Fix typos (libvips#3693)

Found by codespell v2.2.6.

ChangeLog

@@ -2361,7 +2361,7 @@ _ add vips_foreign_get_suffixes()
 
 27/7/07 started 7.12.4
 - proto.h had vars called small, breaking on win32
-- more python fixing, we now have working matricies too
+- more python fixing, we now have working matrices too
 
 17/7/07 started 7.12.3
 - fix to VImage.i for gcc 4.2 (thanks Damir)
@@ -2736,7 +2736,7 @@ _ add vips_foreign_get_suffixes()
 - added "ccittfax4" as a TIFF compression option
 - fix all 64-bit compiler warnings
 - "," allowed as column separator in mask read
-- better at spotting singular matricies
+- better at spotting singular matrices
 - small im_render() tidies
 - glib dependency reduced to just 2.0, but untested ... helps people building
   on older systems who aren't interested in nip2

cplusplus/include/vips/VImage8.h

@@ -1237,7 +1237,7 @@ class VImage : public VObject {
 
 	/**
 	 * Dilate the image with the specified structuring element, see
-	 * VImage::new_matrix(). Stucturing element values can be 0 for
+	 * VImage::new_matrix(). Structuring element values can be 0 for
 	 * black, 255 for white and 128 for don't care. See VImage::morph().
 	 */
 	VImage
@@ -1249,7 +1249,7 @@ class VImage : public VObject {
 
 	/**
 	 * Erode the image with the specified structuring element, see
-	 * VImage::new_matrix(). Stucturing element values can be 0 for
+	 * VImage::new_matrix(). Structuring element values can be 0 for
 	 * black, 255 for white and 128 for don't care. See VImage::morph().
 	 */
 	VImage

libvips/convolution/convi.c

@@ -20,7 +20,7 @@
  * 1/2/95 JC
  *	- use of IM_REGION_ADDR() updated
  *	- output size was incorrect! see comment below
- *	- bug with large non-square matricies fixed too
+ *	- bug with large non-square matrices fixed too
  *	- uses new im_embed() function
  * 13/7/98 JC
  *	- weird bug ... im_free_imask is no longer directly called for close

libvips/create/logmat.c

@@ -257,7 +257,7 @@ vips_logmat_init(VipsLogmat *logmat)
  * of radius
  * @sigma.  The size of the mask is determined by the variable @min_ampl;
  * if for instance the value .1 is entered this means that the produced mask
- * is clipped at values within 10 persent of zero, and where the change
+ * is clipped at values within 10 percent of zero, and where the change
  * between mask elements is less than 10%.
  *
  * The program uses the following equation: (from Handbook of Pattern

libvips/deprecated/mask_dispatch.c

@@ -1,4 +1,4 @@
-/* VIPS function dispatch tables for matricies.
+/* VIPS function dispatch tables for matrices.
  *
  * J. Cupitt, 14/2/95.
  */
@@ -47,7 +47,7 @@ static im_arg_desc one_in_one_out[] = {
 	IM_OUTPUT_DMASK("out")
 };
 
-/* Two matricies in, one out.
+/* Two matrices in, one out.
  */
 static im_arg_desc two_in_one_out[] = {
 	IM_INPUT_DMASK("in1"),

libvips/deprecated/rw_mask.c

@@ -10,9 +10,9 @@
  *	Instead, they set default values of scale=1, offset=0.
  *    - Code tidied up, better error recovery.
  *    -	Bugs fixed in im_dup_*mask. No longer coredump.
- *    -	Bugs fixed in im_write_*mask. Now work for non-square matricies.
+ *    -	Bugs fixed in im_write_*mask. Now work for non-square matrices.
  *    -	im_copy_*mask_matrix, im_copy_matrix_*mask added: copy VIPS mask
- *	structures into Numerical Recipies in C style matricies and vice
+ *	structures into Numerical Recipies in C style matrices and vice
  *	versa. Both structures should have been built before copy attempted.
  *	See im_create_*mask, im_*mat_alloc. The matrix should be indexed by 0
  *	to size-1.

libvips/foreign/csvload.c

@@ -686,7 +686,7 @@ vips_foreign_load_csv_source_init(VipsForeignLoadCsvSource *source)
  *
  * Load a CSV (comma-separated values) file. The output image is always 1
  * band (monochrome), #VIPS_FORMAT_DOUBLE. Use vips_bandfold() to turn
- * RGBRGBRGB mono images into colour iamges.
+ * RGBRGBRGB mono images into colour images.
  *
  * Items in lines can be either floating point numbers in the C locale, or
  * strings enclosed in double-quotes ("), or empty.

libvips/foreign/matlab.c

@@ -44,7 +44,7 @@
 
 + it will not do complex images
 
-+ it will not handle sparse matricies
++ it will not handle sparse matrices
 
 + it loads the first variable in the file with between 1 and 3 dimensions,
   is this sensible behaviour?

libvips/include/vips/interpolate.h

@@ -107,7 +107,7 @@ VIPS_API
 int vips_interpolate_get_window_offset(VipsInterpolate *interpolate);
 
 /* How many bits of precision we keep for transformations, ie. how many
- * pre-computed matricies we have.
+ * pre-computed matrices we have.
  */
 #define VIPS_TRANSFORM_SHIFT (6)
 #define VIPS_TRANSFORM_SCALE (1 << VIPS_TRANSFORM_SHIFT)

libvips/iofuncs/image.c

@@ -184,7 +184,7 @@
  *
  * #VIPS_DEMAND_STYLE_SMALLTILE --- This is the most general demand format.
  * Output is demanded in small (around 100x100 pel) sections. This style works
- * reasonably efficiently, even for bizzarre operations like 45 degree rotate.
+ * reasonably efficiently, even for bizarre operations like 45 degree rotate.
  *
  * #VIPS_DEMAND_STYLE_FATSTRIP --- This operation would like to output strips
  * the width of the image and as high as possible. This option is suitable

libvips/resample/affine.c

@@ -42,7 +42,7 @@
  * 14/4/01 JC
  *	- oops, invert_point() had a rounding problem
  * 23/2/02 JC
- *	- pre-calculate interpolation matricies
+ *	- pre-calculate interpolation matrices
  *	- integer interpolation for int8/16 types, double for
  *	  int32/float/double
  *	- faster transformation

libvips/resample/bicubic.cpp

@@ -79,7 +79,7 @@ typedef VipsInterpolateClass VipsInterpolateBicubicClass;
  * scale + 1 so we can round-to-nearest safely.
  */
 
-/* We could keep a large set of 2d 4x4 matricies, but this actually
+/* We could keep a large set of 2d 4x4 matrices, but this actually
  * works out slower since for many resizes the thing will no longer
  * fit in L1.
  */