本文接上篇《cups-pdf编译安装》，内容都为本身根据项目经验总结，可以转载但请注明出处。作者简书id: weogkoii
开发环境：
Linux 4.4.0-31-generic #50~14.04.1-Ubuntu SMP Wed Jul 13 01:07:32 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
CUPS及周边安装请参考前篇：《CUPS-PDF编译安装》
https://www.jianshu.com/p/4e9b15fe9386
1. lp和lpr命令简介
这两个命令都是CUPS系统自带的打印命令，为了兼容老系统而保留下来的，也是使用起来最方便的。但是，在中文打印时经常出现乱码。命令参考：
https://www.cups.org/doc/man-lp.html
https://www.cups.org/doc/man-lpr.html
已安装好的cups-pdf打印机：
#lpstat -a
cups-pdf 自从 2019年08月02日 星期五 09时13分34秒 开始接受请求
基本用法：
lpr [-P 打印机] [-# 打印份数] -U [用户名] file
lpr -P cups-pdf /tmp/3333.txt
lp [-d 打印机] [-n 打印份数] file
lp -d cups-pdf /tmp/3333.txt
2. lpr命令打印中文txt文档
使用vim随意编辑一个中文txt文档，内容如下：
#cat /tmp/ChinesePrintTest.txt
一须abc是口
调用lpr或lp使用cups-pdf进行打印，得到pdf文件却是乱码：
lpr -P cups-pdf /tmp/ChinesePrintTest.txt
期望的”一须abc是口“没有出现，而是用空白方框代替了4个中文字符，只剩余中间的abc正常打印出来。
3. 解决乱码调试分析过程
初步分析，生成的pdf文档中只有位置中间的abc正常打印了，中文字符均为空白，有可能是不支持中文字符打印。于是上网google百度了很久，发现竟然都说lp/lpr命令不支持中文，要打印pdf的话，最好是用图形化Linux的软件。但是我有点不甘心，因为cups的文档中并未说明不支持中文，而且现在utf8格式都这么多年了，这些命令也应该与时俱进了吧。反正是编译安装的，索性自己看日志看代码尝试解决。
3.1 首先分析cups日志
查看/var/log/cups/error_log，本次打印是把txt文档转换成pdf文档，一共使用4个filters，分别是texttopdf、pdftopdf、pdftops、cups-pdf，4个filter的作用分别如日志所描述：
[Job 124] 4 filters for job:
[Job 124] texttopdf (text/plain to application/pdf, cost 32)
[Job 124] pdftopdf (application/pdf to application/vnd.cups-pdf, cost 66)
[Job 124] pdftops (application/vnd.cups-pdf to application/vnd.cups-postscript, cost 100)
[Job 124] - (application/vnd.cups-postscript to printer/cups-pdf, cost 0)
由于texttopdf是cups-filters项目编译而来的，所以直接查看相应代码。接下来模拟cups调用texttopdf（如果不熟悉这部分，可以去查看相关cups filters的原理文档）。从error_log中获取调用texttopdf的参数：
[Job 124] argv[0]="cups-pdf"
[Job 124] argv[1]="124"
[Job 124] argv[2]="root"
[Job 124] argv[3]="ChinesePrintTest.txt"
[Job 124] argv[4]="1"
[Job 124] argv[5]="finishings=3 number-up=1 job-uuid=urn:uuid:431814bd-dd50-3e6b-6fbc-ae490d3763a9 job-originating-host-name=localhost date-time-at-creati on= date-time-at-processing= time-at-creation=1564970267 time-at-processing=1564970267 document-name-supplied=ChinesePrintTest.txt"
[Job 124] argv[6]="/var/spool/cups/d00124-001"
环境变量：
[Job 124] envp[0]="CUPS_CACHEDIR=/var/cache/cups"
[Job 124] envp[1]="CUPS_DATADIR=/usr/share/cups"
[Job 124] envp[2]="CUPS_DOCROOT=/usr/share/doc/cups"
[Job 124] envp[3]="CUPS_FONTPATH=/usr/share/cups/fonts"
[Job 124] envp[4]="CUPS_REQUESTROOT=/var/spool/cups"
[Job 124] envp[5]="CUPS_SERVERBIN=/usr/lib/cups"
[Job 124] envp[6]="CUPS_SERVERROOT=/etc/cups"
[Job 124] envp[7]="CUPS_STATEDIR=/var/run/cups"
[Job 124] envp[8]="HOME=/var/spool/cups/tmp"
[Job 124] envp[9]="PATH=/usr/lib/cups/filter:/usr/bin:/usr/sbin:/bin:/usr/bin"
[Job 124] envp[10]="SERVER_ADMIN=root@ximi"
[Job 124] envp[11]="SOFTWARE=CUPS/2.2.9"
[Job 124] envp[12]="TMPDIR=/var/spool/cups/tmp"
[Job 124] envp[13]="USER=root"
