Microsoft hosts the Windows source in a monstrous 300GB Git repository
'Microsoft has been working to enhance Git to improve the way it handles vast repositories. Central to this effort is a new project released (in part) as open source Git Virtual File system (GVFS). The premise of GVFS is straightforward enough: rather than fetching all the data at once, only a bare skeleton of the repository needs to be populated up front. The virtualized file system subsequently retrieves additional data on a demand-driven, as-needed basis. Building one particular Windows component, for example, will cause GVFS to fetch the files that make up that component, along with anything that the component depends on, but it will stop short of fetching all the many hundreds of gigabytes the repository contains.' -- source: https://arstechnica.com/information-technology/2017/02/microsoft-hosts-the-w... I remember using SourceSafe way back. It may not have been great, but it was a big improvement over "not having any" version control. :-) Cheers, Peter -- Peter Reutemann Dept. of Computer Science University of Waikato, NZ +64 (7) 858-5174 http://www.cms.waikato.ac.nz/~fracpete/ http://www.data-mining.co.nz/
On Tue, 7 Feb 2017 13:15:12 +1300, Peter Reutemann wrote:
I remember using SourceSafe way back. It may not have been great, but it was a big improvement over "not having any" version control. :-)
I have heard horror stories about SourceSafe, namely the unreliable way it did locking. The Windows source tree consists of 270GB in 3.5 million source files <http://www.phoronix.com/scan.php?page=news_item&px=Microsoft-GVFS-Git-Filesystem>. The obvious question is: what does all that source code do? I suspect if you add up all the sources for the packages in Debian, it would not be that large. Yet they offer far more functionality than Windows does. So what is all the Windows source code for? I don’t think it is a question any Microsoft engineer can answer...
I remember using SourceSafe way back. It may not have been great, but it was a big improvement over "not having any" version control. :-)
I have heard horror stories about SourceSafe, namely the unreliable way it did locking.
Funniest thing about people using SourceSafe for the first time was, that they usually did a complete checkout without realizing that they just blocked the whole company from committing any changes... ;-) But I have to say, CVS was pretty bad as well, relying on permissions in the file system whether your user was able to commit a change or not. *shudder*
The Windows source tree consists of 270GB in 3.5 million source files <http://www.phoronix.com/scan.php?page=news_item&px=Microsoft-GVFS-Git-Filesystem>. The obvious question is: what does all that source code do? I suspect if you add up all the sources for the packages in Debian, it would not be that large. Yet they offer far more functionality than Windows does.
So what is all the Windows source code for? I don’t think it is a question any Microsoft engineer can answer...
Not a single person will know that, that's for sure. Cheers, Peter -- Peter Reutemann Dept. of Computer Science University of Waikato, NZ +64 (7) 858-5174 http://www.cms.waikato.ac.nz/~fracpete/ http://www.data-mining.co.nz/
On 7/02/17 1:29 PM, Lawrence D'Oliveiro wrote:
The Windows source tree consists of 270GB in 3.5 million source files <http://www.phoronix.com/scan.php?page=news_item&px=Microsoft-GVFS-Git-Filesystem>. The obvious question is: what does all that source code do? I suspect if you add up all the sources for the packages in Debian, it would not be that large. Yet they offer far more functionality than Windows does.
The Debian source code is over 40 GB[1]. They don't directly host the source control files, that is done by Github and others, but it wouldn't be unreasonable to suggest that it would be well in excess of 300GB.
So what is all the Windows source code for? I don’t think it is a question any Microsoft engineer can answer...
One would hope that different projects (eg Microsoft Office, Visual Studio, SQL Server) would have their own git repos. git clone on a 300GB repo would bring a lot of networks to their knees, especially for the remote workers. -- Simon [1] http://cdimage.debian.org/debian-cd/current/source/iso-dvd/
On Tue, 7 Feb 2017 13:43:33 +1300, Simon Green wrote:
On 7/02/17 1:29 PM, Lawrence D'Oliveiro wrote:
So what is all the Windows source code for? I don’t think it is a question any Microsoft engineer can answer...
One would hope that different projects (eg Microsoft Office, Visual Studio, SQL Server) would have their own git repos.
This is just for Windows.
participants (3)
-
Lawrence D'Oliveiro -
Peter Reutemann -
Simon Green