Why do more widely-adopted OSS take so long to review patches?
Why do more widely-adopted OSS take so long to review patches?
Take Firefox's 1542044, where the dark mode theme colors aren't exposed when they should. The patch has been up for 4 years.
Or, take this, which is part of a series that added Doxygen parsing to clangd, a language server. It was left unreviewed until late 2022, the patcher went in conversation with the reviewer, but then met radio silence again, long enough to the point where the patch-reviewing service shut down. Clangd currently has only 44 open PRs to review, though it uses the same issue tracker as llvm for some reason.
Aren't we paying them to do all this?
Aren’t we paying them to do all this?
No? This question indicates a fundamental lack of understanding of the social relations, power dynamics, and motivations in open source software.
How do I learn more about these things?
One possible starting point could be the now classic essay The Cathedral and the Bazaar:
I don't know, if there's a more general resource, but in the case of Firefox, the donations are so far away from covering the development costs, that they're not even being used for that. Rather, they earn money from search engine deals and are trying to diversify with Pocket, ads, MDN- and VPN-related services etc..
In the case of LLVM, I don't see how they would get many donations to begin with. Maybe Mozilla chips them some of that leftover donation money (they have been doing that with various smaller OSS projects), but I can't imagine much else.
LLVM is probably largely being kept alive by companies or programming language orgs scratching their own itches.This is a good question. I learned it the slow, hard way, back when Apache was “a patchy server”. Maybe someone can suggest books or online resources for getting up to speed quicker.
Aren't we paying them to do all this?
That's the neat part, actually: we don't.
I could be wrong, but I'm pretty sure a lot of open source software is volunteer based and unpaid.
There might be cases where orgs will lend developers to work on a project, but with the org's interests in mind, so if the patch isn't in their interest, then those devs won't look at it.
If you check, there are lots of recently accepted patches to these projects
There can be lots of reasons, including those involved with background discussions that aren't in the bug report.
Sometimes it can be due to patches having a massive impact that are difficult to test but only fixes minor things, or they simply aren't that important.
I wrote a patch for a large project once though that fixed a critical NSIS installer bug when if a specific beta was installed earlier, you couldn't upgrade (but most people were using the betas). The patch was rejected because they wanted to break it up (but I couldn't because I needed to eliminate a big loop by converting it to small ones). So I gave up..
Project stopped a while later.. But this wasn't a normal scenario
How much have you donated to these projects? Trust me, some projects REALLY don't get many donations or volunteers. From my understanding, NTPd (not sure if its still the case), but it was literally being maintained by only 1 person
Aren’t we paying them to do all this?
Are you? Firefox is free to use.
What I meant was donations. As said below, these might not be much.