Remove wrongly repeated words in comments

Clean up misrepetitions ('if if', 'the the' etc) found throughout the
comments. This has been done manually, after grepping
case-insensitively for duplicate if, is, the, then, do, for, an,
plus a few other typos corrected in fly-by

v2:
    * proper commit message and non-joke title;
    * replace two 'as is' followed by 'is' to 'as-is'.
v3:
    * 'a integer' => 'an integer' and similar (originally spotted by
      Jason Ekstrand, I fixed a few other similar ones while at it)

Signed-off-by: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Reviewed-by: Chad Versace <chad.versace@intel.com>

include/GL/mesa_glinterop.h

@@ -97,7 +97,7 @@ struct mesa_glinterop_device_info {
    /* The callee will overwrite it if it supports a lower version.
     *
     * The caller should check the value and access up-to the version supported
-    * by the the callee.
+    * by the callee.
     */
    /* NOTE: Do not use the MESA_GLINTEROP_DEVICE_INFO_VERSION macro */
    uint32_t version;
@@ -125,7 +125,7 @@ struct mesa_glinterop_export_in {
    /* The callee will overwrite it if it supports a lower version.
     *
     * The caller should check the value and access up-to the version supported
-    * by the the callee.
+    * by the callee.
     */
    /* NOTE: Do not use the MESA_GLINTEROP_EXPORT_IN_VERSION macro */
    uint32_t version;
@@ -190,7 +190,7 @@ struct mesa_glinterop_export_out {
    /* The callee will overwrite it if it supports a lower version.
     *
     * The caller should check the value and access up-to the version supported
-    * by the the callee.
+    * by the callee.
     */
    /* NOTE: Do not use the MESA_GLINTEROP_EXPORT_OUT_VERSION macro */
    uint32_t version;

src/compiler/glsl/ast_to_hir.cpp

@@ -1276,7 +1276,7 @@ void
 ast_expression::set_is_lhs(bool new_value)
 {
    /* is_lhs is tracked only to print "variable used uninitialized" warnings,
-    * if we lack a identifier we can just skip it.
+    * if we lack an identifier we can just skip it.
     */
    if (this->primary_expression.identifier == NULL)
       return;

src/compiler/glsl/glsl_to_nir.cpp

@@ -1950,7 +1950,7 @@ void
 nir_visitor::visit(ir_constant *ir)
 {
    /*
-    * We don't know if this variable is an an array or struct that gets
+    * We don't know if this variable is an array or struct that gets
     * dereferenced, so do the safe thing an make it a variable with a
     * constant initializer and return a dereference.
     */

src/compiler/nir/nir.h

@@ -804,7 +804,7 @@ typedef struct {
 } nir_deref_var;
 
 /* This enum describes how the array is referenced.  If the deref is
- * direct then the base_offset is used.  If the deref is indirect then then
+ * direct then the base_offset is used.  If the deref is indirect then
  * offset is given by base_offset + indirect.  If the deref is a wildcard
  * then the deref refers to all of the elements of the array at the same
  * time.  Wildcard dereferences are only ever allowed in copy_var

src/compiler/nir/nir_instr_set.c

@@ -514,7 +514,7 @@ nir_instr_set_add_or_rewrite(struct set *instr_set, nir_instr *instr)
       nir_instr *match = (nir_instr *) entry->key;
       nir_ssa_def *new_def = nir_instr_get_dest_ssa_def(match);
 
-      /* It's safe to replace a exact instruction with an inexact one as
+      /* It's safe to replace an exact instruction with an inexact one as
        * long as we make it exact.  If we got here, the two instructions are
        * exactly identical in every other way so, once we've set the exact
        * bit, they are the same.

src/compiler/nir/nir_intrinsics.h

@@ -48,9 +48,9 @@ INTRINSIC(copy_var, 0, ARR(0), false, 0, 2, 0, xx, xx, xx, 0)
 
