quakeforge/libs/ui/test/test-passage.c

141 lines
3.8 KiB
C
Raw Normal View History

[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
#include <stdio.h>
#include "QF/ui/view.h"
#include "QF/ui/passage.h"
#include "QF/ecs/component.h"
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
enum {
test_href,
};
static const component_t test_components[] = {
[test_href] = {
.size = sizeof (hierref_t),
.create = 0,//create_href,
.name = "href",
},
};
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
static const char test_text[] = {
"Guarding the entrance to the Grendal "
"Gorge is the Shadow Gate, a small keep "
"and monastary which was once the home "
"of the Shadow cult.\n"
" For years the Shadow Gate existed in "
"obscurity but after the cult discovered "
"the \u00c2\u00ec\u00e1\u00e3\u00eb\u2002\u00c7\u00e1\u00f4\u00e5 "
"in the caves below the empire took notice. "
"A batallion of Imperial Knights were "
"sent to the gate to destroy the cult "
"and claim the artifact for the King.\nasdf",
};
static int __attribute__((pure))
check_non_space (const char *text, psg_text_t *to)
{
int size;
for (uint32_t offs = 0; offs < to->size; offs += size) {
if (!(size = Passage_IsSpace (text + to->text + offs))) {
return 1;
}
}
return 0;
}
static int __attribute__((pure))
check_space_or_nl (const char *text, psg_text_t *to)
{
for (uint32_t offs = 0; offs < to->size; offs++) {
if (text[to->text + offs] == '\n'
|| Passage_IsSpace (text + to->text + offs)) {
return 1;
}
}
return 0;
}
int
main (void)
{
int ret = 0;
ecs_registry_t *registry = ECS_NewRegistry ();
ECS_RegisterComponents (registry, test_components, 1);
registry->href_comp = test_href;
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
passage_t *passage = Passage_ParseText (test_text, registry);
if (passage->hierarchy->childCount[0] != 3) {
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
ret = 1;
printf ("incorrect number of paragraphs: %d\n",
passage->hierarchy->childCount[0]);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
}
if (passage->hierarchy->num_objects != 144) {
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
ret = 1;
printf ("incorrect number of text objects: %d\n",
passage->hierarchy->num_objects);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
}
if (passage->hierarchy->childCount[1] != 49) {
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
ret = 1;
printf ("incorrect number of text objects in first paragraph: %d\n",
passage->hierarchy->childCount[1]);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
}
if (passage->hierarchy->childCount[2] != 90) {
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
ret = 1;
printf ("incorrect number of text objects in second paragraph: %d\n",
passage->hierarchy->childCount[2]);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
}
if (passage->hierarchy->childCount[3] != 1) {
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
ret = 1;
printf ("incorrect number of text objects in third paragraph: %d\n",
passage->hierarchy->childCount[3]);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
}
uint32_t *childIndex = passage->hierarchy->childIndex;
psg_text_t *text_objs = passage->hierarchy->components[0];
psg_text_t *to = &text_objs[childIndex[2] + 0];
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
if (to->size != 2 && (passage->text[to->text] != ' '
&& passage->text[to->text + 1] != ' ')) {
ret = 1;
printf ("second paragram does not begin with double space: %d '%.*s'\n",
to->size, to->size, passage->text + to->text);
}
//if (text_view->bol_suppress) {
// ret = 1;
// printf ("second paragram indent suppressed\n");
//}
for (uint32_t i = 0; i < passage->hierarchy->childCount[0]; i++) {
for (uint32_t j = 0; j < passage->hierarchy->childCount[1 + i]; j++) {
psg_text_t *to = &text_objs[childIndex[1 + i] + j];
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
unsigned is_space = Passage_IsSpace (passage->text + to->text);
if (i == 1 && j == 0) {
// second paragraph indent, tested above
continue;
}
//if ((!!is_space) != text_view->bol_suppress) {
// ret = 1;
// printf ("text/suppress mismatch %d [%d '%.*s'] %d %d\n",
// text_view->bol_suppress, to->size, to->size,
// passage->text + to->text, i, j);
//}
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
if (is_space) {
if (!check_non_space (passage->text, to)) {
continue;
}
} else {
if (!check_space_or_nl (passage->text, to)) {
continue;
}
}
ret = 1;
printf ("mixed space/text/\\n [%d '%.*s'] %d %d\n",
to->size, to->size, passage->text + to->text, i, j);
}
}
Passage_Delete (passage);
ECS_DelRegistry (registry);
[ui] Add a sub-system for parsing text passages A passage object has a list of all the text objects in the given string, where the objects represent either white space or "words", as well as a view_t object representing the entire passage, with paragraphs split into child views of the passage view, and each paragraph has a child view for every text/space object in the paragraph. Paragraphs are split by '\n' (not included in any object). White space is grouped into clumps such that multiple adjacent spaces form a single object. The standard ASCII space (0x20) and all of the Unicode characters marked "WS;<compat> 0020" are counted as white space. Unless a white space object is the first in the paragraph, its view is marked for suppression by the view flow code. Contiguous non-white space characters are grouped into single objects, and their views are not suppressed. All text object views (both white space and "word") have their data pointer set to the psg_text_t object representing the text for that view. This should be suitable for simple text-mode unattributed display. More advanced rendering would probably want to create suitable objects and set the view data pointers to those objects. No assumption is made about text direction. Passage and paragraph views need to have their primary axis sizes set appropriately, as well as their resize flags. Their xlen and ylen are both set to 10, and xpos,ypos is 0,0. Paragraph views need their setgeometry pointer set to the appropriate view_flow_* function. However, they are set up to have their secondary axis set automatically when flowed. Text object views are set up for automatic flowing: grav_flow, 0,0 for xpos,ypos. However, xlen and ylen are also both 0, so need to be set by the renderer before attempting to flow the text.
2022-09-30 10:05:18 +00:00
return ret;
}