【基本解决】docbook中使用catalog.xml时出错：Resolve: sysID xxx, Resolve URI xxx, warning: failed to load external entity

【问题】

折腾docbook，给其添加catalog的支持，以方便映射本地的xsl为一个简单的名称，方便其他地方调用。

但是使用catalog的过程中结果出错：

CLi@PC-CLI-1 ~/develop/docbook/books/VBR/VBR/src
$ XML_CATALOG_FILES="/home/CLi/develop/docbook/config/catalog/catalog.xml"    XML_DEBUG_CATALOG=1    xsltproc.exe --xinclude -o ../output/html/MPEG_VBR.html docbook_html_crl.xsl     MPEG_VBR.xml
Resolve: sysID docbook_html_crl.xsl
-2147483592 Parsing catalog /home/CLi/develop/docbook/config/catalog/catalog.xml
/home/CLi/develop/docbook/config/catalog/catalog.xml:0: element uri: Catalog error : uri entry lacks 'name'
/home/CLi/develop/docbook/config/catalog/catalog.xml added to file hash
Resolve URI docbook_html_crl.xsl
Found URI match docbook_html_crl.xsl
Resolve: sysID /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
Resolve URI /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
warning: failed to load external entity "/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl"
compilation error: file /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl line 17 element import
xsl:import : unable to load /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
Catalogs cleanup
Free catalog entry crl.ent
Free catalog entry docbook_fo.xsl
Free catalog entry docbook_html_crl.xsl
Free catalog entry docbook_fo_crl.xsl
Free catalog entry /home/CLi/develop/docbook/config/catalog/catalog.xml
Free catalog entry

相关的catalog.xml内容为：

<?xml version="1.0" encoding="UTF-8"?>

<!DOCTYPE catalog
   PUBLIC "-//OASIS/DTD Entity Resolution XML Catalog V1.0//EN"
   "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd"
[

<!ENTITY config_dir "/home/CLi/develop/docbook/config">
<!ENTITY config_xls_ns_dir "&config_dir;/docbook-xsl-ns-1.76.1">
<!ENTITY xsl_ns_base_dir "/home/CLi/develop/docbook/tools/docbook-xsl-ns-1.76.1">

<!-- <!ENTITY xsl_ns_base_dir "/cygdrive/e/Dev_Root/docbook/tools/docbook-xsl-ns-1.76.1"> -->

]
>

<catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
    <public publicId="crl.ent"
        uri="&config_dir;/entity/crl.ent"/>
        
    <uri publicId="docbook_html.xsl"
        uri="&xsl_ns_base_dir;/html/docbook.xsl"/>

    <uri name="docbook_fo.xsl"
        uri="&xsl_ns_base_dir;/fo/docbook.xsl"/>

    <uri name="docbook_html_crl.xsl"
        uri="&config_xls_ns_dir;/html/docbook_crl.xsl"/>

    <uri name="docbook_fo_crl.xsl"
        uri="&config_xls_ns_dir;/fo/docbook_crl.xsl"/>
</catalog>

【解决过程】

1.其中，命令中的docbook_html_crl.xsl，是catalog中配置的：

    <uri name="docbook_html_crl.xsl"
        uri="&config_xls_ns_dir;/html/docbook_crl.xsl"/>

可以看出，是可以被正确解析的。

但是，另外一个xsl配置文件docbook_crl.xsl中所引用的：

<xsl:import href="docbook_html.xsl"/>

却无法正确解析，显示错误：

warning: failed to load external entity "/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl"

所以，开始是怀疑，用catalog所定义的“别名”，只能在makefile，命令行等环境中使用，而无法在xsl等文件中使用。

但是却又看到这里：DocBook使用笔记，中说：

只需这样写：

<xsl:import href="html/chunk.xsl"/>

而不用写成繁琐的

<xsl:import href="/usr/share/sgml/docbook/xsl-stylesheets-1.69.1-5.1/html/chunk.xsl"/>

这样做的好处是，可以将样式表的安装路径统一在catalog.xml中规定，不用在每个xsl中重复书写完整路径了。 以后万一要升级样式表，也只需修改catalog.xml中的定义即可，不用去每个xsl中修改。

意思就是这样用的，但是我这里却始终不可以。不知道为何。

2.对于catalog中：

    <public publicId="crl.ent"
        uri="&config_dir;/entity/crl.ent"/>

在别的某个xml文件中，是可以引用成功的：

<!DOCTYPE chapter 
[
<!-- <!ENTITY % crl_ent SYSTEM "../../../../config/entity/crl.ent">%crl_ent; -->
<!ENTITY % crl_ent PUBLIC "crl.ent" 'null'>%crl_ent;
]
>

所以，想到，参考其写法，把docbook_html.xsl的定义换成：

    <public publicId="docbook_html.xsl"
        uri="&xsl_ns_base_dir;/html/docbook.xsl"/>
        