 /*
  * Interpolation of input.  The interp_var_at* intrinsics are similar to the
- * load_var intrinsic acting an a shader input except that they interpolate
+ * load_var intrinsic acting on a shader input except that they interpolate
  * the input differently.  The at_sample and at_offset intrinsics take an
- * aditional source that is a integer sample id or a vec2 position offset
+ * additional source that is an integer sample id or a vec2 position offset
  * respectively.
  */
 

src/compiler/nir/nir_lower_vars_to_ssa.c

@@ -621,7 +621,7 @@ rename_variables_block(nir_block *block, struct lower_variables_state *state)
  *     fully-direct references we see and store them in the
  *     direct_deref_nodes hash table.
  *
- *  2) Walk over the the list of fully-qualified direct derefs generated in
+ *  2) Walk over the list of fully-qualified direct derefs generated in
  *     the previous pass.  For each deref, we determine if it can ever be
  *     aliased, i.e. if there is an indirect reference anywhere that may
  *     refer to it.  If it cannot be aliased, we mark it for lowering to an

src/compiler/nir/nir_lower_wpos_ytransform.c

@@ -159,7 +159,7 @@ lower_fragcoord(lower_wpos_ytransform_state *state, nir_intrinsic_instr *intr)
     *
     * The bias of the y-coordinate depends on whether y-inversion takes place
     * (adjY[1]) or not (adjY[0]), which is in turn dependent on whether we are
-    * drawing to an FBO (causes additional inversion), and whether the the pipe
+    * drawing to an FBO (causes additional inversion), and whether the pipe
     * driver origin and the requested origin differ (the latter condition is
     * stored in the 'invert' variable).
     *

src/compiler/nir/nir_opt_dead_cf.c

@@ -30,7 +30,7 @@
 
 /*
  * This file implements an optimization that deletes statically
- * unreachable/dead code. In NIR, one way this can happen if if an if
+ * unreachable/dead code. In NIR, one way this can happen is when an if
  * statement has a constant condition:
  *
  * if (true) {

src/compiler/spirv/spirv_to_nir.c

@@ -2458,7 +2458,7 @@ vtn_handle_execution_mode(struct vtn_builder *b, struct vtn_value *entry_point,
       b->shader->info.cs.local_size[2] = mode->literals[2];
       break;
    case SpvExecutionModeLocalSizeHint:
-      break; /* Nothing do do with this */
+      break; /* Nothing to do with this */
 
    case SpvExecutionModeOutputVertices:
       assert(b->shader->stage == MESA_SHADER_GEOMETRY);

src/gallium/docs/source/context.rst

@@ -234,7 +234,7 @@ include several layers), this surface need not be bound to the framebuffer.
 
 ``clear_depth_stencil`` clears a single depth, stencil or depth/stencil surface
 with the specified depth and stencil values (for combined depth/stencil buffers,
-is is also possible to only clear one or the other part). While it is only
+it is also possible to only clear one or the other part). While it is only
 possible to clear one surface at a time (which can include several layers),
 this surface need not be bound to the framebuffer.
 

src/gallium/drivers/freedreno/freedreno_texture.h

@@ -55,7 +55,7 @@ struct fd_texture_stateobj;
  *         | fp16[3]   /
  *  0x08:  | padding
  *  0x10:  | int16[0]  \
- *         | int16[1]  |___ swizzled int16 channels for for "small integer"
+ *         | int16[1]  |___ swizzled int16 channels for "small integer"
  *         | int16[2]  |    formats (<= 16 bits per component, integer)
  *         | int16[3]  /
  *  0x18:  | padding

src/gallium/drivers/freedreno/ir3/ir3.c

@@ -573,7 +573,7 @@ void * ir3_assemble(struct ir3 *shader, struct ir3_info *info,
 		}
 	}
 
-	/* need a integer number of instruction "groups" (sets of 16
+	/* need an integer number of instruction "groups" (sets of 16
 	 * instructions on a4xx or sets of 4 instructions on a3xx),
 	 * so pad out w/ NOPs if needed: (NOTE each instruction is 64bits)
 	 */

src/gallium/drivers/llvmpipe/lp_state_derived.c

@@ -116,7 +116,7 @@ compute_vertex_info(struct llvmpipe_context *llvmpipe)
          /*
           * Note that we'd actually want to skip position (as we won't use
           * the attribute in the fs) but can't. The reason is that we don't
-          * actually have a input/output map for setup (even though it looks
+          * actually have an input/output map for setup (even though it looks
           * like we do...). Could adjust for this though even without a map
           * (in llvmpipe_create_fs_state()).
           */

src/gallium/drivers/nouveau/codegen/nv50_ir_util.h

@@ -199,7 +199,7 @@ public:
       virtual void erase();
       virtual bool insert(void *data);
 
