updates to <code> tag http://www.allegro.cc/forums/view/599658 Allegro.cc Forum Thread matthew@allegro.cc (Matthew Leverton) Thu, 26 Mar 2009 03:43:51 +0000

The code tag now looks like:

<code name="filename" start="1">
</code>

where both parameters are optional. The name attribute is meant to be the name or title of the source code. The start attribute is the first line number. If omitted, it is assumed to be 1.

Short code snippets with no name and starting line number will be inline without any chrome.

Lines can be highlighted by appending *** to them.

To turn off line numbers (copy / paste) click the # button. Click the 'E' to get rid of the scrollbars.

allegro.c#SelectExpand
1#include <stdio.h> 2#include <stdlib.h> 3#include <string.h> 4 5#include "allegro5/allegro5.h" 6#include "allegro5/internal/aintern.h" 7#include "allegro5/internal/aintern_dtor.h" 8#include "allegro5/internal/aintern_memory.h" 9 10 11 12#define PREFIX_I "al-main INFO: " 13#define PREFIX_W "al-main WARNING: " 14#define PREFIX_E "al-main ERROR: " 15 16 17/* debugging stuff */
18static int debug_assert_virgin = true;
19static int debug_trace_virgin = true; 20 21static FILE *assert_file = NULL; 22static FILE *trace_file = NULL; 23 24static int (*assert_handler)(AL_CONST char *msg) = NULL; 25int (*_al_trace_handler)(AL_CONST char *msg) = NULL; 26 27 28/* dynamic registration system for cleanup code */ 29struct al_exit_func { 30 void (*funcptr)(void); 31 AL_CONST char *desc; 32 struct al_exit_func *next; 33}; 34 35static struct al_exit_func *exit_func_list = NULL; 36 37 38 39/* _al_add_exit_func: 40 * Adds a function to the list that need to be called on Allegro shutdown. 41 * `desc' should point to a statically allocated string to help with 42 * debugging. 43 */ 44void _al_add_exit_func(void (*func)(void), AL_CONST char *desc) 45{ 46 struct al_exit_func *n; 47 48 for (n = exit_func_list; n; n = n->next) 49 if (n->funcptr == func) 50 return; 51 52 n = _AL_MALLOC(sizeof(struct al_exit_func)); 53 if (!n) 54 return; 55 56 n->next = exit_func_list; 57 n->funcptr = func; 58 n->desc = desc; 59 exit_func_list = n; 60} 61 62 63 64/* _al_remove_exit_func: 65 * Removes a function from the list that need to be called on Allegro 66 * shutdown. 67 */ 68void _al_remove_exit_func(void (*func)(void)) 69{ 70 struct al_exit_func *iter = exit_func_list, *prev = NULL; 71 72 while (iter) { 73 if (iter->funcptr == func) { 74 if (prev) 75 prev->next = iter->next; 76 else 77 exit_func_list = iter->next; 78 _AL_FREE(iter); 79 return; 80 } 81 prev = iter; 82 iter = iter->next; 83 } 84} 85 86 87 88/* _al_run_exit_funcs: 89 * Run all the functions registered with _al_add_exit_func, in reverse order of 90 * registration. 