vndb/manned
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix handling of URLs ending in a ⟩

Browse source 
I've known about this issue before, but didn't realize it was so
widespread. This fixes many links.
This commit is contained in:
Yorhel 2016-10-16 09:11:13 +02:00
parent 7d31f41ba8
commit 17fc298217
2 changed files with 15 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

1
lib/ManUtils/Build.PL
View file

@ -5,6 +5,7 @@ use Module::Build;
Module::Build->new( Module::Build->new(
dist_name => 'ManUtils', dist_name => 'ManUtils',
dist_version_from => 'ManUtils.pm', dist_version_from => 'ManUtils.pm',
dist_abstract => 'Utils for manned.org',
pm_files => { pm_files => {
'ManUtils.pm' => 'lib/ManUtils.pm', 'ManUtils.pm' => 'lib/ManUtils.pm',
}, },

17
lib/ManUtils/ManUtils.xs
View file

@ -169,9 +169,8 @@ static void flushline(ctx_t *x) {
// HTTP(s) URL. // HTTP(s) URL.
// This is just a simple q{https?://[^ ][.,;"\)>]?( |$)} match, doesn't // This is just a simple q{https?://[^ ][.,;"\)>]?( |$)} match, doesn't
// always work right: // always work right, e.g.:
// - troff.1: ⟨http://www.gnu.org/copyleft/fdl.html⟩. <- yes, that's an Unicode character. // - https://manned.org/spu_run/414316a1 -> URL wrapped to new line
// - roff.7: Has quite a few issues with wrapped URLs and situations similar to the above.
// Note: Don't use strncmp() before manually checking for 'http'. The parse // Note: Don't use strncmp() before manually checking for 'http'. The parse
// time is otherwise increased by a factor 2. // time is otherwise increased by a factor 2.
if(s[0] == 'h' && s[1] == 't' && s[2] == 't' && s[3] == 'p' && (strncmp(s, "http://", 7) == 0 || strncmp(s, "https://", 8) == 0)) { if(s[0] == 'h' && s[1] == 't' && s[2] == 't' && s[3] == 'p' && (strncmp(s, "http://", 7) == 0 || strncmp(s, "https://", 8) == 0)) {
@ -189,6 +188,18 @@ static void flushline(ctx_t *x) {
endchr = *sp; endchr = *sp;
*(sp--) = 0; *(sp--) = 0;
} }
// Also catch a Unicode '⟩', which is how groff sometimes ends a .UR, e.g.:
// - https://manned.org/troff/c4467840
// - https://manned.org/pass/78413b49
// - https://manned.org/empathy-accounts/8c05b2c1
// - https://manned.org/urn/8cb83e85
// - https://manned.org/wine/4a699a22
if(*sp == '\xa9' && *(sp-1) == '\x9f' && *(sp-2) == '\xe2') {
sp[1] = endchr;
sp -= 3;
endchr = sp[1];
sp[1] = 0;
}
sv_catpvf(x->dest, "<a href=\"%s\" rel=\"nofollow\">%s</a>", s, s); sv_catpvf(x->dest, "<a href=\"%s\" rel=\"nofollow\">%s</a>", s, s);
*(++sp) = endchr; *(++sp) = endchr;
es = s = sp; es = s = sp;