Forums : Technical Issue Help

Hello,

I recently checked the Preesm project and realized that, although the code is said to have been analyzed 1 day ago, the commits done in the last week are not included.
Is it normal ?

kdesnos over 11 years ago

Karol,

There are two dates of interest... Analyzed 2 days ago based on code collected 2 days ago. The analysis can occur on any date for a variety of reasons. The code collected date is more important and is the date of the least recently updated enlistment. This should coincide with the latest commit for an active project.

Currently looks OK to me. Please remind me here if the project goes astray again.

Thanks!

ssnow-blackduck over 11 years ago

Karol,

Strike that... It does look as if there is something odd going on. Standby.

Thanks!

ssnow-blackduck over 11 years ago

Karol,

Looks like you were the victim of the Allura upgrade phenomenon. The current URL for the subversion has changed and all new commits are being made only to that subversion. The original repository is static and as you saw it on ohloh (r1415 - 2013-05-12).

Added the new URL and removed the old.

Standby while it performs the initial fetch.

Thanks!

ssnow-blackduck over 11 years ago

Karol,

We are left with a couple of options, none of which is perfect. I have left the project with both enlistments in-place. This preserves the history of the project but sacrifices accurate lines-of-code measurements for the current code and all the calculations which depend on them. There are double entries for all the revisions r1113 through r1415. Should you prefer, you can again turn off the older repository and more accurately represent the lines-of-code measurements for the current code but remove reference to all commits before r1113. Your choice.

Let me know what you decide.

Thanks!

ssnow-blackduck over 11 years ago

hi,

I thinks i will keep both repository for now.
There's something I do not understand about this problem: Why is the complete history not contained in the new repository ?

Thanks for your help,

Karol

kdesnos over 11 years ago

Karol,

That is an interesting question. It (r1113 - Create new trunk) seems to have been a convenient place to start fresh, I suppose. In any case you can discuss that with SourceForge management if it's important.

The decision is mainly based on what is more important - accurate history or accurate calculation of the metrics.

Let me know if you find out anything about how the decision was made. We can pass it on to other projects in the same boat or facing the Allura upgrade.

Thanks!

ssnow-blackduck over 11 years ago

I think i found the reason for this bug !

As you said, r1113 corresponds to the creation of a new trunk.
This commit seems to be a missuse of the svn repository consisting in deleting the trunk directory, then renaming the branch directory into trunk.
Consequently, the history of the new trunk directory starts at this revision !

To confirm this, could this be possible to allow the use of
svn://svn.code.sf.net/p/preesm/code instead of svn://svn.code.sf.net/p/preesm/code/trunk as the repository address? I think that using the code folder instead of the trunk, We might get a more accurate and complete history. (This should not increase too much the repo size since we have never used branches intensively in this project)

Thanks !.

kdesnos over 11 years ago

Karol,

It's an interesting idea... I'll explore it today.

Thanks!

ssnow-blackduck over 11 years ago

Karol,

Just checked and there's nothing there in either subversion except trunk. The branches directories are empty in both cases. Also, at least in this case, the system refuses to accept anything besides trunk (perhaps it checks to see the contents of the rest, although I doubt it.)

Sorry!

ssnow-blackduck over 11 years ago