[
https://issues.apache.org/jira/browse/CASSANDRA-373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12745193#action_12745193
]
Bill de hOra edited comment on CASSANDRA-373 at 8/19/09 1:54 PM:
-----------------------------------------------------------------
{quote}
1. Wherever possible, lines should wrap at 75 chars.
{quote}
is this not best done by reducing the comment noise?
Also verbose comments tend to be annoying in practice operationally - hey get
in the way of the the actual elements you need to set/read when you need to
see/read them on a console - tomcat is an example offender and a system i'm
tired of stripping comments from.
{quote}
4. Path locations that conform to the FHS.
{quote}
Not for us to mandate surely.
* The tilde's at the beginning of the comment lines impair reading.
* I don't like putting advice or good rules of thumb in the config - would we
do this kind of thing in the thrift?
* CommitLogSyncDelay : "in millis" but other elements have the InFoo quantity
pattern. Inconsistent.
* MemtableObjectCountInMillions: asking people to type in floats is error
prone, why not MemtableObjectCount?
* CommitLogSync: doesn't document its legal values
* Thrift*: should be wrapped in an holding element. Because, and I'm guessing
here, Thrift will not remain the only, or even the default, transport/api.
* InitialToken: doesn't document what happens when empty
* Seed, ListenAddress, ThriftAddress: don't document whether they take Ipv6
addresses
was (Author: dehora):
{quote}
1. Wherever possible, lines should wrap at 75 chars.
{quote}
is this not best done by reducing the comment noise?
Also verbose comments tend to be annoying in practice operationally - hey get
in the way of the the actual elements you need to set/read when you need to
see/read them on a console - tomcat is example offender and a system i'm tired
of stripping comments from.
{quote}
4. Path locations that conform to the FHS.
{quote}
Not for us to mandate surely.
* The tilde's at the beginning of the comment lines impair reading.
* I don't like putting advice or good rules of thumb in the config - would we
do this kind of thing in the thrift?
* CommitLogSyncDelay : "in millis" but other elements have the InFoo quantity
pattern. Inconsistent.
* MemtableObjectCountInMillions: asking people to type in floats is error
prone, why not MemtableObjectCount?
* CommitLogSync: doesn't document its legal values
* Thrift*: should be wrapped in an holding element. Because, and I'm guessing
here, Thrift will not remain the only, or even the default, transport/api.
* InitialToken: doesn't document what happens when empty
* Seed, ListenAddress, ThriftAddress: don't document whether they take Ipv6
addresses
> storage-conf.xml reformatting
> -----------------------------
>
> Key: CASSANDRA-373
> URL: https://issues.apache.org/jira/browse/CASSANDRA-373
> Project: Cassandra
> Issue Type: Bug
> Reporter: Eric Evans
> Assignee: Eric Evans
> Attachments:
> v1-0001-CASSANDRA-373-re-format-factor-conf-storage-conf.xml.txt,
> v2-0001-CASSANDRA-373-re-format-factor-conf-storage-conf.xml.txt
>
>
> Our sample config (conf/storage-conf.xml) is the canonical source of
> configuration documentation. As such readability should be a priority, and it
> should serve as the best possible basis for customization.
> I propose the following:
> 1. Wherever possible, lines should wrap at 75 chars. The file will be edited
> post-installation by operational personnel, who are often confined to
> standard 80 character terminals.
> 2. Indention of 2 spaces to make maximum use of horizontal real estate.
> 3. More distinctive multi-line comments.
> 4. Path locations that conform to the FHS.
> Patch to follow...
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
