Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:126553
Content-Type: multipart/alternative;
 boundary="----sinikael-?=_1-17410167882810.8660317794734658"
Message-ID: <1741015860846.535394408.882677499@gmail.com>
To: internals@lists.php.net
Subject: Re: [PHP-DEV] PHP True Async RFC
Date: Mon, 03 Mar 2025 15:46:28 +0000
In-Reply-To: <1741011949553.2796973485.548949007@gmail.com>
References: <E0EFE00D-07E7-47A1-89C2-865E90812E7B@daniil.it>
 <1741011949553.2796973485.548949007@gmail.com>
User-Agent: Vivaldi Mail/7.1.3570.50
Precedence: bulk
MIME-Version: 1.0
From: edmond.ht@gmail.com (Edmond Dantes)

------sinikael-?=_1-17410167882810.8660317794734658
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit

Lock/Unlock issue


It seems that this is actually about a database query that puts the Fiber 
into a waiting state specifically, query("UNLOCK").

In that case, everything should work correctly. 

Although there are some dangerous edge cases. The database might be under 
high load, causing the query("UNLOCK") request to wait for too long, 
leading to a timeout. This would trigger another exception, which could 
then be interpreted as a complete failure.

Putting a Fiber into a waiting state inside a finally block does not 
contradict the shutdown mode. However, the programmer must be careful 
inside finally section because if a second exception occurs, it means the 
code cannot properly complete execution.

-- 
Ed.
------sinikael-?=_1-17410167882810.8660317794734658
Content-Type: text/html; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit

<html><head></head><body><span class="viv-signature"></span><blockquote 
style="margin: 0 0 0.80ex; border-left: #0000FF 2px solid; padding-left: 
1ex"><blockquote style="padding-left:1ex;border-left-color:rgb(0, 0, 
255);border-left-style:solid;border-left-width:2px;margin-left:0px;margin-bottom:0.8ex;margin-right:0px;margin-top:0px;"><div><span 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);"></span></div></blockquote></blockquote><br><blockquote 
style="margin: 0 0 0.80ex; border-left: #0000FF 2px solid; padding-left: 
1ex"><blockquote style="padding-left:1ex;border-left-color:rgb(0, 0, 
255);border-left-style:solid;border-left-width:2px;margin-left:0px;margin-bottom:0.8ex;margin-right:0px;margin-top:0px;"><div><span 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);">Lock/Unlock 
issue</span></div></blockquote></blockquote><div><br></div><span 
class="viv-signature-below"><div><span style="background-color: 
var(--colorBgIntense); color: var(--colorFgIntense);">It seems that this is 
actually about a database query that puts the </span><strong 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="77" data-start="68">Fiber</strong><span 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);"> into a waiting state specifically, </span><code 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="130" 
data-start="113">query("UNLOCK")</code><span style="background-color: 
var(--colorBgIntense); color: 
var(--colorFgIntense);">.</span></div><div><span 
class="viv-signature-below"><p data-end="234" data-start="135">In that 
case, everything should work correctly.&nbsp;</p><p data-end="234" 
data-start="135">Although there are some <strong data-end="231" 
data-start="207">dangerous edge cases</strong>.&nbsp;<span 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);">The database might be under </span><strong 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="277" data-start="264">high 
load</strong><span style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);">, causing the </span><code style="background-color: 
var(--colorBgIntense); color: var(--colorFgIntense);" data-end="308" 
data-start="291">query("UNLOCK")</code><span style="background-color: 
var(--colorBgIntense); color: var(--colorFgIntense);"> request to 
</span><strong style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="341" data-start="320">wait for too 
long</strong><span style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);">, leading to a </span><strong 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="367" 
data-start="356">timeout</strong><span style="background-color: 
var(--colorBgIntense); color: var(--colorFgIntense);">. This would trigger 
another exception, which could then be interpreted as a </span><strong 
style="background-color: var(--colorBgIntense); color: 
var(--colorFgIntense);" data-end="464" data-start="444">complete 
failure</strong><span style="background-color: var(--colorBgIntense); 
color: var(--colorFgIntense);">.</span></p><p data-is-only-node="" 
data-is-last-node="" data-end="748" data-start="469">Putting a <strong 
data-end="488" data-start="479">Fiber</strong> into a waiting state inside 
a <code data-end="528" data-start="519">finally</code> block <strong 
data-end="558" data-start="535">does not contradict</strong> the shutdown 
mode. However, the programmer <strong data-end="621" data-start="602">must 
be careful</strong> inside finally section because if a <strong 
data-end="675" data-start="655">second exception</strong> occurs, it means 
the code <strong data-end="747" data-start="702">cannot properly complete 
execution</strong>.</p></span></div><div><span style="background-color: 
var(--colorBgIntense); color: 
var(--colorFgIntense);">--&nbsp;</span></div>Ed.</span></body></html>
------sinikael-?=_1-17410167882810.8660317794734658--