module Jumplinks

[Karinne Cyphers]

Thanks for the advice @dragan... however, after trying both of those suggestions the result and the error is the same :(

[Mike Rockett]

@Karinne Cyphers I've pushed up 1.5.52 which attempts to fix the change that PHP 7.3 made to preg_*, where hyphens need to be escaped if not at the end of a character set in an expression. Not able to test on my side, so please could you update and let me know if it runs? Thanks!

[Karinne Cyphers]

Unfortunately, Using 1.5.52 created a number of issues for us.  We did downgrade our servers to PHP 7.2.x and for us, using 1.5.51 generates far fewer errors.  With 52 we were having an issue where the headers weren't being set and our cookies we use for different forms weren't being set.

[Zeka]

Hi @Mike Rockett

I'm moving WP site to PW where I have such URLs with Cyrillic characters 

2018/11/16/обухівських-школярів-навчать-писати/

and actualy it doesn't work. I have turned on debug mode and it says that there is no match.  

Also, I have tried to use Wildcards like

Source 
{year:num}/{month:num}/{date}/{title:segment}
Destination
[template=archive-item,archive_url=/{year}/{month}/{date}/{title}/]]

In the archive-item template, I have field archive_url that contain full URL in the format '/2018/11/16/обухівських-школярів-навчать-писати/'.

Have I missed something?

Thanks. 

[wbmnfktr]

Is it maybe necessary to enable UTF8 page names first?

https://processwire.com/blog/posts/page-name-charset-utf8/

 

[Mike Rockett]

Hi @Zeka

Did you comment out rule 16a and uncomment 16b in your htaccess? (reference)

[Zeka]

@wbmnfktr @Mike Rockett Thanks, but I'm not sure that you understand me correctly.

On PW site I don't want to use UTF8 name, but I have such URLs on the old site and I want to make a redirect from URLs like '2018/11/16/обухівських-школярів-навчать-писати/' to some destination page.

Currently, I tracked down that it's relative to urldecode/urlencode. 

For example, I have such jumplink

Here is the output of Jumplink debug mode

 

As you can see "request" is not urldecoded, but we are trying match decoded string.

If I change this code on line 204 https://gitlab.com/rockettpw/seo/jumplinks-one/blob/master/ProcessJumplinks.module.php#L204

$this->request = ltrim(@$_SERVER['REQUEST_URI'], '/');

to

$this->request = urldecode(ltrim(@$_SERVER['REQUEST_URI'], '/'));

I get such output

-------------

The next issue is that I can't get wildcards to work with Cyrillic characters

 

[dragan]

It's weird. I have managed to get it working with German umlauts, but not with cyrillic characters.

I guess the culprit is not line 204, but further below:

$input = iconv('utf-8', 'us-ascii//TRANSLIT', $input);

I've tried to implement the fix suggested here, but it still didn't work... (1st user-contributed note).

[elabx]

Hi everyone! Does anyone know if it's possible to map a source url taking into account optional string at the end? For example:

Source -> Destination

example.com/es/this-is-a-nice-url -> example.com/es/this-is-a-nice-url

example.com/es/this-is-a-nice-url-1 -> example.com/posts/this-is-a-nice-url

example.com/es/this-is-a-nice-url-2 -> example.com/posts/this-is-a-nice-url

 

Making the "-1" or "-2" optional.

[Zeka]

On 3/23/2019 at 4:10 PM, dragan said:

It's weird. I have managed to get it working with German umlauts, but not with cyrillic characters.

I guess the culprit is not line 204, but further below:

$input = iconv('utf-8', 'us-ascii//TRANSLIT', $input);

I've tried to implement the fix suggested here, but it still didn't work... (1st user-contributed note).

These regex expressions don't not match Cyrillic characters

https://gitlab.com/rockettpw/seo/jumplinks-one/blob/master/ProcessJumplinks.module.php#L94

https://gitlab.com/rockettpw/seo/jumplinks-one/blob/master/ProcessJumplinks.module.php#L95

https://gitlab.com/rockettpw/seo/jumplinks-one/blob/master/ProcessJumplinks.module.php#L98

https://gitlab.com/rockettpw/seo/jumplinks-one/blob/master/ProcessJumplinks.module.php#L99

They should be changed to something like

[\wа-я]+/ig

 

[Zeka]

6 hours ago, elabx said:

Hi everyone! Does anyone know if it's possible to map a source url taking into account optional string at the end? For example:

Source -> Destination

example.com/es/this-is-a-nice-url -> example.com/es/this-is-a-nice-url

example.com/es/this-is-a-nice-url-1 -> example.com/posts/this-is-a-nice-url

example.com/es/this-is-a-nice-url-2 -> example.com/posts/this-is-a-nice-url

 

Making the "-1" or "-2" optional.

 

Not tested

content/{path}/some-url[-{num}]

[elabx]

4 hours ago, Zeka said:

Not tested

content/{path}/some-url[-{num}]

Didn't work unfortunately ?

