包括文件并缩进每一行

时间:2019-02-21 02:58:32

标签: m4

我想包含一个文件,并缩进所有行。我希望它成为降价文档中的代码块。

基本上我想要这样的东西:

text
include(somefile)
text

具有以下输出:

text
    content
    from
    somefile
text

我翻阅了manual,发现了patsubst。我也找到了这个问题和答案:How to indent a block of text in an m4 macro

适用于不包含逗号的文件:

$ cat textnocomma 
some text.
with sentences.
and no commas.

$ cat patsubstincludenocomma
text
patsubst(include(textnocomma), `^', `    ')
text

$ m4 patsubstincludenocomma
text
    some text.
    with sentences.
    and no commas.

text

但是当我include包含逗号的文件时:

$ cat textcomma 
some text.
with sentences.
and, commas.

$ cat patsubstincludecomma
text
patsubst(include(textcomma), `^', `    ')
text

$ m4 patsubstincludecomma
text
m4:patsubstincludecommanoquote:3: Warning: excess arguments to builtin `patsubst' ignored
some text.
with sentences.
and
text

问题似乎是m4进行宏扩展的幼稚方式。包含的文本中的逗号被解释为patsubst宏的语法。解决方案(应该是简单的):引用包含的文本。

但是,如果我引用include,则仅缩进第一行:

$ cat patsubstincludecommaquote 
text
patsubst(`include(textcomma)', `^', `    ')
text

$ m4 patsubstincludecommaquote
text
    some text.
with sentences.
and, commas.

text

我尝试了引号和文字换行符的不同组合,而不是正则表达式换行符。但是到目前为止,我所得到的只是excess arguments错误消息,或者只是缩进的第一行。

如何包含带有逗号的文本以及其他m4语法,并使其缩进m4?

2 个答案:

答案 0 :(得分:2)

为什么不使用外部命令?

esyscmd(`sed "s,^,    ," textcomma')

答案 1 :(得分:2)

我已经进行了一些研究,可以得出这样的结论:用引号include使patsubst在文本的第一行之外不再识别^(行首)。至少在我的系统上。

$ m4 --version
m4 (GNU M4) 1.4.18
...

观察:

$ cat textnocomma 
some text.
with sentences.
and no commas.

$ cat includenocomma 
foo
patsubst(include(textnocomma), `^', `    ')
bar
patsubst(`include(textnocomma)', `^', `    ')
baz

$ m4 includenocomma 
foo
    some text.
    with sentences.
    and no commas.

bar
    some text.
with sentences.
and no commas.

baz

在将文本定义为“字符串文字”而不是include时也会发生这种情况:

$ cat definestringliterals 
define(`sometext', `first line
second line
third line')dnl
foo
patsubst(sometext, `^', `    ')
bar
patsubst(`sometext', `^', `    ')
baz

$ m4 definestringliterals 
foo
    first line
    second line
    third line
bar
    first line
second line
third line
baz

以下是支持该观察的问答:How to match newlines in GNU M4 _properly_

奇怪的是,如果将字符串文字直接放在patsubst中,则不会发生这种情况:

$ cat patsubststringliterals 
foo
patsubst(first line
second line
third line, `^', `    ')
bar
patsubst(`first line
second line
third line', `^', `    ')
baz

$ m4 patsubststringliterals 
foo
    first line
    second line
    third line
bar
    first line
    second line
    third line
baz

使用括号“引用”文本不存在此问题。但现在我的文字周围有括号:

$ cat textcomma 
some text.
with sentences.
and, commas.

$ cat includecomma 
foo
patsubst(include(textcomma), `^', `    ')
bar
patsubst(`include(textcomma)', `^', `    ')
baz
patsubst((include(textcomma)), `^', `    ')
qux

$ m4 includecomma 
foo
m4:includecomma:2: Warning: excess arguments to builtin `patsubst' ignored
some text.
with sentences.
and
bar
    some text.
with sentences.
and, commas.

baz
    (some text.
    with sentences.
    and, commas.
    )
qux

所以我想这是一个错误。如果引用include,则patsubst将不再识别第一行中的^。但是,必须使用引号include来防止文本中的逗号被解释为语法。