91 */ 92void _al_run_exit_funcs(void) 93{ 94 while (exit_func_list) { 95 void (*func)(void) = exit_func_list->funcptr; 96 _al_remove_exit_func(func); 97 (*func)(); 98 } 99} 100 101 102 103/* debug_exit: 104 * Closes the debugging output files. 105 */ 106static void debug_exit(void) 107{ 108 if (assert_file) { 109 fclose(assert_file); 110 assert_file = NULL; 111 } 112 113 if (trace_file) { 114 fclose(trace_file); 115 trace_file = NULL; 116 } 117 118 debug_assert_virgin = true; 119 debug_trace_virgin = true; 120} 121 122 123 124/* al_assert: 125 * Raises an assert (uses ASCII strings). 126 */ 127void al_assert(AL_CONST char *file, int line) 128{ 129 static int asserted = false; 130 int olderr = errno; 131 char buf[128]; 132 char *s; 133 134 if (asserted) 135 return; 136 137 /* todo, some day: use snprintf (C99) */ 138 sprintf(buf, "Assert failed at line %d of %s", line, file); 139 140 if (assert_handler) { 141 if (assert_handler(buf)) 142 return; 143 } 144 145 if (debug_assert_virgin) { 146 s = getenv("ALLEGRO_ASSERT"); 147 148 if (s) 149 assert_file = fopen(s, "w"); 150 else 151 assert_file = NULL; 152 153 if (debug_trace_virgin) 154 _al_add_exit_func(debug_exit, "debug_exit"); 155 156 debug_assert_virgin = false; 157 } 158 159 if (assert_file) { 160 fprintf(assert_file, "%s\n", buf); 161 fflush(assert_file); 162 } 163 else { 164 asserted = true; 165 166/* 167 if ((system_driver) && (system_driver->assert)) { 168 system_driver->assert(buf); 169 } 170 else { 171 */ 172// al_uninstall_system(); 173#ifndef ALLEGRO_MSVC 174 fprintf(stderr, "%s\n", buf); 175#endif 176 abort(); 177 // } 178 } 179 180 errno = olderr; 181} 182 183 184 185/* al_trace: 186 * Outputs a trace message (uses ASCII strings). 187 */ 188void al_trace(AL_CONST char *msg, ...) 189{ 190 int olderr = errno; 191 char buf[512]; 192 char *s; 193 194 /* todo, some day: use vsnprintf (C99) */ 195 va_list ap; 196 va_start(ap, msg); 197 vsprintf(buf, msg, ap); 198 va_end(ap); 199 200 if (_al_trace_handler) { 201 if (_al_trace_handler(buf)) 202 return; 203 } 204 205 if (debug_trace_virgin) { 206 s = getenv("ALLEGRO_TRACE"); 207 208 if (s) 209 trace_file = fopen(s, "w"); 210 else 211 trace_file = fopen("allegro.log", "w"); 212 213 if (debug_assert_virgin) 214 _al_add_exit_func(debug_exit, "debug_exit"); 215 216 debug_trace_virgin = false; 217 } 218 219 if (trace_file) { 220 fwrite(buf, sizeof(char), strlen(buf), trace_file); 221 fflush(trace_file); 222 } 223 224 errno = olderr; 225} 226 227 228 229/* register_assert_handler: 230 * Installs a user handler for assert failures. 231 */ 232void register_assert_handler(int (*handler)(AL_CONST char *msg)) 233{ 234 assert_handler = handler; 235} 236 237 238 239/* register_trace_handler: 240 * Installs a user handler for trace output. 241 */ 242void register_trace_handler(int (*handler)(AL_CONST char *msg)) 243{ 244 _al_trace_handler = handler; 245}