-      // move item to a another list, no consistency with its iterators though
+      // move item to another list, no consistency with its iterators though
       void moveToList(DLList&);
 
    private:

src/gallium/drivers/nouveau/nouveau_buffer.c

@@ -163,7 +163,7 @@ nouveau_transfer_staging(struct nouveau_context *nv,
    return tx->map;
 }
 
-/* Copies data from the resource into the the transfer's temporary GART
+/* Copies data from the resource into the transfer's temporary GART
  * buffer. Also updates buf->data if present.
  *
  * Maybe just migrate to GART right away if we actually need to do this. */

src/gallium/drivers/r300/compiler/radeon_dataflow.c

@@ -850,7 +850,7 @@ static void init_get_readers_callback_data(
  * of the loop it reads the value written by instruction 0 and in all other
  * iterations it reads the value written by instruction 3.
  *
- * @param read_cb This function will be called for for every instruction that
+ * @param read_cb This function will be called for every instruction that
  * has been determined to be a reader of writer.
  * @param write_cb This function will be called for every instruction after
  * writer.

src/gallium/drivers/r300/compiler/radeon_vert_fc.c

@@ -134,7 +134,7 @@ static void lower_bgnloop(
 	} else {
 		fc_state->PredStack[fc_state->LoopDepth] =
 						fc_state->PredicateReg;
-		/* Copy the the current predicate value to this loop's
+		/* Copy the current predicate value to this loop's
 		 * predicate register */
 
 		/* Use the old predicate value for src0 */

src/gallium/drivers/softpipe/sp_state_derived.c

@@ -176,14 +176,14 @@ softpipe_compute_vertex_info(struct softpipe_context *softpipe)
             /*
              * Note that we'd actually want to skip position (as we won't use
              * the attribute in the fs) but can't. The reason is that we don't
-             * actually have a input/output map for setup (even though it looks
+             * actually have an input/output map for setup (even though it looks
              * like we do...). Could adjust for this though even without a map.
              */
          } else {
             /*
              * Note that we'd actually want to skip position (as we won't use
              * the attribute in the fs) but can't. The reason is that we don't
-             * actually have a input/output map for setup (even though it looks
+             * actually have an input/output map for setup (even though it looks
              * like we do...). Could adjust for this though even without a map.
              */
             draw_emit_vertex_attr(vinfo, EMIT_4F, vs_index);

src/gallium/drivers/svga/svga_format.c

@@ -1916,7 +1916,7 @@ svga_format_size(SVGA3dSurfaceFormat format,
    *block_width = format_cap_table[format].block_width;
    *block_height = format_cap_table[format].block_height;
    *bytes_per_block = format_cap_table[format].block_bytes;
-   /* Make sure the the table entry was valid */
+   /* Make sure the table entry was valid */
    if (*block_width == 0)
       debug_printf("Bad table entry for %s\n", svga_format_name(format));
    assert(*block_width);

src/gallium/drivers/svga/svga_tgsi_vgpu10.c

@@ -1762,7 +1762,7 @@ alloc_immediate_float4(struct svga_shader_emitter_v10 *emit,
 
 
 /**
- * Allocate space for a int[4] immediate.
+ * Allocate space for an int[4] immediate.
  * \return  the index/position of the immediate.
  */
 static unsigned

src/gallium/drivers/swr/rasterizer/core/backend.h

@@ -208,13 +208,13 @@ INLINE void generateInputCoverage(const uint64_t *const coverageMask, uint32_t (
 
     mask[0] = _mm256_set_epi8(-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0xC, 0x8, 0x4, 0x0,
                               -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, 0xC, 0x8, 0x4, 0x0);
-    // pull out the the 8bit 4x2 coverage for samples 0-7 into the lower 32 bits of each 128bit lane
+    // pull out the 8bit 4x2 coverage for samples 0-7 into the lower 32 bits of each 128bit lane
     __m256i packedCoverage0 = _simd_shuffle_epi8(sampleCoverage[0], mask[0]);
 
