Regex matching summary blocks in C#

Regex matching summary blocks in C#

[Samvid Mistry]

Hey,

I am trying to write an overlay that will simplify the large `<summary>`
blocks around comments in C# and just keep the content. For example,

```
/// <summary>
/// Gets the value of the counter.
/// </summary>
```

will get simplified to just

```
// Gets the value of the counter.
```

I have written a regex that can match the summary blocks, that is

"\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(.*\n\\)*?\\s-*///\\s-*</summary>"

The first group captures the indentation so that the overlay is indented
properly. The second group is supposed to capture the text of the block.
It works fine for single line blocks but for block spanning multiple
lines, the second capture group only captures the last line of comment.
The remaining lines get captured by `///\\s-*<summary>\\s-*\n` somehow.

For example, in

```
    /// <summary>
    /// Initializes a new instance of the
    /// <see cref="Service"/>
    /// class.
    /// </summary>
```

the second capture group only captures ` /// class.` and all lines above
it are captured by the former part of the regex. Not sure what I am
doing wrong here. I haven't put any wildcards that will capture anything
other than whitespace around summary tag. Appreciate any help with
this.

PS: This is my first time using any mailing list. Let me know if I
should be following any conventions while posting.

-- 
Thanks,
Samvid

Re: Regex matching summary blocks in C#

[Yuri Khan]

On Sun, 3 Dec 2023 at 19:20, Samvid Mistry <mistrysamvid@gmail.com> wrote:

> I have written a regex that can match the summary blocks, that is
>
> "\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(.*\n\\)*?\\s-*///\\s-*</summary>"
>
> The first group captures the indentation so that the overlay is indented
> properly. The second group is supposed to capture the text of the block.
> It works fine for single line blocks but for block spanning multiple
> lines, the second capture group only captures the last line of comment.

You have a *? iterator around your second group, so only one line is
captured. I’d try making the sub-regex that matches a single line into
a non-capturing group, then putting a capturing group around the
iterator:

    \\(\\(?:.*\n\\)*?\\)

(BTW what is your plan after you get the whole inside of the <summary>
in a single group? Your description makes it look like you’re going to
want to do something to display /// as //.)

Re: Regex matching summary blocks in C#

[Samvid Mistry]

Yuri Khan <yuri.v.khan@gmail.com> writes:

>
> You have a *? iterator around your second group, so only one line is
> captured. I’d try making the sub-regex that matches a single line into
> a non-capturing group, then putting a capturing group around the
> iterator:
>
>     \\(\\(?:.*\n\\)*?\\)

IIUC, you are suggesting I do this

"\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(\\(:?.*\n\\)*?\\)\\s-*///\\s-*</summary>"

That doesn't seem to change anything.

>
> (BTW what is your plan after you get the whole inside of the <summary>
> in a single group? Your description makes it look like you’re going to
> want to do something to display /// as //.)

I'm removing the <(/)summary> tags from view. There are other ceremonies
like documentation for parameters and return values, which I'm
also planning to simplify. It is partly for aesthetics and partly for me
to try my hand at overlays.

-- 
Thanks,
Samvid

Re: Regex matching summary blocks in C#

[Yuri Khan]

On Sun, 3 Dec 2023 at 21:23, Samvid Mistry <mistrysamvid@gmail.com> wrote:

> >     \\(\\(?:.*\n\\)*?\\)
>
> IIUC, you are suggesting I do this
>
> "\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(\\(:?.*\n\\)*?\\)\\s-*///\\s-*</summary>"

\\(?:, not \\(:? . The former marks the group as non-capturing. The
latter introduces a capturing group starting with a colon.

Re: Regex matching summary blocks in C#

[Samvid Mistry]

Yuri Khan <yuri.v.khan@gmail.com> writes:

> On Sun, 3 Dec 2023 at 21:23, Samvid Mistry <mistrysamvid@gmail.com> wrote:
>
>> >     \\(\\(?:.*\n\\)*?\\)
>>
>> IIUC, you are suggesting I do this
>>
>> "\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(\\(:?.*\n\\)*?\\)\\s-*///\\s-*</summary>"
>
> \\(?:, not \\(:? . The former marks the group as non-capturing. The
> latter introduces a capturing group starting with a colon.

That was my bad. Your expression seems to have worked. Thanks!

Would you mind expaining a bit about why it works? Per my current
understanding, both mine and yours should have the same net effect,
except that marking the inner group non-capturing should make the regex
evaluation a bit faster because emacs need not remember the matched
text. But clearly, my understanding isn't correct.

-- 
Thanks,
Samvid

Re: Regex matching summary blocks in C#

[tomas]

[-- Attachment #1: Type: text/plain, Size: 1095 bytes --]

On Sun, Dec 03, 2023 at 08:26:51PM +0530, Samvid Mistry wrote:
> 
> Yuri Khan <yuri.v.khan@gmail.com> writes:
> 
> > On Sun, 3 Dec 2023 at 21:23, Samvid Mistry <mistrysamvid@gmail.com> wrote:
> >
> >> >     \\(\\(?:.*\n\\)*?\\)
> >>
> >> IIUC, you are suggesting I do this
> >>
> >> "\\(\\s-*\\)///\\s-*<summary>\\s-*\n\\(\\(:?.*\n\\)*?\\)\\s-*///\\s-*</summary>"
> >
> > \\(?:, not \\(:? . The former marks the group as non-capturing. The
> > latter introduces a capturing group starting with a colon.
> 
> That was my bad. Your expression seems to have worked. Thanks!
> 
> Would you mind expaining a bit about why it works? Per my current
> understanding, both mine and yours should have the same net effect,
> except that marking the inner group non-capturing should make the regex
> evaluation a bit faster because emacs need not remember the matched
> text. But clearly, my understanding isn't correct.

Not Yuri here, but I think your expression does match several times,
and it keeps the last match. Yuri's captures the whole bunch just
once.

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]