Alexey Zakhlestin's blog - Latest Comments

Re: DNS SRV-records support in HTTP-browsers

Sergey Sergeev — Mon, 04 Jan 2021 06:25:43 -0000

Let's do it 22, maybe after my death they will implement this.

Re: DNS SRV-records support in HTTP-browsers

RelatedTitle — Mon, 29 Jul 2019 23:10:38 -0000

They didn't.

Re: DNS SRV-records support in HTTP-browsers

Noman — Fri, 04 Jan 2019 09:53:36 -0000

Hi i want a sexy girlfriend<spoiler>

Re: DNS SRV-records support in HTTP-browsers

Noman — Fri, 04 Jan 2019 09:52:22 -0000

Hii
Helo

Re: DNS SRV-records support in HTTP-browsers

Filippos Vasilakis — Wed, 25 Oct 2017 16:49:39 -0000

still hoping for it!

Re: regexps in PHP, again

Clinton Shane Wright — Wed, 13 Sep 2017 16:49:53 -0000

Hi there,

I would like to add an extra function to your test. I tested the same source code but added the mb_ereg_match function, and that ran every so slightly faster than mb_ereg_search.

preg_match: 1.4547030925751
mb_ereg: 1.8813030719757
mb_ereg_search: 0.79150390625
mb_ereg_match: 0.78851294517517

Re: DNS SRV-records support in HTTP-browsers

Mike Sadler — Wed, 19 Apr 2017 02:42:31 -0000

Yes... you can use an IP address, but in the context of SRV records (where you're trying to even further abstract the actual IP by use of FQDN's pointing to address:port, rather than just address)... it seems like an odd requirement. If you e.g. 'always' know the IP, you don't need SRV records, except to make the URI for the resource look 'pretty' (i.e. to abstract the Port, not to load balance). I'd ALWAYS have a target pointing to an 'A' record... sometimes a CNAME... but A records are safer if you do - or might - ever use certificates for the resource concerned (CN must match FQDN). Of course we use very low TTL's for records for load balanced systems. The trailing dot is a BIND thing... your DNS and, more particularly, its UI may or may not require it... most are automatic these days.

Re: DNS SRV-records support in HTTP-browsers

Mike Sadler — Wed, 19 Apr 2017 02:32:44 -0000

Ummmm.... wrong. SIP is w-a-y more time sensitive than... well, pretty much everything and SRV is almost mandatory for it. So, browser publishers don't support it because they are slack, or it needs refinement or whatever... not because SRV isn't the 'right' protocol to offer multi-address, multi-port load sharing without the expense of e.g L7 switches. An 'RFC' is exactly that, so you comment or you support. Doing nothing sucks. Sorry.

Re: DNS SRV-records support in HTTP-browsers

Elo — Tue, 14 Feb 2017 02:30:12 -0000

The reason this is not supported is that it can't serve the purpose you purport.

It works for SMTP because delays of a few minutes are not an issue. But for a timing sensitive application such as browsing, what good would this do?

In your example, the browser might first select www1.example.com. If that didn't respond in 21 seconds, the browser would then go for www2.example.com. If that didn't respond at the 42 second mark, the browser would go for www-backup.example.com.

Who exactly is going to wait 42 seconds for a page to load?

The load balancing you get with multiple A records is the best you can do for applications that require responsiveness. You can't provide responsive fault tolerance using just DNS.

And then, what exactly should the browser consider a failure that requires a fallback? What if the TCP connection can be established, but there's no HTTP response? What if there's an HTTP response, but it's an error?

This is not a bug in Mozilla. This is the absence of a feature; in this case, a braindead misfeature that ought not to be proposed.

Re: DNS SRV-records support in HTTP-browsers

Shmulik Klein — Wed, 07 Dec 2016 07:19:16 -0000

Can I use IP address in the TARGET part of a SRV record ? Who adds the
trailing dot to make the hostename absolute ? is it the DNS server or
the client ?

http://stackoverflow.com/qu...

Re: DNS SRV-records support in HTTP-browsers

Shmulik Klein — Wed, 07 Dec 2016 04:19:15 -0000

Can I use IP address in the TARGET part of a SRV record ? Who adds the trailing dot to make the hostename absolute ? is it the DNS server or the client ?

http://stackoverflow.com/qu...

Re: DNS SRV-records support in HTTP-browsers

network101 — Thu, 03 Nov 2016 03:14:32 -0000

Make that 15 years.

Re: DNS SRV-records support in HTTP-browsers

Yura — Thu, 28 Jul 2016 09:17:04 -0000

Hope they will fix it before 20 years anniversary.

Re: BeOS + SmallTalk

Lala Khaylilla — Wed, 14 Jan 2015 03:23:47 -0000

It?s Actually A Cool And Useful Piece Of Info. I Am Satisfied That You Shared This Useful Information With Us. Please Keep Us Up To Date Like This. Thanks For Sharing. taruhan bola kita

Re: Chronograph 1.4.0

Dinamita Lee — Tue, 07 Oct 2014 15:06:07 -0000

Thank you, powerfull and simple aplication.

Re: DNS SRV-records support in HTTP-browsers

jing — Wed, 06 Aug 2014 03:54:36 -0000

it is 2014 now, seem like FF still not not srv

Re: XSLCache in PECL

Fri, 06 Sep 2013 14:17:40 -0000

i agree. i actually thought i saw documentation (perhaps unofficial), that specifically noted XSLCache was configured for Apache (but did not suggest it wasn't configured for Nginx). it didn't make much sense because, as you mention, XSLCache operates at the PHP level. figured i would ask here before installing.

since i can't find good documentation about the extension (the nytimes site is jacked), i would like to know if this extension allows parameters to be passed to the stylesheets as the XSLTProcessor. i suppose i should just download and read the code.

thank you for maintaining this extension by the way.

Re: XSLCache in PECL

Alexey Zakhlestin — Thu, 05 Sep 2013 13:05:29 -0000

You mean http://wiki.nginx.org/HttpX... ? That's an option when you need to do full-page transoformations during output phase. But XSLT is used for a lot of other tasks (cross-schema transformations of input-data, mostly) and it should happen fast. So, solution on PHP level still makes sense

Re: XSLCache in PECL

Thu, 05 Sep 2013 12:34:34 -0000

Nginx?

Re: Pake: PHP project build system

dgdg — Tue, 20 Aug 2013 00:38:59 -0000

dgggddgdddgd

Re: on PHP’s webserver

raymond — Mon, 05 Aug 2013 04:44:37 -0000

interesting! some more details?

Re: regexps in PHP, again

Dutch Adachi — Sat, 24 Nov 2012 08:59:05 -0000

Thanks, I was almost going to replace series of mb_ereg use in my code to preg_match because of many advices in stack overload etc., There is another post on http://d.hatena.ne.jp/ichii... which shows comparable results

Re: Syck for PHP goes Beta (in PECL)

Surajit Debnath — Fri, 19 Oct 2012 01:57:30 -0000

how to install in windows 7 ? any one can describe...?

Re: DNS SRV-records support in HTTP-browsers

brain Steven — Thu, 04 Aug 2011 03:10:52 -0000

But it is not committed and waits for approval. I doubt, that is the longest-living bug in mozilla, but, still, 10 years is a great age for a bug. Let’s party!
_______________
Steven
DNS Look up

Re: on PHP’s webserver

Pavel Burminsky — Fri, 24 Jun 2011 04:24:38 -0000

Thats great. My working process on web forms will be now easiest.