When making this post, I realized that there is a slight problem with multiline coloring (e.g., comments) that disables highlight ability and the original colors. That will be fixed later.

Refresh JS/CSS files.

]]> no-reply@allegro.cc (Matthew Leverton) Sat, 21 Mar 2009 02:57:49 +0000

Cool! I was gonna say it was ugly but then I refreshed the js/css. Very pretty.

cool.c#SelectExpand
125 int c = 46; 126
127 c = 43
128 129 if(c == 43)

The error is on line 127.

]]> no-reply@allegro.cc (ImLeftFooted) Sat, 21 Mar 2009 03:00:39 +0000

It takes a long time (two seconds roughly) to turn the line numbers off in IE7. Turning them back on is significantly faster.

A nice set of additions.

EDIT: Dustin's post removes the line #s fast.

]]> no-reply@allegro.cc (someone972) Sat, 21 Mar 2009 03:02:51 +0000

Hiding and showing line numbers fast enough on Matthew's example breaks the scroll bar (Firefox 3.0.7).
Other than that, this is pretty sweet!

]]> no-reply@allegro.cc (Schyfis) Sat, 21 Mar 2009 03:50:53 +0000

Cool stuff.

When I disable line numbers, the font size changes from 10 to 8. Using Firefox 3.1b3 on Windows.

]]> no-reply@allegro.cc (torhu) Sat, 21 Mar 2009 04:05:44 +0000

Very nice! I really like all the new work you've put into the site lately Matthew!
This is by far the best forum system I've used!

Schyfis said:

Hiding and showing line numbers fast enough on Matthew's example breaks the scroll bar

Same here

]]> no-reply@allegro.cc (Jonatan Hedborg) Sat, 21 Mar 2009 04:22:20 +0000

Now add support for multiple languages.

]]> no-reply@allegro.cc (BAF) Sat, 21 Mar 2009 04:48:23 +0000

Very nice! Now it looks like a programming forum! Could you add some way of selecting all code instead of having to scroll around? That is the only thing I would want.

The only problem I find is that when scrolling with the well the page upwards or downwards, the top of the code box somehow jumps above the navigation bar of the site, so you get some strange effect.

]]> no-reply@allegro.cc (ReyBrujo) Sat, 21 Mar 2009 04:50:23 +0000

Here's what it looks like in konqueror (4.3 svn), I'm sure you can see a couple flaws.

{"name":"accnewcodeboxkonqi.png","src":"\/\/djungxnpq2nug.cloudfront.net\/image\/cache\/2\/d\/2d9e97fe0d1c247dffc4b31cadee19bd.png","w":814,"h":425,"tn":"\/\/djungxnpq2nug.cloudfront.net\/image\/cache\/2\/d\/2d9e97fe0d1c247dffc4b31cadee19bd"}accnewcodeboxkonqi.png

]]> no-reply@allegro.cc (Thomas Fjellstrom) Sat, 21 Mar 2009 05:06:56 +0000

Use a better browser.

]]> no-reply@allegro.cc (BAF) Sat, 21 Mar 2009 05:43:32 +0000

edit:nm

edit2:
it was happening briefly with WebKit too, but now its not happening in either WebKit or KHTML.

]]> no-reply@allegro.cc (Thomas Fjellstrom) Sat, 21 Mar 2009 05:44:52 +0000

I find it really odd that everything there's an update, my cache seems to have been cleared out enough that I don't need to hit Ctrl+F5...

test.c#SelectExpand
1int c; 2int y; 3 4int main(int n, char **args) {
5 int c;
6 return 0; 7}

]]> no-reply@allegro.cc (MiquelFire) Sat, 21 Mar 2009 07:01:46 +0000
main.cpp#SelectExpand
1#include <iostream> 2 3int main(int argc, char *argv[]) 4{
5 std::cout << "Allegro.cc FTW!\n" << std::endl;
6 7 return 0; 8}

]]> no-reply@allegro.cc (bamccaig) Sat, 21 Mar 2009 07:08:38 +0000

The collapse/expand feature could stand to be a little more obvious. I don't think newbies would catch on without prompting.

]]> no-reply@allegro.cc (Kibiz0r) Sat, 21 Mar 2009 07:31:52 +0000
someone972 said:

It takes a long time (two seconds roughly) to turn the line numbers off in IE7. Turning them back on is significantly faster.

IE is a horrible web browser. If you have any choice, you shouldn't use it. Let's take a look at the benchmark of the said operation:

  1. Safari 4: 14 ms

  2. Opera 9.64: 50 ms

  3. FireFox 3: 61 ms

  4. IE 8: 817 ms

This is just a typical string build during a DOM traversal.

Safari is so fast you don't even realize anything has happened. Opera and Firefox are both nearly as fast—neither of them feel lagged. Then IE 8, Microsoft's latest and greatest browser, steps up to the plate and takes nearly a second.

But keep using IE.

]]> no-reply@allegro.cc (Matthew Leverton) Sat, 21 Mar 2009 08:37:21 +0000

Hmm... I would like the top (filename, numbering on/off) to be at the bottom too if you expand the code and scrolled to the bottom. Also, although I am not sure if it is possible at all, can I generate a URL to point to the line 104 of the code you posted at the top?

