[migration][web] Scripted Conversions.

[Dave Fisher <da...@comcast.net>]

Yesterday I worked on conversion scripts for the website. They are now in ooo/site/trunk/tools/.

readme.txt
web-list.txt
kenai2website.sh
urlrewrite.sh
urlrewrite.sed
fetch-all-web.sh

The skeleton is now in good shape to handle Kenai based html. A change may be needed to preserve some PDL license comments. Also the footer may need to be changed for PDL for some pages. I have focused on two processes to be used on a project by project basis.

(1) kenai2website.sh ${project}

This is used to move a project from Kenai's svn to the correct place in ooo/site/trunk/content/openofficeorg/${project}

A dos2unix is applied to fix some mixed line endings. If there is an AL 2.0 compatible version of  dos2unix then I'd like to include it.

If the project exists in AOOo then a diff is done to find files that differ. This can be because they were changed in Kenai. More likely this will be customization that was required. The why project will require it.

(2) urlrewite.sh ${project}

This applies a sed script to rewrite urls in the html for a project.
s/http:\/\/www.openoffice.org/\/openofficeorg\/www/g
s/http:\/\/download.openoffice.org/\/openofficeorg\/download/g
s/http:\/\/support.openoffice.org/\/openofficeorg\/support/g
s/http:\/\/contributing.openoffice.org/\/openofficeorg\/contributing/g
s/http:\/\/projects.openoffice.org/\/openofficeorg\/projects/g
s/http:\/\/why.openoffice.org/\/openofficeorg\/why/g

Other tools can be applied to find and properly take care of various licenses and copyrights.

The project should consider using Apache RAT at some point when we are ready to check for AL 2.0 headers as applied.

(3) fetch-all-web.sh - I plan to convert to perform.sh.

It will run another script like kenai2website.sh or urlrewrite.sh for each project in a project list file like web-list.txt.

Once a project is copied then I recommend local builds until the new project is correct before checking in the project. Raphael did this live and some intervention was required to clear the buildbot for the project.

One key part is finding bad url references to images and adding images to content/openofficeorg/ from www.openoffice.org/branding/images/.

Regards,
Dave

Re: [migration][web] Scripted Conversions.

["Marcus (OOo)" <ma...@wtnet.de>]

Thanks for your great work.

I've a recommendation when adding+committing the "development" project.

As there are billions of release notes I would only commit the latest 
into SVN and pack the others into a ZIP file; split into Beta, RC, 
nightly build, etc.

Release notes from OOo 2.0 (from 2006?) are really no longer 
interesting. However, within the ZIP files we still have them and could 
link to.

Marcus



Am 08/29/2011 01:30 AM, schrieb Dave Fisher:
> Yesterday I worked on conversion scripts for the website. They are now in ooo/site/trunk/tools/.
>
> readme.txt
> web-list.txt
> kenai2website.sh
> urlrewrite.sh
> urlrewrite.sed
> fetch-all-web.sh
>
> The skeleton is now in good shape to handle Kenai based html. A change may be needed to preserve some PDL license comments. Also the footer may need to be changed for PDL for some pages. I have focused on two processes to be used on a project by project basis.
>
> (1) kenai2website.sh ${project}
>
> This is used to move a project from Kenai's svn to the correct place in ooo/site/trunk/content/openofficeorg/${project}
>
> A dos2unix is applied to fix some mixed line endings. If there is an AL 2.0 compatible version of  dos2unix then I'd like to include it.
>
> If the project exists in AOOo then a diff is done to find files that differ. This can be because they were changed in Kenai. More likely this will be customization that was required. The why project will require it.
>
> (2) urlrewite.sh ${project}
>
> This applies a sed script to rewrite urls in the html for a project.
> s/http:\/\/www.openoffice.org/\/openofficeorg\/www/g
> s/http:\/\/download.openoffice.org/\/openofficeorg\/download/g
> s/http:\/\/support.openoffice.org/\/openofficeorg\/support/g
> s/http:\/\/contributing.openoffice.org/\/openofficeorg\/contributing/g
> s/http:\/\/projects.openoffice.org/\/openofficeorg\/projects/g
> s/http:\/\/why.openoffice.org/\/openofficeorg\/why/g
>
> Other tools can be applied to find and properly take care of various licenses and copyrights.
>
> The project should consider using Apache RAT at some point when we are ready to check for AL 2.0 headers as applied.
>
> (3) fetch-all-web.sh - I plan to convert to perform.sh.
>
> It will run another script like kenai2website.sh or urlrewrite.sh for each project in a project list file like web-list.txt.
>
> Once a project is copied then I recommend local builds until the new project is correct before checking in the project. Raphael did this live and some intervention was required to clear the buildbot for the project.
>
> One key part is finding bad url references to images and adding images to content/openofficeorg/ from www.openoffice.org/branding/images/.
>
> Regards,
> Dave