| layout | title |
|---|---|
default |
SoC 2022 Applicant Microprojects |
First make sure you read and understand our general guidelines and suggestions for microprojects.
There are some suggestions on how you can find some microprojects on your own in the document.
"git diff" shows the function name corresponding to each hunk after the @@ ... @@ line. For common languages (C, HTML, Ada, Matlab, ...), the way to find the function name is built-in Git's source code as regular expressions (see userdiff.c). A few languages are common enough to deserve a built-in driver, but are not yet recognized. For example, shell.
This project requires a very good knowledge of regular expressions.
It is easy though to find examples of how this can be done by searching the code base and the mailing list archive, as this has already been done for a number of languages.
See for example what Junio did in ffcb4e94d3 (bisect: do not run show-branch just to show the current commit, 2021-07-27).
If you can't find one please tell us, along with the command you used to search, so that we can remove this microproject idea.
Find one test script that verifies the presence/absence of
files/directories with 'test -(e|f|d|...)' and replace them with the
appropriate test_path_is_file, test_path_is_dir, etc. helper
functions.
If you can't find one please tell us, along with the command you used to search, so that we can remove this microproject idea.
The Git project uses a large collection of integration tests written in
Shell to guard against regressions when adding new features or fixing
bugs. The scripts in question can be found in the t directory
here.
While it is perfectly OK to use pipes when writing integration tests, we must be careful to avoid writing a pipeline that suppresses the exit code of a Git process, like so:
git <subcommand> | <some other command>
...since the exit code of git <subcommand> would be suppressed by the
pipe. If git <subcommand> crashed, we would not catch it in the above
example when running the integration suite.
Other examples to avoid include:
# bad:
<some command> $(git <subcommand>)
# also bad:
<some command> <<EOF
... some text ...
$(git <subcommand>)
EOF
...since the exit code of git <subcommand> is hidden behind the
subshell in both instances.
On the other hand, both of the following examples are OK, since neither
hides the exit code of running git <subcommand>:
# good:
var=$(git <subcommand>)
# also good:
<some command> | <some other command> | git <subcommand>
(provided that neither <some command> or <some other command> are
git).
See the commit c6f44e1da5 for example, and then do the same thing in one other test script.
If you can't find one please tell us, along with the command you used to search, so that we can remove this microproject idea.
Pick one field of a structure that (1) is of signed integral type and (2) is used as a collection of multiple bits. Discuss if there is a good reason why it has to be a signed integral field and change it to an unsigned type otherwise. [thread]
Even though the amount of code to write is small, these projects involve a lot of prior work to understand the specification and deal with all potential corner-cases.
A number of our test scripts have been written a long time ago in a style that is now outdated.
In the following email it is explained in details how to modernize and clean up the t7001 test script:
https://lore.kernel.org/git/CAPig+cQpUu2UO-+jWn1nTaDykWnxwuEitzVB7PnW2SS_b7V8Hg@mail.gmail.com/
t7001 is not the only test script where similar changes could be made though.
Find one test script that needs some of the same changes and make them. Please make sure that the test script is not already being worked on by asking on the mailing list before starting to work on it.
There should be only one kind of change per commit. For example if one of your commits indents test bodies with TABs, instead of spaces, then this should be the only kind of change in this commit.