然后在docbook_crl.xsl中引用：

<!DOCTYPE stylesheet [
<!ENTITY % docbook_html PUBLIC "docbook_html.xsl" 'null'>%docbook_html;
]>

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
                version='1.0'>
<xsl:import href="&docbook_html;"/>
</xsl:stylesheet> 

结果出错：

CLi@PC-CLI-1 ~/develop/docbook/books/VBR/VBR/src
$ XML_CATALOG_FILES="/home/CLi/develop/docbook/config/catalog/catalog.xml"    XML_DEBUG_CATALOG=1    xsltproc.exe --xinclude -o ../output/html/MPEG_VBR.html docbook_html_crl.xsl MPEG_VBR.xml
Resolve: sysID docbook_html_crl.xsl
-2147483592 Parsing catalog /home/CLi/develop/docbook/config/catalog/catalog.xml
/home/CLi/develop/docbook/config/catalog/catalog.xml:0: element uri: Catalog error : uri entry lacks 'name'
/home/CLi/develop/docbook/config/catalog/catalog.xml added to file hash
Resolve URI docbook_html_crl.xsl
Found URI match docbook_html_crl.xsl
Resolve: pubID docbook_html.xsl sysID /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/null
Resolve URI /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/null
/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl:6: warning: failed to load external entity "/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/null"
<!ENTITY % docbook_html PUBLIC "docbook_html.xsl" 'null'>%docbook_html;
                                                                       ^
Entity: line 1:
 %docbook_html;
               ^
/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl:20: parser error : Entity 'docbook_html' not defined
<xsl:import href="&docbook_html;"/>
                                ^
compilation error: file /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl line 20 element import
xsl:import : recursion detected on imported URL /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl
Catalogs cleanup
Free catalog entry crl.ent
Free catalog entry docbook_fo.xsl
Free catalog entry docbook_html_crl.xsl
Free catalog entry docbook_fo_crl.xsl
Free catalog entry /home/CLi/develop/docbook/config/catalog/catalog.xml
Free catalog entry

所以，还是搞不懂，如何实现在别的xsl文件中，引用catalog定义的某个xsl文件的别名。

3.参考：Failed to load external entity "xhtml_onechunk"，去试了试xmllint.：

CLi@PC-CLI-1 ~/develop/docbook/books/VBR/VBR/src
$ xmllint.exe --valid --noout /home/CLi/develop/docbook/config/catalog/catalog.xml

说明我的catalog是valid的。

4.后来参考：Re: [xslt] [xml] Problem with xsltproc on Windows，把xsl_ns_base_dir去添加file://前缀：

<!ENTITY xsl_ns_base_dir "file:///home/CLi/develop/docbook/tools/docbook-xsl-ns-1.76.1">

结果也还是同样错误。

5.试了试把默认的/etc/xml/catalog也加上，结果还是出错：

CLi@PC-CLI-1 ~/develop/docbook/books/VBR/VBR/src
$ XML_CATALOG_FILES="/etc/xml/catalog /home/CLi/develop/docbook/config/catalog/catalog.xml" \
   xsltproc.exe --xinclude -o ../output/html/MPEG_VBR.html docbook_html_crl.xsl MPEG_VBR.xml>    XML_DEBUG_CATALOG=1 \
>    xsltproc.exe --xinclude -o ../output/html/MPEG_VBR.html docbook_html_crl.xsl MPEG_VBR.xml
Resolve: sysID docbook_html_crl.xsl
-2147483592 Parsing catalog /etc/xml/catalog
/etc/xml/catalog added to file hash
/home/CLi/develop/docbook/config/catalog/catalog.xml not found in file hash
-2147483592 Parsing catalog /home/CLi/develop/docbook/config/catalog/catalog.xml
/home/CLi/develop/docbook/config/catalog/catalog.xml added to file hash
Resolve URI docbook_html_crl.xsl
Found URI match docbook_html_crl.xsl
Resolve: sysID /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
Resolve URI /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
warning: failed to load external entity "/home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl"
compilation error: file /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_crl.xsl line 19 element import
xsl:import : unable to load /home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl
Catalogs cleanup
Free catalog entry -//OASIS//ENTITIES DocBook
Free catalog entry -//OASIS//ELEMENTS DocBook
Free catalog entry -//OASIS//DTD DocBook
Free catalog entry http://www.oasis-open.org/docbook/
Free catalog entry http://www.oasis-open.org/docbook/
Free catalog entry ISO 8879:1986
Free catalog entry http://docbook.sourceforge.net/release/xsl/
Free catalog entry http://docbook.sourceforge.net/release/xsl/
Free catalog entry http://docbook.sourceforge.net/release/xsl-ns/
Free catalog entry http://docbook.sourceforge.net/release/xsl-ns/
Free catalog entry http://glade.gnome.org/glade-2.0.dtd
Free catalog entry /etc/xml/catalog
Free catalog entry crl.ent
Free catalog entry docbook_html.xsl
Free catalog entry docbook_fo.xsl
Free catalog entry docbook_html_crl.xsl
Free catalog entry docbook_fo_crl.xsl
Free catalog entry /home/CLi/develop/docbook/config/catalog/catalog.xml
Free catalog entry
Free catalog entry

