[
http://jira.codehaus.org/browse/MERCURY-119?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Oleg Gusakov updated MERCURY-119:
---------------------------------
Description:
This is a todo list to keep track/share optimization ideas
* VirtualRepositoryReader - refactor readArtifact() into 3 pieces, eliminate
readArtifactsNoBatch(). Pieces
** read qualified artifacts (we know the repository they are in)
** read virtual artifacts - SNAPSHOT/LATEST/RELEASE
** read the rest: not qualified real versions
* implement negative cache [MERCURY-116]. Work started, need to add creation of
negative record in case of bad metadata bypass
* ? cache version lists for resolved artifacts. This will help to prevent
re-parsing of POMs in MavenDependencyProcessor
* ? make storing of pom bytes in the artifact optional / compress them
* add evict() call to the metadata cache
* implement LRU strategy in the metadata in-memory cache
* pool HttpClient instances in the VirtualRepositoryReader and
RemoteRepoWriterM2 instead of closing them
* review remote repository strategy regarding reading available versions. It
now tries to replace virtuals with actual version thus trying to read metadata
for all virtuals. Time consuming on clean local repository and with lack of
negative metadata caching
was:
This is a todo list to keep track of optimization ideas
* VirtualRepositoryReader - refactor readArtifact() into 3 pieces, eliminate
readArtifactsNoBatch(). Pieces
** read qualified artifacts (we know the repository they are in)
** read virtual artifacts - SNAPSHOT/LATEST/RELEASE
** read the rest: not qualified real versions
* implement negative cache [MERCURY-116]. Work started, need to add creation of
negative record in case of bad metadata bypass
* ? cache version lists for resolved artifacts. This will help to prevent
re-parsing of POMs in MavenDependencyProcessor
* ? make storing of pom bytes in the artifact optional / compress them
* add evict() call to the metadata cache
* implement LRU strategy in the metadata in-memory cache
* pool HttpClient instances in the VirtualRepositoryReader and
RemoteRepoWriterM2 instead of closing them
* review remote repository strategy regarding reading available versions. It
now tries to replace virtuals with actual version thus trying to read metadata
for all virtuals. Time consuming on clean local repository and with lack of
negative metadata caching
> Mercury optimization
> ---------------------
>
> Key: MERCURY-119
> URL: http://jira.codehaus.org/browse/MERCURY-119
> Project: Mercury
> Issue Type: Improvement
> Components: Misc/All
> Affects Versions: 1.0-alpha-8
> Reporter: Oleg Gusakov
> Assignee: Oleg Gusakov
>
> This is a todo list to keep track/share optimization ideas
> * VirtualRepositoryReader - refactor readArtifact() into 3 pieces, eliminate
> readArtifactsNoBatch(). Pieces
> ** read qualified artifacts (we know the repository they are in)
> ** read virtual artifacts - SNAPSHOT/LATEST/RELEASE
> ** read the rest: not qualified real versions
> * implement negative cache [MERCURY-116]. Work started, need to add creation
> of negative record in case of bad metadata bypass
> * ? cache version lists for resolved artifacts. This will help to prevent
> re-parsing of POMs in MavenDependencyProcessor
> * ? make storing of pom bytes in the artifact optional / compress them
> * add evict() call to the metadata cache
> * implement LRU strategy in the metadata in-memory cache
> * pool HttpClient instances in the VirtualRepositoryReader and
> RemoteRepoWriterM2 instead of closing them
> * review remote repository strategy regarding reading available versions. It
> now tries to replace virtuals with actual version thus trying to read
> metadata for all virtuals. Time consuming on clean local repository and with
> lack of negative metadata caching
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
