code to flatten the clip path
before passing it to the fill operation. This is necessary to
ensure consistent results between the banded and non-banded
cases (the non-banded case also flattens a clip path before
passing it to the fill operation).
src/gxdhtserial.c
1. Updated gx_ht_read_tf and gx_ht_read_and_install to properly
use the reference count in a transfer function.
2. Fixed various other errors discovered when this code was first
used.
src/gxdhtserial.h
Fixed typo.
src/gxpath.c
Initialize the new bbox_accurate field of the gx_path structure to
false in gx_path_init_contents.
src/gxpath2.c
Modified gx_path_bbox to use the bbox_accurate rather than bbox_set
field to determine if the bounding box of a path needs to be
explicitly calculated. If bbox_set is true, the bounding box
in the bbox field will contain the current path, but may be larger
than necessary. Due to a problem in the clip/fill code, the larger
bounding box may generate slightly inconsistent results.
src/gzpath.h
Added the bbox_accurate field to the gx_path_s structure. This field
is used by the clip accumulator device, which avoids the need to
overload the bbox_set field (which is also used by the setbbox
operator).
src/lib.mak
Updated varius dependency lists.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2934 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2925 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
files.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2745 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
RCSfile and Revision pair.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2241 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
Artifex Software's commercial release). References file LICENSE.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2234 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
clist,
causing segment fault during playback. Changes use proper number
of device color components to be written to clist and read from it.
SourceForge bug #211449.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@1251 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
Ghostscript to AFPL Ghostscript, as well as some internal references to the
program name.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@729 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
self-consistent state.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@673 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
state.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@671 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
and banding don't do the right thing yet.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@669 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
additions are now implemented (at least as stubs) and documented, as is
bookkeeping for the new scalar graphics state members (but not for the soft
masks or the transparency group stack).
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@658 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
FUNCTIONALITY IS NOT IMPLEMENTED YET.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@657 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
that doesn't like ptr = buf - 1 where buf is an array.
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@206 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
git-svn-id: http://svn.ghostscript.com/ghostscript/trunk@2 a1074d23-0009-0410-80fe-cf8c14f379e6
|
|
git-svn-id: http://svn.ghostscript.com/ghostpcl/trunk/ghostpcl@1089 06663e23-700e-0410-b217-a244a6096597
|
|
git-svn-id: http://svn.ghostscript.com/ghostpcl/trunk/ghostpcl@1040 06663e23-700e-0410-b217-a244a6096597
|
