使用 sed 或 perl 在每个否定环视的特定行添加反斜杠

Question

在我们高效的 Splunk 环境中，我们在一个特殊的应用程序中拥有一百多个自行开发的用例，我们希望将其进一步导出到专用于其他客户等的集群。 Savedsearches.conf 分为默认文件夹和本地文件夹，我们希望通过

splunk btool --app=our_fancy_usecases savedsearches list > searches.conf

预先合并这些文件夹。由此产生的配置文件将如下所示：

[Stanza #1]
some.config = some settings
description = just a few words
some.other.config = more settings
search = this is 
| a well
| formatted search
[Stanza #2]
some.config = 1
description = This is taking 
a lot of words
to explain
this time.
some.other.config = 2
search = a lot 
of searching 
to do
[This is yet another stanza]
the.end = nigh

该 btool 的问题是丢失了原始配置文件中的反斜杠（“”），这些反斜杠显示特定行实际上是多行值。每个设置都可以是多行值，但不一定需要如此。当反斜杠丢失时，Splunk 将仅读取该设置的配置，直到行结束，即，对于第一个节，Splunk 中的搜索最终将仅为

this is

，缺少以下两行。也就是说，配置文件应该如下所示：

[Stanza #1]
some.config = some settings
description = just a few words
some.other.config = more settings
search = this is \
| a well\
| formatted search
[Stanza #2]
some.config = 1
description = This is taking \
a lot of words\
to explain\
this time.
some.other.config = 2
search = a lot \
of searching \
to do
[This is yet another stanza]
the.end = nigh

遗憾的是，btool 缺乏保留这些反斜杠的功能，所以我试图用 sed 或任何可以完成这项工作的工具将它们恢复。

我首先研究如何使用正则表达式获取需要更改的行。节始终包含在方括号

[

]

中，并且设置始终在

=

后面有一个空格，在实际值之前还有另一个空格。我想出了这个负面的看法：

\n(?!^[A-Za-z0-9._]+ = .*$|^\[.*\])

对于顶部的配置示例，这似乎很好地抓住了我想在 regex101 中更改的内容。然而#sed 不允许环视。 #Perl 应该是一个选择，但这就是我失败的地方：

perl -pe 's/\n(?![A-Za-z0-9._]+ = .*$|\[.*\]$)/\\\n/' test.conf > test.out ; cat test.out

[Stanza #1]\
some.config = some settings\
description = just a few words\
some.other.config = more settings\
search = this is\
| a well\
| formatted search\
[Stanza #2]\
some.config = 1\
description = This is taking\
a lot of words\
to explain\
this time.\
some.other.config = 2\
search = a lot\
of searching\
to do\
[This is yet another stanza]\
the.end = nigh\

遗憾的是，这根本无法按预期工作，只是在每一行添加一个反斜杠。我也用#sed 摆弄了一下：

sed '/^[A-Za-z0-9._]\+ = .*$\|^\[.*\]$/! s/.*/&\\/' test.conf > test.out ; cat test.out
[Stanza #1]
some.config = some settings
description = just a few words
some.other.config = more settings
search = this is
| a well\
| formatted search\
[Stanza #2]
some.config = 1
description = This is taking
a lot of words\
to explain\
this time.\
some.other.config = 2
search = a lot
of searching\
to do\
[This is yet another stanza]
the.end = nigh

关于如何实现我的目标有什么想法吗？

Answer 1

perl  -ne 'chomp;
           print "\\" if ! / = |^\[/;
           print "\n" if $. != 1;
           print;
           END { print "\n" }' < file

• -n

  逐行读取输入并运行每行代码；
• chomp 从输入中删除最后的换行符；
• 如果“下”行不是设置或节，则打印反斜杠；
• 除了第一行之前打印换行符；
• 打印“下一行”；
• 最后打印换行符。