<HTML><BODY><div class="js-helper js-readmsg-msg">
<style type="text/css"></style>
<div>
<base target="_self" href="https://e.mail.ru/">
<div id="style_15202693030000000579_BODY"><div class="class_1520275397">
<div class="js-helper_mailru_css_attribute_postfix js-readmsg-msg_mailru_css_attribute_postfix">
<style></style>
<div>
<div id="style_15202680940000000206_BODY_mailru_css_attribute_postfix"><br>
>Can we change the default log_nonblock to nil, which means the<br>
>default is derived from log type, true for pipe/syslog, false<br>
>for file/stdout?<br>
><br>
<br>Yes, we can. So this conversion from nil to default depending on type<br>must be executed in lua or in box_check_config()?<br>
<br>>This would spare us from the useless warning. The patch will go to<br>
>2.x after that.<br>
><br>
<br>Do you mean the patch with raising errors?<br>
<br>
>What was the issue with moving the check for valid combinations<br>
>to box_check_config()? <br>
><br>
<br>
It may be not so easy as in lua to add new cases. It requires add new `if`s for each case.<br>
It may look ugly.<br>
<br>
>How does the check work when log_nonblock or some other<br>
>combination becomes invalid after initial configuration, e.g. in<br>
>this scenario:<br>
><span style="font-family: Arial, Tahoma, Verdana, sans-serif;" data-mce-style="font-family: Arial, Tahoma, Verdana, sans-serif;">box.cfg{}</span><br style="font-family: Arial, Tahoma, Verdana, sans-serif;" data-mce-style="font-family: Arial, Tahoma, Verdana, sans-serif;"><span style="font-family: Arial, Tahoma, Verdana, sans-serif;" data-mce-style="font-family: Arial, Tahoma, Verdana, sans-serif;">>box.cfg{log_nonblock=true}?</span><br style="font-family: Arial, Tahoma, Verdana, sans-serif;" data-mce-style="font-family: Arial, Tahoma, Verdana, sans-serif;"><br>It would not fail as parameter `log` is not specified.<br><br>
>After giving this some thought, the only reason we're doing it here <br>
>is that we failed to refactor say.c and separate log string parser api <br>
>so that it can be called independently from box_check_config. Am<br>
>I right?<br>
><br><br>
For checking loggers it's ok to implement it in box_check_config.<br>But if there are more constraints, it may be inconvenient.<br>
><br>
><cut> a lot of changes which will become unnecessary if we change<br>
>log_nonblock default after moving this problem to 2.0.<br>
><br>>-- <br>
>Konstantin Osipov, Moscow, Russia, <span class="js-phone-number_mailru_css_attribute_postfix"><span class="js-phone-number">+7 903 626 22 32</span></span><br>
><a href="http://tarantool.org" target="_blank" rel=" noopener noreferrer">http://tarantool.org</a> - <a href="http://www.twitter.com/kostja_osipov" target="_blank" rel=" noopener noreferrer">www.twitter.com/kostja_osipov</a><br>
<br>
<br>
-- <br>
Thank you,<br>
Ilya Markov<br>
</div>
</div>
</div>
</div></div>
<base target="_self" href="https://e.mail.ru/">
</div>
</div></BODY></HTML>