about process to decrease size of idempiere repository.

[Hiep Lq]

Hi everyone.

at the moment size of idempiere repository is very big (1.5G)

I test a solution for resize it. result is nice. new size is decreased to 500M

but it's effect to every one current syncing with idempiere repository, 

we will must re-clone and re-apply private patch.

I description process at below wiki, please let me know it's worth to process or not?

http://wiki.idempiere.org/en/Decrease_Idempiere_Repository

[HansAnton Riess]

I think it make sense. As I know there is a restriction in bitbucket of 2GB, so this activity is worth to execute.

[Carlos Antonio Ruiz Gomez]

Great Hiep, thanks for taking care of this.

One question - are the commit IDs preserved after the conversion?

[Hiep Lq]

Hi Carlos.

No, commit id don't preserved.

in idempiere, what's used for commit id? it's importance for maintain it?

[Carlos Antonio Ruiz Gomez]

yes, that's a bad effect  :-(   all historic external references (i.e. forums, JIRA tickets and blogs, etc) will be broken.

[redhuan d. oon]

I search a bit for some tips how to avoid losing the IDs while reducing and got this, which i quote a bit to give an idea if this maybe helpful link to pursue:
http://stackoverflow.com/questions/25399705/reduce-size-of-git-repository-on-bitbucket

On your original repo, you ran a git filter-branch command that copied commits to new, modified commits (with different trees and hence, at some point, different true names which led to different parent IDs in subsequent commits, and so on). However, if you copy those copied commits by applying a filter that this time does nothing, then the new tree objects will be the same as the old ones. If the new parent is the same, and the author, committer, and message also all remain the same, the new commit-ID for the copy will be the same as the old ID.

end quote.

[Hiep Lq]

Hi Carlos.

for jira link, not sure, but in document said it's use issue key for link, so we can re-run a sync process, or make a plug-in for this job.

https://bitbucket.org/idempiere/idempiere/admin/links

for other place (bloc, forum,..) just keep old repository for reference.

@Red, thanks for your link, it's not exact apply to mercurial but let me idea.

as i understand about post in stackoverflow.

when convert has 2 type of patch
  1. patch with modification will re-commit => can't maintain ID

  2. patch without modification, will use apply function => maintain ID

we can improve convert extension for implement it, but maybe it make conflict when commit some next other patch.

other idea relate improve convert extension

current when import a patch we can change (userName, date of patch,...)

in case we can manual set ID, just change extension to do it.

[Heng Sin Low]

One of the reason that the repository size growth rapidly is the current practise of storing large binary files ( zip, jar, etc ) in the repository. That needs to be fix after the repository have been shrink ( the zk project in dev branch have a fix for that ), otherwise the same issue will comes up again in future.

--
You received this message because you are subscribed to the Google Groups "iDempiere" group.
To unsubscribe from this group and stop receiving emails from it, send an email to idempiere+...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/idempiere/c9b31a54-bc44-43d8-b2f9-08037585dd75%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Hiep Lq]

sure Hengsin.

use get ant method same zk library or use "Maven Ant Tasks" for achieve full dependency.

please reference:

https://maven.apache.org/ant-tasks/

i make a test project for collect all dependency for jasper report and it work perfect.

https://bitbucket.org/hieplq/org.idempiere.report.jasper.library/overview