Re: xml2enc does not appear to be functional with mod-proxy-html 3.1.2

[Martin Gerdes <martingrds@xxxxxxxxxxxxxx>]

I think I now know what is going wrong in principle:

>ldd -r /usr/lib/apache2/modules/mod_xml2enc.so
undefined symbol: ap_filter_flush       (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_bucket_type_eos   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_bucket_type_flush (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_strcasestr (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_register_output_filter_protocol    (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_write     (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_pass_brigade       (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_array_push        (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_regexec    (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_create    (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_partition (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_dynamic_fn_register       (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_length    (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_xlate_conv_buffer (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_pstrndup  (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_check_cmd_context  (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_add_output_filter  (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_palloc    (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_cleanup   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_pregcomp   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_fflush     (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: xmlDetectCharEncoding (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: xmlGetCharEncodingName        (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: xmlAddEncodingAlias   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_pstrdup   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: xmlParseCharEncoding  (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_remove_output_filter       (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_pstrcat   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_brigade_flatten   (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: ap_log_rerror (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_array_make        (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_optional_hook_get (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_xlate_open        (/usr/lib/apache2/modules/mod_xml2enc.so)
undefined symbol: apr_bucket_transient_create   (/usr/lib/apache2/modules/mod_xml2enc.so)
        linux-vdso.so.1 =>  (0x00007ffff61fe000)
        libc.so.6 => /lib/libc.so.6 (0x00007f32edc1e000)
        /lib64/ld-linux-x86-64.so.2 (0x00007f32ee17b000)

If I understand the output correctly, the sybols provided by libxml2 are undefined, and it does not depend on libxml2 as it should.
Furthermore, of the 3 referenced libraries, only libc.so.6 exists (/lib/ld-linux-x86-64.so.2 does, but /lib64/ld-linux-x86-64.so.2 does not).

I did get warnings about the missing symbols when building the .deb package. But I ignored them thinking that a used but undefined symbol would throw an error somewhere along the line...!?

Not sure which of these problems is the decisive on, or how to solve them (I do have package libxml2-dev installed, which defines all these symbols under /usr/include/libxml2/, and contains /usr/lib/libxml2.so), but I will try to get those symbols defined. Hopefully that will do it.

All help and suggestions appreciated.

Martin

2009/10/28 Martin Gerdes <martingrds@xxxxxxxxxxxxxx>

Background:
I am running Apache as provided for Debian Lenny (2.2.9-10+lenny4) as
a reverse proxy. To that end, I installed libapache2-mod-proxy-html
(3.0.0-1) for URL rewriting.
As the original content is in charset ISO-8859-1, I set the option
"ProxyHTMLCharsetOut *" to get the webpage output as ISO-8859 instead
of UTF-8.

Everything worked just fine, except that a certain packet crashed the
proxy-html module (Paket went into the proxy, no paket came out;
according to the log, a process crashed at precisely that time...)

I emailed the developer, who advised me to try the most recent version
(3.1.2) for which unfortunately no debian package exists. So I spent
half a day figuring out how to update the source debian archive with
the most recent source code. I managed it, installed it, and it
worked, and the module did not crash anymore. :-)

However, after continuing to play around with mod-proxy-html I noticed
that the output stays as utf-8 (having been converted from ISO-8859-1)
instead of being converted back to ISO-8859-1.
Note that 3.0.0 worked with precisely the same configuration (except
that it did not need xml2enc).
I am loading xml2enc, and mod-proxy-html notices that I do: If I
comment the line "LoadModule xml2enc_module
/usr/lib/apache2/modules/mod_xml2enc.so" out, I get Lines "[Wed Oct 28
10:42:04 2009] [warn] [client 192.168.0.9] No i18n support found.
Install mod_xml2enc if required , referer: ...".
Loading the module xml2enc, I am not getting error messages of any
kind in the apache log file.

If I turn on verbose Logging, and set log level to info, I get Lines
"[Wed Oct 28 11:04:22 2009] [info] [client 192.168.0.9] H: matched
/dwhfe/, substituting /testdwh/, referer: ...", "[Wed Oct 28 11:04:30
2009] [info] [client 192.168.0.9] No content-type; bailing out of
proxy-html filter, referer: ..." and similar entries, so the logging
works.
However, if I grep the file for "charset" (cat
/var/log/apache2/error.log | grep -i "charset") I don't get any
matches, although according to the documention I should.

I have tried changing the value of ProxyHTMLCharsetOut from * to
ISO-8859-1, and to set "ProxyHTMLMeta On". No change.

Unless I have compiled the module incorrectly (I have used the command
"apxs2 -c -I/usr/include/libxml2 mod_xml2enc.c") I conclude that the
output conversion just plainly is not working anymore.

If I have compiled it incorrectly, I would appreciate being given the
proper command to use, as I could not find any documentation to do so
(thus I adjusted the command the debian module used to build
mod_html_proxy).

Any help or suggestions would be much appreciated. :-(

Martin