[vfs] Enchancing extraction

[hrjk <ha...@zeniitti.net>]

Hi.

I was wondering if you've planned to enchance extraction process of VFS? I
know, that extracting packages isn't the main driver behind developing VFS
library. However, VFS works unlike Java.util.zip (the umlaut issues) and
can handle several file types.

We've been using VFS, and with small zip packages (less entries)
extraction is efficient. However, when the package size (and entry count)
increases, the extraction time is increased almost exponentially. For
example, one large (>1Gb) zip took 7 minutes with org.apache.tools.zip and
1:07 minutes with VFS. So over 1 hour more... This is a huge issue for us
atleast.

So basicly I was wondering if it would be possible to change VFS to easily
"skip" the caching process when resolving file, but merely request next
entry from zip or any other file. I think this would help the efficiency
issues, or am I missing something? What do you think of this?

BR,
hrjk


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-user-help@jakarta.apache.org

Re: [vfs] Enchancing extraction

[Mario Ivankovits <ma...@ops.co.at>]

Hi!
> I will surely inform you, if I find any clues during development. I don't
> however have the time to concentrate fully on this (employer wouldn't like
> that :).
>   
Ok, I understand.

>> Anyway, please open a jira ticket for this at
>> http://issues.apache.org/jira
>>     
>
> I don't seem to have access to Jakarta project (where Commons VFS reside).
>   
Try this url, should work, but for sure, you have to register with jira
to create a new ticket.

http://issues.apache.org/jira/browse/VFS

Ciao,
Mario


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-user-help@jakarta.apache.org

Re: [vfs] Enchancing extraction

[hrjk <ha...@zeniitti.net>]

> Hmmm .... without a big debugging session I am not sure how to solve
> this. I'll see if I find some time for this - or maybe you can step in
> here? At least where the time will be lost will be great.

I will surely inform you, if I find any clues during development. I don't
however have the time to concentrate fully on this (employer wouldn't like
that :).

> Anyway, please open a jira ticket for this at
> http://issues.apache.org/jira

I don't seem to have access to Jakarta project (where Commons VFS reside).

Br,
hrjk


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-user-help@jakarta.apache.org

Re: [vfs] Enchancing extraction

[Mario Ivankovits <ma...@ops.co.at>]

Hi!
> So basicly I was wondering if it would be possible to change VFS to easily
> "skip" the caching process when resolving file, but merely request next
> entry from zip or any other file.
It would be great to check where the time is really lost, I dont think
its a caching issue.

The problem might be, that VFS uses a two pass strategy to handle zip
files, one to create the toc and then to extract these files, now
extracting them not in the same order like they are stored might issue a
big performance impact.

Hmmm .... without a big debugging session I am not sure how to solve
this. I'll see if I find some time for this - or maybe you can step in
here? At least where the time will be lost will be great.

Anyway, please open a jira ticket for this at http://issues.apache.org/jira

Thanks!

Ciao,
Mario


---------------------------------------------------------------------
To unsubscribe, e-mail: commons-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-user-help@jakarta.apache.org