notgne2/tuxpaint-pencil-sharpener
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

now --lang works (must not use stale pointers)

Browse source
This commit is contained in:
Albert Cahalan 2009-11-21 21:17:19 +00:00
parent f296909c75
commit 1c980026f0
2 changed files with 19 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

1
src/i18n.c
View file

@ -741,6 +741,7 @@ static void do_locale_option(const char *const arg)
int setup_i18n(const char *restrict lang, const char *restrict locale)
{
printf("lang %p, locale %p\n", lang, locale);
printf("lang \"%s\", locale \"%s\"\n", lang, locale);
if(lang)
set_langstr(lang);

20
src/tuxpaint.c
View file

@ -18319,7 +18319,19 @@ void rec_undo_label(void)
return;
}
if (have_to_rec_label_node && current_label_node != NULL) /* FIXME: Bill added the "current_label_node != NULL" test to avoid crashing in the following situation: (1) use label tool, (2) start new drawing, (3) change to paint tool & try to draw. Is this test correct, or is the crash here an artifact of a different bug? -bjk 2009.10.05 */
// FIXME: Bill added the "current_label_node != NULL" test to avoid
// crashing in the following situation: (1) use label tool, (2) start new
// drawing, (3) change to paint tool & try to draw. Is this test correct,
// or is the crash here an artifact of a different bug? -bjk 2009.10.05
//
// It's all wrong to have a separate undo stack anyway. We need a way
// for arbitrary code to supply callback functions and parameters when
// creating an undo entry. One obvious function is a destructor for the
// private data, for when it drops off the far end of the stack or gets
// wiped out by an undo,draw combo. Others might be for when the level
// stops being current or for when the level becomes current again.
if (have_to_rec_label_node && current_label_node != NULL)
{
current_label_node->save_undoid = cur_undo;
text_undo[cur_undo] = 1;
@ -18926,7 +18938,11 @@ static void parse_file_options(struct cfginfo *restrict tmpcfg, const char *file
char *arg = strchr(str,'=');
if(arg)
*arg++ = '\0';
parse_one_option(tmpcfg,str,arg,filename);
// FIXME: leaking mem here, but the trouble is that these
// strings get mixed in with ones from .data and .rodata
// and free() isn't smart about the situation -- also some
// of the strings end up being kept around
parse_one_option(tmpcfg,str,strdup(arg),filename);
}
fclose(fi);
}