]]> no-reply@allegro.cc (ReyBrujo) Sat, 21 Mar 2009 09:07:28 +0000

Fixed some bugs.

]]> no-reply@allegro.cc (Matthew Leverton) Sat, 21 Mar 2009 10:57:00 +0000

No way to do inline code snippets yet?

]]> no-reply@allegro.cc (X-G) Sat, 21 Mar 2009 13:29:15 +0000

Thanks for fixing the font size, but any chance of courier new 10 instead of 8? Maybe I'm just old, but reading text that small is not a joy.

]]> no-reply@allegro.cc (torhu) Sat, 21 Mar 2009 20:06:07 +0000

You can use custom CSS to change things like that. 

div.source-code div.inner td,
div.source-code div.inner pre {
  font-size: 10pt !important;
}

]]> no-reply@allegro.cc (Matthew Leverton) Sat, 21 Mar 2009 20:19:01 +0000

Cool, thanks.

]]> no-reply@allegro.cc (torhu) Sat, 21 Mar 2009 20:33:37 +0000

Neat, now it looks precious.

]]> no-reply@allegro.cc (ReyBrujo) Sat, 21 Mar 2009 21:48:07 +0000
ReyBrujo said:

precious

Looks at your avatar
My precioussss....

]]> no-reply@allegro.cc (Darizel) Sat, 21 Mar 2009 22:20:40 +0000
X-G said:

No way to do inline code snippets yet?

Just <code> with no new lines. e.g., while (false) create_bitmap(640, 480)

]]> no-reply@allegro.cc (Matthew Leverton) Sun, 22 Mar 2009 10:50:24 +0000

test: int test(char bar);***

seems not to work :(

int test(char bar);

works...

]]> no-reply@allegro.cc (Thomas Fjellstrom) Sun, 22 Mar 2009 13:03:35 +0000

Matthew Leverton said:

Just <code> with no new lines. e.g., while (false) create_bitmap(640, 480)

Hmm. I could've sworn it didn't work when I tried it the first time.

]]> no-reply@allegro.cc (X-G) Sun, 22 Mar 2009 13:51:12 +0000

Hmm... I don't see the point in having inline code highlighted...

]]> no-reply@allegro.cc (ReyBrujo) Sun, 22 Mar 2009 20:12:54 +0000
ReyBrujo said:

Hmm... I don't see the point in having inline code highlighted...

To autolink to a function: blit().

]]> no-reply@allegro.cc (CGamesPlay) Sun, 22 Mar 2009 20:32:42 +0000

I mean, Thomas wanted to highlight a single inline line.

]]> no-reply@allegro.cc (ReyBrujo) Sun, 22 Mar 2009 20:35:47 +0000

Yeah, that would be a great way to annoy the crap out of people

]]> no-reply@allegro.cc (CGamesPlay) Sun, 22 Mar 2009 20:36:30 +0000

Well the ***s are just highlighted as operators.. Maybe they should be removed if they are there?

]]> no-reply@allegro.cc (Thomas Fjellstrom) Mon, 23 Mar 2009 02:18:06 +0000
ReyBrujo said:

Hmm... I don't see the point in having inline code highlighted...

For the same reason we highlight other code... To make it easier to read and/or catch mistakes. :-/

]]> no-reply@allegro.cc (bamccaig) Mon, 23 Mar 2009 02:30:35 +0000

Highlighting inline code is obnoxious and not supported. Removing the highlight characters would give the false impression that it is.

]]> no-reply@allegro.cc (Matthew Leverton) Mon, 23 Mar 2009 03:43:47 +0000

I just noticed that in the thread review when replying to a post the code boxes have large spaces in between lines.

Other than that I'm quite pleased with the update :).

]]> no-reply@allegro.cc (someone972) Thu, 26 Mar 2009 03:43:51 +0000