如何生成可靠的逐字书目数据

Question

我正在尝试.bib从 git 更改日志生成一个 形状的“biblatex”数据库。典型的(?)项目如下所示：

@gitcommit{f1750d6,
 author = {Brent Longborough},
 date = {2013-12-15},
 title = {Complete name changes
Decomplicate top-level PROCs (&VERSION -> &CONFIG)
},

title该 MWE 生成的字段可以包含任何事物，并且我需要避免任何额外的外部命令（即不允许使用 sed、awk）：

\documentclass{article} 
\begin{document}
{\catcode`\%=12
\immediate\write18{
 git --no-pager log --reverse --pretty='format:@gitcommit{%h,%n 
 author = {%an},%n 
 date = {%ad},%n 
 title = {%B},%n
 commithash = {%H} 
 }' --date=short > \jobname.gll}
}
\end{document}

我应该尝试生成什么来代替title = {xxxx},，以及如何将任何适当的 TeX 命令注入到\write18生成的git命令行中？

Answer 1

biblatex具有恰当命名的字段类型verbatim，可能适合此处的要求。文档州（第 14 页）

逐字字段以逐字模式处理，可能包含特殊字符。逐字字段的典型示例是file和doi。

所以我们需要做的title就是verbatim通过

\DeclareDatamodelFields[type=field,datatype=verbatim]{
  title,
}

我们可以出发了。

当然，有些东西甚至连verbatim字段都无法忍受，到目前为止，我只能想到不匹配的花括号，但可能还有更多。不过，我们不需要担心大多数其他事情。s%和\s 完全没问题。

MWE（cave filecontentsem：这将覆盖gitlog.dbx，gitlog.bbx和\jobname.gll）

\documentclass{article} 
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{filecontents}
\begin{filecontents*}{gitlog.dbx}
\DeclareDatamodelEntrytypes{gitcommit}
\DeclareDatamodelFields[type=field,datatype=literal]{
  commithash,
}
\DeclareDatamodelFields[type=field,datatype=verbatim]{
  title,
}
\DeclareDatamodelEntryfields[gitcommit]{
  title,
  author,
  date,
  commithash,
}
\end{filecontents*}

\begin{filecontents*}{gitlog.bbx}
\ProvidesFile{gitlog.bbx}[gitlog bibliography file 2015/11/17]

\defbibenvironment{bibliography}
  {\list
     {\printfield{entrykey}}
     {\setlength{\labelwidth}{0pt}%
      \setlength{\itemindent}{-\leftmargin}%
      \setlength{\itemsep}{\bibitemsep}%
      \setlength{\parsep}{\bibparsep}%
      \renewcommand*{\makelabel}[1]{%
    \hspace\labelsep\ttfamily##1}}}
  {\endlist}
  {\item}

\newbibmacro*{begentry}{}
\newbibmacro*{finentry}{\finentry}

\DeclareBibliographyDriver{gitcommit}{%
  \usebibmacro{bibindex}%
  \usebibmacro{begentry}%
  \usebibmacro{title}%
  \newblock%
  \usebibmacro{author}%
  \newunit
  \usebibmacro{date}%
  \newunit
  \usebibmacro{commithash}
  \usebibmacro{finentry}}

\newbibmacro*{commithash}{%
    \printtext{%
       \printfield{commithash}%
}}

\endinput
\end{filecontents*}

\usepackage[%
    datamodel=gitlog,
    bibstyle=gitlog,
    sorting=none,
    date=iso8601,
    firstinits=true,
]{biblatex}


\begin{filecontents*}{\jobname.gll}
@gitcommit{f1750d6,
 author = {Brent Longborough},
 date = {2013-12-15},
 title = {Complete name changes
Decomplicate top-level PROCs (&VERSION -> &CONFIG)
},
}
@gitcommit{9dce109,
 author = {Brent Longborough},
 date = {2015-11-15},
 title = {Initial commit &&&&&& %
},
 commithash = {9dce10970bb6be976ce59f76dd28e52abeb3b103} }
@gitcommit{f558cdd,
 author = {Brent Longborough},
 date = {2015-11-15},
 title = {First code to build pseudo-bibfile
},
 commithash = {f558cddc5a95acb1c081ead6f6e81cb9941941a0} }
@gitcommit{240f4c9,
 author = {Brent Longborough},
 date = {2015-11-15},
 title = {Added data model (not working) - now we step back a bit
},
 commithash = {240f4c9dbb22631f6098fbe45c409ef58cdddb09} }
@gitcommit{b4629f6,
 author = {Brent Longborough},
 date = {2015-11-16},
 title = {First working version full of restrictions
},
 commithash = {b4629f6bb7504b2d6e4a677294a761e1d1d5c659} }
@gitcommit{d651d24,
 author = {Brent Longborough},
 date = {2015-11-17},
 title = {Gradually improving things (especially understanding)
},
 commithash = {d651d249a3a14fd270a0f40e5d8124f6c552cc5b} }
@gitcommit{4c1dd2c,
 author = {Brent Longborough},
 date = {2015-11-17},
 title = {Making an MWE for TeX.SX},
 commithash = {4c1dd2c540ff8a2f2976d972d15071c1f54f9379},
verba = {4c1dd2c540ff8a2f2976d972d15071c1f54f9379} }
\end{filecontents*}

\addbibresource{\jobname.gll}
\begin{document}
\nocite{*}
\printbibliography[title={Change Log},type=gitcommit]
\end{document}