On Mon, May 07, 2001 at 07:13:28PM +0000, Bjorn Reese wrote:
I am failing to understand the following sequence in HTMLtree.c
/ htmlNodeDump. More accurately, I cannot see the point with the
last three lines.
if (!htmlIsAutoClosed(doc, cur)) {
xmlBufferWriteChar(buf, "</");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">");
}
xmlBufferWriteChar(buf, "</");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">");
Dohhh :-(
htmlIsAutoClosed returns true "if a tags is autoclosed by one of it's children" (if 'autoclosed' implies that an end-tag is inserted then I suspect that two end-tags will be inserted in this case, but I have not verified this). If htmlIsAutoClosed returns false, then two identical end-tags will be inserted. Why? My problem is that much of my HTML output have two end-tags per start-tag, which is fatal for embedded tags. Removing the last three lines above, seemed to solve this problem. As I was under time pressure, I did not had time to understand or experiment with the code, nor did I run the regression tests.
Problem is that this code doesn't show up in regression tests because they use a different set of routines (htmlNodeDumpOutput i.e. saving to an I/O buffer not to a memory buffer ... some cleanup might be needed :-\). I have cleaned up htmlNodeDump(), considering that htmlNodeDumpOutput() is the tested routine and that its output is correct. I commited the change to CVS, patch enclosed Daniel -- Daniel Veillard | Red Hat Network http://redhat.com/products/network/ veillard redhat com | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
Attachment:
htmldump.diff
Description: Text document