[Job 124] envp[14]="CUPS_MAX_MESSAGE=2047"
[Job 124] envp[15]="CUPS_SERVER=/var/run/cups/cups.sock"
[Job 124] envp[16]="CUPS_ENCRYPTION=IfRequested"
[Job 124] envp[17]="IPP_PORT=631"
[Job 124] envp[18]="CHARSET=utf-8"
[Job 124] envp[19]="LANG=zh_CN.UTF-8"
[Job 124] envp[20]="PPD=/etc/cups/ppd/cups-pdf.ppd"
[Job 124] envp[21]="RIP_MAX_CACHE=128m"
[Job 124] envp[22]="CONTENT_TYPE=text/plain"
[Job 124] envp[23]="DEVICE_URI=cups-pdf:/"
[Job 124] envp[24]="PRINTER_INFO=cups-pdf"
[Job 124] envp[25]="PRINTER_LOCATION="
[Job 124] envp[26]="PRINTER=cups-pdf"
[Job 124] envp[27]="PRINTER_STATE_REASONS=none"
[Job 124] envp[28]="CUPS_FILETYPE=document"
[Job 124] envp[29]="FINAL_CONTENT_TYPE=application/vnd.cups-postscript"
[Job 124] envp[30]="AUTH_I****"
3.2 使用shell脚本模拟cups调用texttopdf，将stdout重写向到文件中，这个文件中就包括转换的pdf格式文件。查看其中内容：
6 stream
7 q
8 0 g
9 BT
10 18 745.884 Td
11 105.522 Tz
12 /FN02 11.333 Tf <00000000> Tj
13 ET
14 0 g
15 BT
16 46.800 745.884 Td
17 105.522 Tz
18 /FN00 11.333 Tf <004400450046> Tj
19 ET
20 0 g
21 BT
22 68.400 745.884 Td
23 105.522 Tz
24 /FN02 11.333 Tf <00000000> Tj
25 ET
26 Q
27 endstream
可以看到在stream中，原本的中文字符变成了<00000000>，只剩余<004400450046>（abc，ASCII码16进制）。从这里可以判断，应该是这一步转换出了问题，原来txt文件中的中文，在生成的pdf格式文件中已经被0000所代替了。找到了问题点，再继续分析出问题的原因。
3.3 查看txt文件格式和编码
:set ff
fileformat=unix
:set fileencoding
fileencoding=utf-8
经过查看，文件格式和编码都没有问题
3.4 查看texttopdf源码
为什么会把中文弄成空格呢？经过查看代码终于是有所发现，texttopdf没有找到中文字体？！为什么呢，vim都已经可以显示中文字符了，那为什么texttopdf却又找不到字体呢？
查看cups-filters的README发现一点线索：
TEXTTOPDF
=========
This implements a texttopdf filter, and is derived from cups' texttops.
To configure:
-------------
- texttopdf uses CUPS_DATADIR/charset/pdf.utf-8 for font configuration
(when utf-8 was requested as charset). The font names given there are
used as fontconfig selectors; the best matching font, that is both
monospaced and in a supported format (TTC, TTF or OTF) will then be used.
- As a special exception, all fontnames that start with a '.' or '/' are
considered filenames, and fontconfig is skipped; the name is used directly
for loading the font file.
- Implementation note: TrueType Collections (.TTC) are internally handled
by appending '/' and the index of the font inside the collection to
the filename (e.g. to use the second font of uming.ttc, the filename
uming.ttc/1 must be given to the fontembed-library).
By appending the index-field returned from fontconfig, this is completely
transparent to the user (but currently not widely tested).
- You may look at the two examples: pdf.utf-8.simple and pdf.utf-8.heavy.
To use:
-------
The filter is called just like any other cups filter. Have a
look at test.sh for example.
Known Issues
------------
- Text extraction does not work (at least for pdftotext from xpdf)
for the resulting pdfs.
- OTF(CFF) embedding currently does not subset the fonts.
