lists.gnu.org/archive web search is broken?

lists.gnu.org/archive web search is broken?

[Ag Ibragimov]

How do you search for messages in emacs-devel archive (web form)? I
think it's broken.
I can never find relevant messages, take for example this one:

https://lists.gnu.org/archive/html/emacs-devel/2022-06/msg01229.html

The content isn't relevant. But let's try to find that page using
"search".

I've attempted to use the following search strings:

1) +subject:"Looking at function"
- No document matching your query.

2) +message-id:<86edz8k6q1.fsf@mail.linkov.net>
- No document matching your query.

3) +from:juri@linkov.net
- Total 865 documents matching your query.;
Sorted in reverse chronological order, the latest indexed document dates
Thu, 18 Mar 2021. The original message is from Tue, 28 Jun 2022.

4) Apparently, the search can take `+date:` argument, but I couldn't
figure out the proper formatting for the date. I couldn't even get it to
show messages posted on a specific date.

Does anyone use the archive? What's the point of having a web-page
for it, if you can't even search for stuff without downloading it
locally? Or maybe "How to search" instructions need updating?

Re: lists.gnu.org/archive web search is broken?

[Visuwesh]

[வியாழன் ஜூலை 21, 2022] Ag Ibragimov wrote: 
 
> Does anyone use the archive? What's the point of having a 
> web-page for it, if you can't even search for stuff without 
> downloading it locally? Or maybe "How to search" instructions 
> need updating? 

It is indeed broken (try searching in bug-gnu-emacs, you will have more
fun).  Right now, a good alternative is https://yhetil.org.  It only has
emacs-devel and bug-gnu-emacs archived though.

Re: lists.gnu.org/archive web search is broken?

[Eli Zaretskii]

> From: Ag Ibragimov <agzam.ibragimov@gmail.com>
> Date: Thu, 21 Jul 2022 19:07:20 -0500
> 
> Does anyone use the archive?

I do.

> What's the point of having a web-page for it, if you can't even
> search for stuff without downloading it locally? Or maybe "How to
> search" instructions need updating?

The search is indeed problematic, but it isn't entirely dysfunctional.

Anyway, the archive is not maintained by us, it's a separate project.
Please send the problems you find to mailman@gnu.org, that's where
people who maintain the GNU email archives are.

Thanks.

Re: lists.gnu.org/archive web search is broken?

[Eli Zaretskii]

> From: Visuwesh <visuweshm@gmail.com>
> Cc: emacs-devel@gnu.org
> Date: Fri, 22 Jul 2022 10:44:46 +0530
> 
> [வியாழன் ஜூலை 21, 2022] Ag Ibragimov wrote: 
>  
> > Does anyone use the archive? What's the point of having a 
> > web-page for it, if you can't even search for stuff without 
> > downloading it locally? Or maybe "How to search" instructions 
> > need updating? 
> 
> It is indeed broken (try searching in bug-gnu-emacs, you will have more
> fun).

To search bug-gnu-emacs, my advice is to use the debbugs search
capability, available from the https://debbugs.gnu.org/ page.

Re: lists.gnu.org/archive web search is broken?

[Juri Linkov]

> How do you search for messages in emacs-devel archive (web form)? I
> think it's broken.

Indeed, the web search is broken - there is a huge delay before the
messages get indexed.

> Does anyone use the archive? What's the point of having a web-page
> for it, if you can't even search for stuff without downloading it
> locally? Or maybe "How to search" instructions need updating?

The nicest way is indeed to download locally from
https://lists.gnu.org/archive/mbox/emacs-devel/
https://lists.gnu.org/archive/mbox/bug-gnu-emacs/

This gives the power of using grep with any regexps, and even
special mailbox searching programs like `grepmail` that can match
on headers, mail body and MIME attachments.

Re: lists.gnu.org/archive web search is broken?

[Visuwesh]

[வெள்ளி ஜூலை 22, 2022] Eli Zaretskii wrote: 

>> > Does anyone use the archive? What's the point of having a 
>> > web-page for it, if you can't even search for stuff without 
>> > downloading it locally? Or maybe "How to search" instructions 
>> > need updating?  
>>  It is indeed broken (try searching in bug-gnu-emacs, you will 
>> have more fun). 
> 
> To search bug-gnu-emacs, my advice is to use the debbugs search 
> capability, available from the https://debbugs.gnu.org/ page. 

Yep, this is what I was doing before I found out about yhetil.org.

Re: lists.gnu.org/archive web search is broken?

[Daniel Martín]

Juri Linkov <juri@linkov.net> writes:

>
> The nicest way is indeed to download locally from
> https://lists.gnu.org/archive/mbox/emacs-devel/
> https://lists.gnu.org/archive/mbox/bug-gnu-emacs/
>
> This gives the power of using grep with any regexps, and even
> special mailbox searching programs like `grepmail` that can match
> on headers, mail body and MIME attachments.

Based on this approach, I have a small Bash function that uses Wget to
download the complete mailing list history into a single mbox file.

function download_mailing_list() {
    url=$1
    list_name=$2
    wget -r -nH --cut-dirs=2 --no-parent -A "*-*" --reject="index.html*" $url
    cd $list_name
    cat * > $list_name.mbox
}

Example:

download_mailing_list https://lists.gnu.org/archive/mbox/emacs-devel/ emacs-devel

will download all the history of emacs-devel and generate a single
emacs-devel.mbox file.  (Caveat emptor, you might want to convert it to
a more efficient format.)

Re: lists.gnu.org/archive web search is broken?

[Ag Ibragimov]

That's all pretty cool, but how do you suggest sharing a link to a
thread with someone who doesn't want to download the entire archive?

We need to fix the search, it's ridiculous to have it broken like
that. I emailed mailman@gnu.org, hope someone replies.

Thank you, guys.

Daniel Martín <mardani29@yahoo.es> writes:

> Juri Linkov <juri@linkov.net> writes:
>
>>
>> The nicest way is indeed to download locally from
>> https://lists.gnu.org/archive/mbox/emacs-devel/
>> https://lists.gnu.org/archive/mbox/bug-gnu-emacs/
>>
>> This gives the power of using grep with any regexps, and even
>> special mailbox searching programs like `grepmail` that can match
>> on headers, mail body and MIME attachments.
>
> Based on this approach, I have a small Bash function that uses Wget to
> download the complete mailing list history into a single mbox file.
>
> function download_mailing_list() {
>     url=$1
>     list_name=$2
>     wget -r -nH --cut-dirs=2 --no-parent -A "*-*" --reject="index.html*" $url
>     cd $list_name
>     cat * > $list_name.mbox
> }
>
> Example:
>
> download_mailing_list https://lists.gnu.org/archive/mbox/emacs-devel/ emacs-devel
>
> will download all the history of emacs-devel and generate a single
> emacs-devel.mbox file.  (Caveat emptor, you might want to convert it to
> a more efficient format.)