Support multi-char arg separators.

PR: #8274

Pass core_globals to new inline functions

Submitted by: Sebastian Bergmann

Move code from macros to inline functions to ease debugging

Reset the state, if there is garbage in an HTML tag.

PR: #7854

Validators are happy now.

Reinstantiate single quotes properly.

Add url_adapt_single_url() API which modifies a single URL. Might be used
to modify the Location: URL.

Hack for #7262, not fixed yet

Initialize tags properly.

Submitted by: Andreas Karajannis <Andreas.Karajannis@gmd.de>

Back out ws changes

make compile again

Enable users to set the HTML tags to rewrite
through a configuration directive

Ignore close tags and use unsigned char type.

Recognize single quotes around cdata.

Move smart_str functions to an extra file.

Unify memory allocation in smart_str functions

Cleanup SCANNER_DEBUG and don't NUL terminate work buffer

Use emalloc and friends. Since we reuse all buffers, this does not affect
the speed.

The smart_str_* functions don't NUL terminate strings anymore. We can
rely on the length of the s

Use emalloc and friends. Since we reuse all buffers, this does not affect
the speed.

The smart_str_* functions don't NUL terminate strings anymore. We can
rely on the length of the strings completely.

show more ...

Remove this pseudo optimization from the previous rewrite.

Since we leave/enter the loop more often, the cost of maintaining
c_arg/c_val is higher than simply copying the tag/arg string

Remove this pseudo optimization from the previous rewrite.

Since we leave/enter the loop more often, the cost of maintaining
c_arg/c_val is higher than simply copying the tag/arg string once.

show more ...

Another 5%-10% speed increase achieved by

- making ctx->val read-only
- adding a special appendc function for appending single characters

Add support for URLs containing NAME info (for scrolling; #name).

2nd Rewrite of the main scanner engine. This version is cleaner and faster.

It handles the following tags currently:

<FRAME .. SRC=URL ..>
<INPUT .. TYPE=IMAGE SRC=URL ..>
<

2nd Rewrite of the main scanner engine. This version is cleaner and faster.

It handles the following tags currently:

<FRAME .. SRC=URL ..>
<INPUT .. TYPE=IMAGE SRC=URL ..>
<A .. HREF=URL ..>
<AREA .. HREF=URL ..>
<IMG .. SRC=URL ..>

Additional ones can be added within seconds.

The support for HTML forms has been significantly improved. The scanner
will now add proper Hidden Fields for the Session ID.

show more ...

Reduce the amount of copying in the mainloop. We copy tag/arg only, if we
need to preserve them for a new loop iteration (after we leave mainloop).
Otherwise, we can just let them point to th

Reduce the amount of copying in the mainloop. We copy tag/arg only, if we
need to preserve them for a new loop iteration (after we leave mainloop).
Otherwise, we can just let them point to the work area.

show more ...

Revert part of last commit. I'll refine it later.

Another speedup by avoiding string copies.

15% speed gain in test case (rewriting 1M URLs).

Remove nextiter label

s/DEBUG/SCANNER_DEBUG/. "Convert" one macro to inline function to be more
consistent. Also use continue.