     __m256i packedCoverage1;
     if(T::MultisampleT::numSamples > 8)
     {
-        // pull out the the 8bit 4x2 coverage for samples 8-15 into the lower 32 bits of each 128bit lane
+        // pull out the 8bit 4x2 coverage for samples 8-15 into the lower 32 bits of each 128bit lane
         packedCoverage1 = _simd_shuffle_epi8(sampleCoverage[1], mask[0]);
     }
 

src/gallium/drivers/swr/rasterizer/core/state.h

@@ -627,7 +627,7 @@ struct SWR_STREAMOUT_STATE
 
     // The stream masks specify which attributes are sent to which streams.
     // These masks help the FE to setup the pPrimData buffer that is passed
-    // the the Stream Output Shader (SOS) function.
+    // the Stream Output Shader (SOS) function.
     uint32_t streamMasks[MAX_SO_STREAMS];
 
     // Number of attributes, including position, per vertex that are streamed out.

src/gallium/drivers/swr/rasterizer/core/threads.cpp

@@ -445,7 +445,7 @@ void WorkOnFifoBE(
 
         for (uint32_t tileID : macroTiles)
         {
-            // Only work on tiles for for this numa node
+            // Only work on tiles for this numa node
             uint32_t x, y;
             pDC->pTileMgr->getTileIndices(tileID, x, y);
             if (((x ^ y) & numaMask) != numaNode)

src/gallium/drivers/vc4/kernel/vc4_validate_shaders.c

@@ -163,7 +163,7 @@ check_tmu_write(uint64_t inst,
 			return false;
 		}
 
-		/* We assert that the the clamped address is the first
+		/* We assert that the clamped address is the first
 		 * argument, and the UBO base address is the second argument.
 		 * This is arbitrary, but simpler than supporting flipping the
 		 * two either way.

src/gallium/drivers/vc4/vc4_qir_schedule.c

@@ -407,7 +407,7 @@ choose_instruction(struct schedule_state *state)
                 }
 
                 /* If we would block on the previously chosen node, but would
-                 * block less on this one, then then prefer it.
+                 * block less on this one, then prefer it.
                  */
                 if (chosen->unblocked_time > state->time &&
                     n->unblocked_time < chosen->unblocked_time) {

src/gallium/include/pipe/p_video_codec.h

@@ -134,17 +134,17 @@ struct pipe_video_buffer
    void (*destroy)(struct pipe_video_buffer *buffer);
 
    /**
-    * get a individual sampler view for each plane
+    * get an individual sampler view for each plane
     */
    struct pipe_sampler_view **(*get_sampler_view_planes)(struct pipe_video_buffer *buffer);
 
    /**
-    * get a individual sampler view for each component
+    * get an individual sampler view for each component
     */
    struct pipe_sampler_view **(*get_sampler_view_components)(struct pipe_video_buffer *buffer);
 
    /**
-    * get a individual surfaces for each plane
+    * get an individual surfaces for each plane
     */
    struct pipe_surface **(*get_surfaces)(struct pipe_video_buffer *buffer);
 

src/gallium/state_trackers/nine/device9.c

@@ -236,7 +236,7 @@ NineDevice9_ctor( struct NineDevice9 *This,
         NineUnknown_ConvertRefToBind(NineUnknown(This->state.rt[i]));
     }
 
-    /* Initialize a dummy VBO to be used when a a vertex declaration does not
+    /* Initialize a dummy VBO to be used when a vertex declaration does not
      * specify all the inputs needed by vertex shader, on win default behavior
      * is to pass 0,0,0,0 to the shader */
     {

src/gbm/main/gbm.c

@@ -361,7 +361,7 @@ gbm_bo_create(struct gbm_device *gbm,
  *   GBM_BO_IMPORT_EGL_IMAGE
  *   GBM_BO_IMPORT_FD
  *
- * The the gbm bo shares the underlying pixels but its life-time is
+ * The gbm bo shares the underlying pixels but its life-time is
  * independent of the foreign object.
  *
  * \param gbm The gbm device returned from gbm_create_device()

src/glx/indirect_glx.c

@@ -169,7 +169,7 @@ indirect_unbind_context(struct glx_context *gc, struct glx_context *new)
    if (gc == new)
       return;
    
