Gentoo Linux Installer Meeting Summary: 12 Sep 2004esammer stated the purpose for the meeting: "the reason for this meeting is to discuss current status, releng integration efforts, live cd requirements, and a release date". zhen took the floor and stated that if the Gentoo Linux Installer can be released as an ebuild there will be no problem integrating it into a LiveCD. He also stated that the Releng project is flexible to do whatever the Installer project requires, including creation of X and CLI LiveCDs. zhen opened the floor to questions regarding Releng coordination. samyron asked about execution points on the LiveCDs, where the installer should start in the LiveCD boot process, and what should occur after the installer exits. zhen answered that the execution points will vary based on the specific LiveCDs and that it is up to the Installer project to determine the execution specifics. esammer noted that the LiveCD requirements would be a future issue, and that the only problem he could forsee is being able to run the installer before unpacking the stage since as zhen stated earlier "python is not on the livecd and [releng] would like to keep it that way for size constraints". klieber pointed out "we also need to support both guided and automatic installations from the liveCD. We can't just assume everyone wants a guided install". esammer noted that there is now a point of contact with the Releng project and that this line of communication could be helpful in determining when and how the installer should be run from the LiveCDs and what libraries and utilities will be required. In further discussion of the LiveCD integration, variant suggested that running the command installer at any point should start the installer. In response spyderous noted "that doesn't really work for an automated one, though. maybe it should check for existence of some kickstart-like file". samyron noted "we have to *standardize* on a filename/location". esammer suggested that a special LiveCD may be needed for automated deployments that might probe devices and network interfaces for a configuration. codeman asked about parted being on the LiveCDs and esammer pointed out that parted has not been decided on as the partitioning tool since there are still some architecture issues to be overcome and that it can be discussed outide of the meeting. esammer asked zhen what release the Installer project was aiming for, and zhen answered "esammer: anything 2005". zhen stated that 2005 was not yet planned and would not be planned until after the 2004.3 release. klieber suggested setting an internal date of Dec 31st and codeman answered that a non-partitioning release could be possible by that date. It was then suggested by klieber that if an x86 specific installer could be released sooner than a multi-arch installer, that would be preferable. It was then discussed to drop to a shell on architectures it cannot partition in order to allow manual partitioning. The release date discussion was then postponed in order to discuss current code status. codeman reported his success at installing a laptop using some hacked together scripts to drive the installer with only 3 interventions. samyron reported on the status of the GLIClientController class, with agaffney suggesting the capability to bypass parts of or the whole of the GLIClientController class to allow installation from within a chroot. samyron expressed a desire to review the arch templates as the current design seems unnatural, but it was reiterated that they currently work. samyron then summarized the remaining code work, estimating the installer to be 70% complete:
esammer noted that he could implement the notification system very quickly. klieber suggested a public beta before releasing on a LiveCD, and the idea was generally accepted. esammer asked "can we be tested and working for 2005.0?", with samyron answering "i'm pretty confident", and codeman answering "we can do it". klieber also brought up the topic of documentation, with samyron stating he would write the API documentation, and esammer stating the end-user documentation should wait until things solidify more, with codeman adding there were many people who volunteered to help with documentation who did not show up to the meeting. The release date discussion ended with the target release date as end of year, and a possibility of a public beta by the end of November. No additional questions were brought forth, and the meeting was adjourned at 1930 UTC. The contents of this document are licensed under the Creative Commons - Attribution / Share Alike license. |
|
