As you know, I’ve been working on branches this long holiday weekend. Today, I managed to get the display of a port on a branch to correctly. Along the way, I discovered a downside to storing raw data in the database. Take for example this output: This is the output of make -V RUN_DEPENDS. You …
A downside of storing the exact results of ‘make -V’ Read More »
In my previous post, I wrote about the need to copy a port from head to a branch in some circumstances, specifically when a master-port is updated and the slave ports do not already exist in the FreshPorts database on that branch. This is a side-effect of FreshPorts only taking action for commits, and not …
I used to think branches were simple when it came to FreshPorts, but I was wrong. Consider slave ports for example. If we get a commit for databases/mariadb100-server, we have to update the slave ports, but only the slave ports on the branch, not the slave ports on head. That sounds easy enough, but keep …
Today, kwm mentioned that http://www.freshports.org/Mk/bsd.port.mk was out of date, showing the most recent commit as back in March. My first concern was the system had stopped recording such commits. To check on that, I looked at https://svnweb.freebsd.org/ports/head/Mk/bsd.port.mk and searched my email for the most recent commit message. Extracting the message id from that email, I …
In the past few days, I have implemented the full pathnames solution on dev. It was first implemented on gelt, a server setup explicitly for developing full pathnames. I am so glad that I keep notes when making changes such as this. Sometimes it is days or weeks between testing out a new approach and …
I think the work to move to full pathnames is proceeding well. I created another website, gelt.freshports.org, and I’m doing all the work over there. What I learn and the code changes will eventually be applied to the other websites. The differences should be subtle, and barely noticeable. Most of the changes are internal. Compare …
Before the last post was published, I forgot to add in the fix for the pathname problem. freshports.org=# select ElementPathnameUpdateChildren(453745); It run. A long time. I let it run. For hours. About 16 hours. Then I killed it. I checked the procedure. It made sense. Then I checked the update statement: freshports.org=# explain analyse UPDATE …
I started looking at how to create these new directories as mentioned yesterday. This starts the process: $ psql freshports.org psql (9.0.8, server 9.0.9) Type “help” for help. freshports.org=# select Element_Add(‘/base/head’, ‘D’); element_add ————- 453742 (1 row) freshports.org=# select Element_Add(‘/doc/head’, ‘D’); element_add ————- 453743 (1 row) freshports.org=# select Element_Add(‘/ports/head’, ‘D’); element_add ————- 0 (1 row) …
We learn as we go. We create new designs based on past experiences and knowledge to date. Back when FreshPorts started, we catered only for ports. Now we cover the ports and src trees. Now that we are on svn, we can do much more than when on cvs. One change I would like to …