-   /* We are either switching to no context, away from a indirect
+   /* We are either switching to no context, away from an indirect
     * context to a direct context or from one dpy to another and have
     * to send a request to the dpy to unbind the previous context.
     */

src/gtest/include/gtest/internal/gtest-port.h

@@ -176,7 +176,7 @@
 //                         string.
 //
 // Integer types:
-//   TypeWithSize   - maps an integer to a int type.
+//   TypeWithSize   - maps an integer to an int type.
 //   Int32, UInt32, Int64, UInt64, TimeInMillis
 //                  - integers of known sizes.
 //   BiggestInt     - the biggest signed integer type.

src/gtest/src/gtest.cc

@@ -1461,7 +1461,7 @@ std::string CodePointToUtf8(UInt32 code_point) {
   return str;
 }
 
-// The following two functions only make sense if the the system
+// The following two functions only make sense if the system
 // uses UTF-16 for wide string encoding. All supported systems
 // with 16 bit wchar_t (Windows, Cygwin, Symbian OS) do use UTF-16.
 

src/mesa/drivers/dri/i965/brw_device_info.h

@@ -73,7 +73,7 @@ struct brw_device_info
     * In general, you can find shader thread maximums by looking at the "Maximum
     * Number of Threads" field in the Intel PRM description of the 3DSTATE_VS,
     * 3DSTATE_GS, 3DSTATE_HS, 3DSTATE_DS, and 3DSTATE_PS commands. URB entry
-    * limits come from the "Number of URB Entries" field in the the
+    * limits come from the "Number of URB Entries" field in the
     * 3DSTATE_URB_VS command and friends.
     *
     * These fields are used to calculate the scratch space to allocate.  The

src/mesa/drivers/dri/i965/brw_fs.cpp

@@ -1180,7 +1180,7 @@ fs_visitor::emit_general_interpolation(fs_reg *attr, const char *name,
 
          /* Data starts at suboffet 3 in 32-bit units (12 bytes), so it is not
           * 64-bit aligned and the current implementation fails to read the
-          * data properly. Instead, when there is is a double input varying,
+          * data properly. Instead, when there is a double input varying,
           * read it as vector of floats with twice the number of components.
           */
          if (attr->type == BRW_REGISTER_TYPE_DF) {

src/mesa/drivers/dri/i965/brw_nir_opt_peephole_ffma.c

@@ -113,7 +113,7 @@ get_mul_for_src(nir_alu_src *src, int num_components,
       break;
 
    case nir_op_fmul:
-      /* Only absorb a fmul into a ffma if the fmul is is only used in fadd
+      /* Only absorb a fmul into a ffma if the fmul is only used in fadd
        * operations.  This prevents us from being too aggressive with our
        * fusing which can actually lead to more instructions.
        */

src/mesa/drivers/dri/i965/brw_performance_monitor.c

@@ -1016,7 +1016,7 @@ wrap_bookend_bo(struct brw_context *brw)
 }
 
 /* This is fairly arbitrary; the trade off is memory usage vs. extra overhead
- * from wrapping.  On Gen7, 32768 should be enough for for 128 snapshots before
+ * from wrapping.  On Gen7, 32768 should be enough for 128 snapshots before
  * wrapping (since each is 256 bytes).
  */
 #define BOOKEND_BO_SIZE_BYTES 32768

src/mesa/drivers/x11/xm_buffer.c

@@ -451,7 +451,7 @@ xmesa_MapRenderbuffer(struct gl_context *ctx,
 
          assert(xrb->pixmap);
 
-         /* Install error handler for XGetImage() in case the the window
+         /* Install error handler for XGetImage() in case the window
           * isn't mapped.  If we fail we'll create a temporary XImage.
           */
          mesaXErrorFlag = 0;

src/mesa/main/dlist.c

@@ -6026,7 +6026,7 @@ save_MultiTexCoord4fv(GLenum target, const GLfloat * v)
 
 
 /**
- * Record a GL_INVALID_VALUE error when a invalid vertex attribute
+ * Record a GL_INVALID_VALUE error when an invalid vertex attribute
  * index is found.
  */
 static void

src/mesa/main/formatquery.c

@@ -559,7 +559,7 @@ _is_internalformat_supported(struct gl_context *ctx, GLenum target,
     */
    GLint buffer[1];
 
