Welcome to the 50th edition of Git Rev News, a digest of all things Git. For our goals, the archives, the way we work, and how to contribute or to subscribe, see the Git Rev News page on git.github.io.
This edition covers what happened during the month of March 2019.
Fabio Aiuto asked for support on the Git mailing list as he was trying to build the first commit of Git made by Linus in April 2005, but was getting “undefined reference to symbol” errors related to the OpenSSL library.
Santiago Torres replied suggesting him to use a toolchain from around the time the commit was made and using an earlier OpenSSL library.
Fabio replied that he would have liked to “to debug around that simple version, to understand how everything works”.
Santiago suggested playing with libgit2, that has a smaller code base, and its python bindings that “abstract the memory management and other low-level stuff”.
Jeff King, alias Peff, suggested linking with -lcrypto instead of
-lssl and also adding -lz to the Makefile or the command line.
He said that make LIBS='-lcrypto -lz' works for him and that he
“used periodically check that Git v1.0 can fetch happily from GitHub”.
asymptosis chimed in to suggest working through the “Git Internals” chapter in the Git Book to understand how Git works. He also pointed to a Git re-implementation as simple bash scripts he had started to understand that chapter.
Jonathan Nieder recommended reading the Hacking Git chapter in the user manual documentation that comes with Git.
Peff then reported that he managed again to build Git version 1.0.0
from December 2005 with make NO_OPENSSL=Nope and to use it to fetch
from GitHub.
Fabio replied that he also succeeded in building the first commit of Git made by Linus.
Fabio, Santiago and Peff then discussed the fact that it was still possible to build a 15 year old codebase with a modern toolchain. Peff explained that it’s possible because “Git has very few external dependencies”, and other than that, Git “just depends on a reasonable C compiler and a POSIX libc, both of which have been standardized for decades”.
Then Fabio reported a segfault in the version he had built. Peff reproduced it and showed how to work around it.
Fabio thanked Peff and asymptosis. He said he would “go on studying Git this way, and follow all the improvements that were made along its history”.
Various
Light reading
Git tools and sites
This edition of Git Rev News was curated by Christian Couder <christian.couder@gmail.com>, Jakub Narębski <jnareb@gmail.com>, Markus Jansen <mja@jansen-preisler.de> and Gabriel Alcaras <gabriel.alcaras@telecom-paristech.fr> with help from Luca Milanesio.