- Text wrapping in pretty-printing mode does not respect double-wide
characters (CJK), and thus produce wrong results (wrap too late)
for lines where they occur. The fix is not trivial, since all the
pretty-printing processing is done without knowledge of / prior to
the font configuration (which is where single or double width
code-ranges are specified).
- The hebrew example in test5.pdf shows one of our limitations:
Compose glyphs are not composed with the primary glyph but printed
as separate glyphs.
原来texttopdf使用CUPS_DATADIR/charset/pdf.utf-8文件配置自身的字体和字符集，看样子是这个文件配置不对。
3.5 研究pdf.utf-8配置格式
这个配置文件是一段注释，但是说得不是特别清楚，还是根据代码反推才搞明白。原来texttopdf是读取fontconfig字体配置。需要texttopdf中使用的字体都需要在这个配置文件中添加，相应字体才能被加载到texttopdf中正常使用。打开文件一看，果然没有中文字体：
charset utf8
#
# This file defines the font mappings used for Unicode/UTF-8 text printing
# through PDF.
#
# Each line consists of:
#
# first last direction width normal bold italic bold-italic
#
# First and last are the first and last glyphs in the font mapping
# that correspond to that font; contrary to PostScript printing
# they only select the font. To find the glyph the complete unicode
# character will be looked up in the (3,1) resp. (3,0) cmap of the
# TrueType font. The glyph values are hexadecimal.
#
# Direction is the string "ltor" or "rtol", indicating left-to-right or
# right-to-left text.
#
# Width is the string "single" or "double"; double means that the glyphs
# are twice as wide as ASCII characters in the Courier typeface.
#
# "Normal", "bold", "italic", and "bold-italic" are the typefaces to use
# for each presentation. If characters are only available in a single
# style then only one typeface should be listed (e.g. "Symbol")
#
# Each font that is listed will be used (and downloaded if needed) when
# printing.
#
0000 04FF ltor single monospace monospace:bold monospace:oblique monospace:bold:oblique
0500 05FF rtol single monospace
“#”号开头的都是注释可以直接跳过。重点在最后面两排，表示utf8编码范围使用的对应字体。
3.5 fontconfig配置字体
用fc-list查看当前已经配置字体
/usr/share/fonts/truetype/dejavu/DejaVuSerif-Bold.ttf: DejaVu Serif:style=Bold
/usr/share/fonts/truetype/arphic/uming.ttc: AR PL UMing TW MBE:style=Light
/usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf: DejaVu Sans Mono:style=Book
/usr/share/fonts/truetype/dejavu/DejaVuSans.ttf: DejaVu Sans:style=Book
/usr/share/fonts/truetype/wqy/wqy-microhei.ttc: 文泉驿微米黑,文泉驛微米黑,WenQuanYi Micro Hei:style=Regular
/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf: DejaVu Sans:style=Bold
/usr/share/fonts/truetype/dejavu/DejaVuSansMono-Bold.ttf: DejaVu Sans Mono:style=Bold
/usr/share/fonts/truetype/arphic/uming.ttc: AR PL UMing TW:style=Light
/usr/share/fonts/truetype/arphic/uming.ttc: AR PL UMing CN:style=Light
/usr/share/fonts/truetype/arphic/uming.ttc: AR PL UMing HK:style=Light
/usr/share/fonts/truetype/wqy/wqy-microhei.ttc: 文泉驿等宽微米黑,文泉驛等寬微米黑,WenQuanYi Micro Hei Mono:style=Regular
/usr/share/fonts/truetype/dejavu/DejaVuSerif.ttf: DejaVu Serif:style=Book
3.6 修改pdf.utf-8配置文件
参考pdf.utf-8.heavy文件，配置中文字体，添加下面一行在配置文件中
3000 9FFF ltor double /usr/share/fonts/truetype/arphic/uming.ttc/0
7. 成功打印中文
再次执行打印，查看pdf格式文件：
6 stream
7 q
8 0 g
9 BT
10 18 745.884 Td
11 127.059 Tz
12 /FN02 11.333 Tf <086149ba> Tj
13 ET
14 0 g
15 BT
16 46.800 745.884 Td
17 105.522 Tz
18 /FN00 11.333 Tf <004400450046> Tj
19 ET
20 0 g
21 BT
22 68.400 745.884 Td
23 127.059 Tz
24 /FN02 11.333 Tf <1dad0e10> Tj
25 ET
26 Q
27 endstream
这一次，相应的中文编码不再是0000了，变成实际对应的utf-8编码。再查看打印出来的文件：
终于使用lp/lpr命令成功打印中文了！
