Stripping the URL fragment

Lagrange will retain URL fragments when parsing gemtext, but will strip them when making requests or when a DocumentWidget's URL is set. This allows opening URLs with fragments in an external browser.

IssueID #128

4 files changed, 11 insertions, 7 deletions

diff --git a/src/gmrequest.c b/src/gmrequest.c
index a1cf1a8e..8626403f 100644
--- a/src/gmrequest.c
+++ b/src/gmrequest.c
@@ -493,7 +493,7 @@ void deinit_GmRequest(iGmRequest *d) {
 }
 
 void setUrl_GmRequest(iGmRequest *d, const iString *url) {
-    set_String(&d->url, url);
+    set_String(&d->url, urlFragmentStripped_String(url));
     /* Encode hostname to Punycode here because we want to submit the Punycode domain name
        in the request. (TODO: Pending possible Gemini spec change.) */
     punyEncodeUrlHost_String(&d->url);
diff --git a/src/gmutil.c b/src/gmutil.c
index 0f1bc803..72074278 100644
--- a/src/gmutil.c
+++ b/src/gmutil.c
@@ -49,11 +49,12 @@ void init_Url(iUrl *d, const iString *text) {
     iRegExpMatch m;
     init_RegExpMatch(&m);
     if (matchString_RegExp(urlPattern_, text, &m)) {
-        d->scheme = capturedRange_RegExpMatch(&m, 2);
-        d->host   = capturedRange_RegExpMatch(&m, 4);
-        d->port   = (iRangecc){ d->host.end, d->host.end };
-        d->path   = capturedRange_RegExpMatch(&m, 5);
-        d->query  = capturedRange_RegExpMatch(&m, 6);
+        d->scheme   = capturedRange_RegExpMatch(&m, 2);
+        d->host     = capturedRange_RegExpMatch(&m, 4);
+        d->port     = (iRangecc){ d->host.end, d->host.end };
+        d->path     = capturedRange_RegExpMatch(&m, 5);
+        d->query    = capturedRange_RegExpMatch(&m, 6);
+        d->fragment = capturedRange_RegExpMatch(&m, 8); /* starts with a hash */
         /* Check if the authority contains a port. */
         init_RegExpMatch(&m);
         if (matchRange_RegExp(authPattern_, d->host, &m)) {
@@ -92,6 +93,7 @@ void stripDefaultUrlPort_String(iString *d) {
 }
 
 const iString *urlFragmentStripped_String(const iString *d) {
+    /* Note: Could use `iUrl` here and leave out the fragment. */
     const size_t fragPos = indexOf_String(d, '#');
     if (fragPos != iInvalidPos) {
         return collect_String(newRange_String((iRangecc){ constBegin_String(d),
@@ -264,6 +266,7 @@ const iString *absoluteUrl_String(const iString *d, const iString *urlMaybeRelat
         appendRange_String(absolute, orig.path);
     }
     appendRange_String(absolute, rel.query);
+    appendRange_String(absolute, rel.fragment);
     normalize_String(absolute);
     cleanUrlPath_String(absolute);
     return absolute;
diff --git a/src/gmutil.h b/src/gmutil.h
index 3fd1268d..1caf2445 100644
--- a/src/gmutil.h
+++ b/src/gmutil.h
@@ -96,6 +96,7 @@ struct Impl_Url {
     iRangecc port;
     iRangecc path;
     iRangecc query;
+    iRangecc fragment;
 };
 
 void            init_Url                (iUrl *, const iString *text);
diff --git a/src/ui/documentwidget.c b/src/ui/documentwidget.c
index 7fffc214..79a77f61 100644
--- a/src/ui/documentwidget.c
+++ b/src/ui/documentwidget.c
@@ -3345,7 +3345,7 @@ void deserializeState_DocumentWidget(iDocumentWidget *d, iStream *ins) {
 
 void setUrlFromCache_DocumentWidget(iDocumentWidget *d, const iString *url, iBool isFromCache) {
     d->flags &= ~showLinkNumbers_DocumentWidgetFlag;
-    set_String(d->mod.url, url);
+    set_String(d->mod.url, urlFragmentStripped_String(url));
     /* See if there a username in the URL. */
     parseUser_DocumentWidget_(d);
     if (!isFromCache || !updateFromHistory_DocumentWidget_(d)) {