6.折腾了半天，终于想到了一个办法来验证docbook_html.xsl是否可以被正确解析。

通过命令：

CLi@PC-CLI-1 ~/develop/docbook/books/VBR/VBR/src
$ XML_CATALOG_FILES="/home/CLi/develop/docbook/config/catalog/catalog.xml"    XML_DEBUG_CATALOG=1    xsltproc.exe --xinclude -o ../output/html/MPEG_VBR.html docbook_html.xsl MPEG_VBR.xml
Resolve: sysID docbook_html.xsl
-2147483592 Parsing catalog /home/CLi/develop/docbook/config/catalog/catalog.xml
/home/CLi/develop/docbook/config/catalog/catalog.xml added to file hash
Resolve URI docbook_html.xsl
Found URI match docbook_html.xsl
Resolve: pubID crl.ent sysID null
Found public match crl.ent
Catalogs cleanup
Free catalog entry crl.ent
Free catalog entry docbook_html.xsl
Free catalog entry docbook_fo.xsl
Free catalog entry docbook_html_crl.xsl
Free catalog entry docbook_fo_crl.xsl
Free catalog entry /home/CLi/develop/docbook/config/catalog/catalog.xml
Free catalog entry

就验证了，docbook_html.xsl，是可以在命令行的环境中正确解析的。

而这点，和DocBook 5 快速起步教程 中的解释也是一致的，即catalog中定义的别名，用于xsltproc所使用。

而之前的错误证明了，catalog中的xsl文件别名，是没法在别的xsl文件中引用的，至少是没法被xsl:import的。

【总结】

目前的理解是，catalog中定义的别名，可以在命令行环境中（给xsltproc，makefile，脚本等）所使用。

但是无法用于别的xsl中，即别的xsl中调用该文件别名，是找不到的。默认是就变成了当前路径去找该文件了。

比如我此处的，catalog中定义了：

docbook_html.xsl = file:///home/CLi/develop/docbook/tools/docbook-xsl-ns-1.76.1/html/docbook.xsl

而在xsltproc中是可以被识别的，但是在另外某个xsl文件docbook_crl.xsl中引用：

<xsl:import href="docbook_html.xsl"/>

就会找不到，被识别为docbook_crl.xsl所在的当前路径下面找该文件：

file:///home/CLi/develop/docbook/config/docbook-xsl-ns-1.76.1/html/docbook_html.xsl

因此才会找不到。

解决办法很简单，那就是xsl中，还是用之前的绝对路径去引用对应的xsl文件，而不能调用catalog中的别名。

【后记】

又找到的权威的解释：

1.其中，上述可以通过xsl:import的话，是可以的，但是是针对Xalan的：

14.Using catalogs with Xalan

2.想要实现去除hardcode，即上述的用（短的）别名代替实际路径中的文件的，是不可以的：

9. Hardcoded import paths

所以，看来还是没法很方便的实现对于xsl中import导入的文件，用短路径或别名的。

【后记2】

另外，之前就参考了：

Using catalogs with xsltproc

刚刚看到这句：

The catalog processing mechanism in xsltproc will properly handle short references such as fo/docbook.xsl for the stylesheet name on the command line.

那就的确说明了，xsltproc使用catalog的话，那是只支持命令行中的别名替换和解析的，而不支持别的xsl文件中所引用的xsl文件的别名的。

所以，我甚至去试了给catalog添加rewriteURI：

    <group  xml:base="file:///home/CLi/develop/docbook/tools/docbook-xsl-ns-1.76.1/" >
        <rewriteURI  uriStartString="/tool/xsl_ns/html/"  
                    rewritePrefix="html/" />
    
        <uri name="docbook_html.xsl"
            uri="html/docbook.xsl" />
                        
        <uri name="docbook_fo.xsl"
            uri="fo/docbook.xsl"/>
    </group>

然后在docbook_crl.xsl中引用：

<xsl:import href="/tool/xsl_ns/html/docbook.xsl"/>

结果也还是无法成功的。

前后的实验结果，证实了，xsltproc中，是不支持xsl文件中引用catalog中的别名的。

转载请注明：在路上 » 【基本解决】docbook中使用catalog.xml时出错：Resolve: sysID xxx, Resolve URI xxx, warning: failed to load external entity