ClearCase：查找仅具有一个特定标签而不是更多标签的文件

发布于 2024-07-25 18:43:05 字数 343 浏览 3 评论 0原文

我想在 ClearCase 中查找标有特定标签但没有设置任何其他标签的文件。

例如，如果我有这样标记的文件：

file1 LBL_A, LBL_B
file2 LBL_A

我希望有一个查询只提供 file2 而不是 file1。

有没有办法用cleartool find 来做到这一点？如果这不可能通过单个查询来完成，我也很乐意了解如何分几步完成此操作的任何想法（我将从 perl 脚本调用cleartool，因此可以轻松保存文件列表暂时并对其运行进一步的命令）。

预先非常感谢！

扬

原文

I'd like to find files in ClearCase that are labeled with a specific label but that do not have any other labels set.

For example, if I have files labeled like this:

file1 LBL_A, LBL_B
file2 LBL_A

I'd like to have a query that gives me just file2 and not file1.

Is there a way to do this with cleartool find? If this is not possible to do with a single query, I'd also be happy for any ideas how to do this in several steps (I'll be calling cleartool from a perl script, so it will be easy to save lists of files temporarily and run further commands on them).

Thanks a lot in advance!

Jan

分享到QQ

分享到微博

如果你对这篇内容有疑问，欢迎到本站社区发帖提问参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。

发布评论

需要登录才能够评论，你可以免费注册一个本站的账号。

歌入人心 2024-08-01 18:43:05

假设 LBL_A 是您想要运行的（唯一）标签，

cleartool find /some/dir -version 'lbtype(LBL_A)' -print | xargs cleartool describe -fmt "%n: %l"

应将其作为输出提供

file1: (LBL_A, LBL_B)
file2: (LBL_A)

，然后您可以在 Perl 脚本中检查或通过 sed -n 's/$.*$: (LBL_A)/\1/ 进行过滤p'（假设文件名中没有冒号）。

更新：正如 VonC 正确指出的那样，上面的命令对于包含空格的文件将失败。要处理该问题，请运行为：

cleartool find ... -print | tr '\012' '\000' | xargs -0 cleartool ....

它将换行符转换为 ascii null，然后让 xargs 使用它作为分隔符。

Assuming LBL_A is the (only) label you want running

cleartool find /some/dir -version 'lbtype(LBL_A)' -print | xargs cleartool describe -fmt "%n: %l"

should give

file1: (LBL_A, LBL_B)
file2: (LBL_A)

as output which you then can check in your perl script or filter through sed -n 's/$.*$: (LBL_A)/\1/p' (assuming no colons in filenames).

Update: As VonC correctly points out, the command above will fail for files with spaces in. To handle that run as:

cleartool find ... -print | tr '\012' '\000' | xargs -0 cleartool ....

which will translate newlines into ascii null and then have xargs use that as delimiter.

回复收藏 0 原文

じее 2024-08-01 18:43:05

您可以直接执行此操作，而无需通过管道：

cleartool find . -ver "lbtype(LBL_A) && !lbtype(LBL_B)" -print

You can do this directly without having to pipe:

cleartool find . -ver "lbtype(LBL_A) && !lbtype(LBL_B)" -print

回复收藏 0 原文

顾忌 2024-08-01 18:43:05

赫洛夫达尔的回答说明，在这种情况下，您必须找到比您需要的更多的元素，然后过滤它们（尽管所有 ClearCase 查找功能）。

注意：

 cleartool find . -type f -element 'lbtype_sub(LBL_A)' -print

可以直接为您提供元素（而不是在这种情况下可能不感兴趣的版本）。您可以使用 -version 和格式指令找到版本，如下所述。

在 fmt_ccase，您可以定制输出以精确获取 Perl 脚本需要继续执行的内容：

 cleartool find . -type f -element 'lbtype_sub(LBL_A)' -exec 'cleartool describe -fmt "%En %Cl\n" \"$CLEARCASE_XPN\"' | grep -v ","

-fmt "%En %l\n" 将显示元素的完整路径（而不是版本：/a/b/myFile@@/main/myVersion）=> /a/b/myFile'. '\n` 确保每行一个结果。
-version 和 -fmt "%n %l\n" 将显示版本
\"$CLEARCASE_XPN\"：双引号围绕找到的版本的扩展路径，确保名称中带有空格的文件仍然有效。
grep -v ","：如果有逗号，则表示“多个标签”
%Cl：避免显示全部列表的标签。无论如何，如果不止一个，你就不感兴趣！

因此，要查找确切的版本：

 cleartool find . -type f -version 'lbtype_sub(LBL_A)' -exec 'cleartool describe -fmt "%n %Cl\n" \"$CLEARCASE_XPN\"' | grep -v ","|awk '{sub(/ \(.*/,"");print}'

注意：
上面的代码适用于 unix 语法。 Windows 语法为：

cleartool find . -type f -element "lbtype(LBL_A)" -exec "cleartool describe -fmt \"%n %Cl\n\" \"%CLEARCASE_XPN%\"" | grep -v "," | gawk "{gsub(/ \(.*,"");print}"

，它将列出仅具有一个（正确）标签的文件的版本。

awk '{sub(/ \(.*/,"");print}' 会将“myFile@@/main/myVersion (LBL_A)”转换为“myFile @@/main/myVersion"

hlovdal's answer illustrates that, in this case, you have to find more elements than you need, and then filter them (despite all the ClearCase find features).

Note:

 cleartool find . -type f -element 'lbtype_sub(LBL_A)' -print

could give you directly the elements (and not the version which may not be interesting in this case). You can find the version with -version and a format directive as explained below.

With the help of fmt_ccase, you can tailor the output to get precisely what your perl script will need to go on:

 cleartool find . -type f -element 'lbtype_sub(LBL_A)' -exec 'cleartool describe -fmt "%En %Cl\n" \"$CLEARCASE_XPN\"' | grep -v ","

-fmt "%En %l\n" will display the full path of the element (instead of the version: /a/b/myFile@@/main/myVersion) => /a/b/myFile'. the '\n` ensure one result per line.
-version and -fmt "%n %l\n" would display the version
\"$CLEARCASE_XPN\": the double quotes arount the extended path of the version found ensure a file with spaces in its name will still work.
grep -v ",": if there is any comma, that means "more than one label"
%Cl: avoid displaying the all list of labels. Anyway, if there are more than one, you are not interested!

So, for finding the exact version:

 cleartool find . -type f -version 'lbtype_sub(LBL_A)' -exec 'cleartool describe -fmt "%n %Cl\n" \"$CLEARCASE_XPN\"' | grep -v ","|awk '{sub(/ \(.*/,"");print}'

Note:
The above works with unix syntax. The windows syntax would be:

cleartool find . -type f -element "lbtype(LBL_A)" -exec "cleartool describe -fmt \"%n %Cl\n\" \"%CLEARCASE_XPN%\"" | grep -v "," | gawk "{gsub(/ \(.*,"");print}"

, which would list the version of files with only one (correct) label.