[Peter Knight]

Feature request, Mike. I know you're busy so just putting it out there.

I am logging 404 hits to a site that get's a ton of spam links. An old SEO spam issue which I am cleaning up.

Anyway, I notice that the jumplinks database table was becoming huge and had bloated to thousands of enties making the database itself very large.

That better house-keeping is my responsibility but wondered if a feature might be useful to optionally auto-delete these over a certain threshold. Or possibly send an email or notification that the logs were becoming quite big etc.

 

[Mike Rockett]

On 4/10/2019 at 6:45 PM, elabx said:

Hi everyone! Does anyone know if it's possible to map a source url taking into account optional string at the end? For example:

Source -> Destination

example.com/es/this-is-a-nice-url -> example.com/es/this-is-a-nice-url

example.com/es/this-is-a-nice-url-1 -> example.com/posts/this-is-a-nice-url

example.com/es/this-is-a-nice-url-2 -> example.com/posts/this-is-a-nice-url

 

Making the "-1" or "-2" optional.

Unfortunately, this isn't built in as a feature, but I'm thinking the non-capture feature might work for you. Not tested, but try using <[-]{1}[\d*]> at the end of the source.

On 4/13/2019 at 2:31 PM, Peter Knight said:

Feature request, Mike. I know you're busy so just putting it out there.

I am logging 404 hits to a site that get's a ton of spam links. An old SEO spam issue which I am cleaning up.

Anyway, I notice that the jumplinks database table was becoming huge and had bloated to thousands of enties making the database itself very large.

That better house-keeping is my responsibility but wondered if a feature might be useful to optionally auto-delete these over a certain threshold. Or possibly send an email or notification that the logs were becoming quite big etc.

 

Thanks Peter – I recall wanting to do this for v2, where you choose how many you'd like to retain in the log.

I think I'm gonna need to find some time to finish v2. Will likely need to make a few changes here and there (not up to speed with all the new core-dev) before I proceed with the frontend work that's still due. 

[Mike Rockett]

Just a note on unicode support: this is vastly improved in v2, once rule 16b is in use. Not planning on bringing it to v1, unfortunately.

[MrSnoozles]

Hey Mike,

is there anything the community could do to support you with v2? The new version is highly anticipated ?

[elabx]

Thanks for the reply @Mike Rockett!

In a completely different scneario, I am trying this as source: 

[propiedades]/detalle_propiedad/1363/some-segment[/][?b=1]

(As destination it's a plain url)

So that it would match:

• detalle_propiedad/1363/casa-en-preventa-colonia-maya?b=1
• detalle_propiedad/1363/casa-en-preventa-colonia-maya/?b=1
• propiedades/detalle_propiedad/1363/casa-en-preventa-colonia-maya/?b=1

Does anyone know if I am correct in this assumption?

EDIT: Solved it with two jumplinks like this:

• {segments}/{num:num}/{segment}/?b=1
• {segments}/{num:num}/{segment}?b=1

I also think there might be something the community could do to help? Funding? "Crowdfunding"? Though maybe Mike's slammed with client work or sth and that's it haha

[Mike Rockett]

On 4/17/2019 at 1:20 AM, elabx said:

I also think there might be something the community could do to help? Funding? "Crowdfunding"? Though maybe Mike's slammed with client work or sth and that's it haha

Under normal circumstances, contributions to OSS help out with continued development – issue on my side is exactly that: slammed with work. With that said, there are several public holidays coming up (four-day weeks for two/three weeks), so I think I can slot in some time to get things going here and at maybe even put up a testing version of sorts next month. Further on down the line, when it's stable, we'd definitely need to look at getting someone onboard to assist with maintenance.

[Mike Rockett]

@elabx Also, v2 will support the optional trailing slash by default (configurable).

[ceberlin]

Hi Mike, the update 1.5.53->1.5.54 with the PW "Upgrades" plugin does not work (worked before). It re-installes the old version.

[wbmnfktr]

Can confirm this behaviour.

[Mike Rockett]

Ah, my apologies - will sort this out. ?

[Mike Rockett]

And done – was a simple case of me forgeting I've switched to tag-based releases.

[teppo]

Hey @Mike Rockett – just wondering if you could add this module to Packagist as well? Would make installing via Composer nice and easy ?

Again setting type as pw-module and adding wireframe-framework/processwire-composer-installer as a dependency would be perfect. Thanks in advance for considering!

[coxharrison]

Hi everyone,

I oversaw a site migration recently. A large part of the legacy site's URL structure changed and we installed Jumplinks to manage all the redirects.

Since the launch I've been monitoring 404s in Google Analytics and we appear to getting quite a few (around 40 per week). What's odd is that the affected URLs all have jumplinks in places, and when I test them the redirects work perfectly fine. 

I can't see why Analytics would be picking these up but I'm quite concerned- if anyone as any insight it would be greatly appreciated.

Best

Harry