-   /* At this point a internalformat is valid if it is valid as a texture or
+   /* At this point an internalformat is valid if it is valid as a texture or
     * as a renderbuffer format. The checks are different because those methods
     * return different values when passing non supported internalformats */
    if (_mesa_base_tex_format(ctx, internalformat) < 0 &&

src/mesa/main/imports.h

@@ -73,7 +73,7 @@ extern "C" {
 
 /**
  * Sometimes we treat GLfloats as GLints.  On x86 systems, moving a float
- * as a int (thereby using integer registers instead of FP registers) is
+ * as an int (thereby using integer registers instead of FP registers) is
  * a performance win.  Typically, this can be done with ordinary casts.
  * But with gcc's -fstrict-aliasing flag (which defaults to on in gcc 3.0)
  * these casts generate warnings.

src/mesa/main/texobj.c

@@ -1196,7 +1196,7 @@ invalidate_tex_image_error_check(struct gl_context *ctx, GLuint texture,
  * glCreateTextures should throw errors if target = 0. This is not exposed to
  * the rest of Mesa to encourage Mesa internals to use nameless textures,
  * which do not require expensive hash lookups.
- * \param target  either 0 or a a valid / error-checked texture target enum
+ * \param target  either 0 or a valid / error-checked texture target enum
  */
 static void
 create_textures(struct gl_context *ctx, GLenum target,

src/mesa/program/ir_to_mesa.cpp

@@ -1975,7 +1975,7 @@ ir_to_mesa_visitor::visit(ir_texture *ir)
       ir->coordinate->accept(this);
 
    /* Put our coords in a temp.  We'll need to modify them for shadow,
-    * projection, or LOD, so the only case we'd use it as is is if
+    * projection, or LOD, so the only case we'd use it as-is is if
     * we're doing plain old texturing.  Mesa IR optimization should
     * handle cleaning up our mess in that case.
     */

src/mesa/state_tracker/st_atom_msaa.c

@@ -54,7 +54,7 @@ static void update_sample_mask( struct st_context *st )
          /* there's lot of ways how to do this. We just use first few bits,
             since we have no knowledge of sample positions here. When
             app-supplied mask though is used too might need to be smarter.
-            Also, there's a interface restriction here in theory it is
+            Also, there's an interface restriction here in theory it is
             encouraged this mask not be the same at each pixel. */
          sample_mask = (1 << nr_bits) - 1;
          if (st->ctx->Multisample.SampleCoverageInvert)

src/mesa/state_tracker/st_cb_fbo.c

@@ -266,7 +266,7 @@ st_new_renderbuffer(struct gl_context *ctx, GLuint name)
 
 
 /**
- * Allocate a renderbuffer for a an on-screen window (not a user-created
+ * Allocate a renderbuffer for an on-screen window (not a user-created
  * renderbuffer).  The window system code determines the format.
  */
 struct gl_renderbuffer *

src/mesa/state_tracker/st_glsl_to_tgsi.cpp

@@ -3877,7 +3877,7 @@ glsl_to_tgsi_visitor::visit(ir_texture *ir)
       ir->coordinate->accept(this);
 
       /* Put our coords in a temp.  We'll need to modify them for shadow,
-       * projection, or LOD, so the only case we'd use it as is is if
+       * projection, or LOD, so the only case we'd use it as-is is if
        * we're doing plain old texturing.  The optimization passes on
        * glsl_to_tgsi_visitor should handle cleaning up our mess in that case.
        */
@@ -5832,7 +5832,7 @@ emit_wpos(struct st_context *st,
     *
     * The bias of the y-coordinate depends on whether y-inversion takes place
     * (adjY[1]) or not (adjY[0]), which is in turn dependent on whether we are
-    * drawing to an FBO (causes additional inversion), and whether the the pipe
+    * drawing to an FBO (causes additional inversion), and whether the pipe
     * driver origin and the requested origin differ (the latter condition is
     * stored in the 'invert' variable).
     *

src/mesa/state_tracker/st_mesa_to_tgsi.c

@@ -830,7 +830,7 @@ emit_wpos(struct st_context *st,
     *
     * The bias of the y-coordinate depends on whether y-inversion takes place
     * (adjY[1]) or not (adjY[0]), which is in turn dependent on whether we are
-    * drawing to an FBO (causes additional inversion), and whether the the pipe
+    * drawing to an FBO (causes additional inversion), and whether the pipe
     * driver origin and the requested origin differ (the latter condition is
     * stored in the 'invert' variable).
     *