guix-patches
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[bug#47282] [PATCH 00/13] node going forward


From: Timothy Sample
Subject: [bug#47282] [PATCH 00/13] node going forward
Date: Tue, 30 Mar 2021 01:24:48 -0400
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/27.1 (gnu/linux)

Hi Jelle,

Jelle Licht <jlicht@fsfe.org> writes:

> So, some people seem to be interested in this one; please review and test.

Now that I’ve finally taken the time to dig into what you’ve done here –
I must say it’s very impressive!

I’ve taken the presumptuous step of re-rolling the series.  The reason
is that all the “(delete 'build)” bits were bothering me.  I decided to
have the build system check the “package.json” file for a build script
before trying to run it.  Since that change required changing all the
other patches, I thought it would be easier to just post the updated
patches.  Also, I’m hoping to spare you some trouble (since you’ve
already gone to a lot!).

Of course, this approach gave me free reign to pick nits.  :)  Below is
a list of bigger things that I changed, but I also adjusted some commit
messages, indentation, descriptions, and other minor things.

    • Add the check for a “build” script as explained above, and adjust
      the “npm-build-system” packages accordingly.

    • Rename “libuv-node” to “libuv-for-node”, as this style is used for
      similar packages.  I also changed the name to just “libuv” and
      marked it hidden.

    • Change the “Fix incorrect import semantics” comments to “Fix
      imports for esbuild”.  To me, if TypeScript’s tsc likes the
      imports, they are correct TypeScript (despite the esbuild bug
      report).

    • Set the llhttp version to 2.1.3, and add a patch to fix
      CVE-2020-8287.  The resulting C source files are identical to the
      ones shipped with Node.js 14.16.0.  This makes the tests a little
      simpler, allowing the removal of the HTTP method superset change
      and fixing the reading one byte failure.

    • Fix the SIGXFSZ failure by fixing a “/bin/sh” in the test.

The final result is still a little messy, but I don’t think we should
hold this back any longer.  It’s a significant step forward, and it puts
us in better shape to improve things incrementally.

WDYT?  Let me know if I made anything worse!  :)  If the altered patches
look good to you, I suggest you go ahead and push them.


-- Tim





reply via email to

[Prev in Thread] Current Thread [Next in Thread]