It’s always seemed poor manners to me to name a fork by appending/incrementing a number - it creates an impression that your fork is more official than it is, and (if the original maintainer ever comes back) puts them in a weird position of having to skip a number the next time they want to do their own major release.
and (if the original maintainer ever comes back) puts them in a weird position of having to skip a number the next time they want to do their own major release.
It does not though? It can create a bit of confusion as e.g. pypdf's history can attest, but you can have foo 2.0 and foo2 2.0 as separate packages with separate versioning.
And at the same time it really helps with discovery, looking for a successor to a package it's much easier to find a foo2 or foo-next than something named completely differently.
he thought that ours could not get popular because it’s on Codeberg instead of on GitHub.
Not that I'm writing any Python nowadays, but... this would be a reason for me to be vary of Pydantic's fork. I prefer my dependencies on reliable platforms that aren't slop infested. GitHub is neither of those.
This is a very stupid, shortsighted reason that will not age well.
It makes me sad every time I hear a popular person making that kind of choice. People staying on X because that's where their audience is (guess why they're there?). People staying on GitHub because they want popularity/stars. People staying with CloudFlare because it's the popular default. And they get the internet they deserve...
The decision to port nearly 1,000 old issues - both open and closed - from the httpx repository is a bit weird. The full comment history of each has been recreated but with one user as the author of every comment (albeit with a note crediting the original author), eg https://github.com/pydantic/httpx2/issues/362
HTTPX maintainer @lovelydinosaur had previously shut down the old issues section on https://github.com/encode/httpx/issues/947 so I understand the desire to make all of that material available again but this is a bit of a weird way to do it. I would have gone with a static set of archived pages instead.
httpx v1.0 development seems to have paused, but it was threatening to ship quite substantial breaking changes - which would have been a big problem for the Python ecosystem for reasons I explained at length here.
In that comment I actually suggested they ship their new changes as a new package called httpx2 to allow users to have the two incompatible versions installed at the same time.
Gaelan | a day ago
It’s always seemed poor manners to me to name a fork by appending/incrementing a number - it creates an impression that your fork is more official than it is, and (if the original maintainer ever comes back) puts them in a weird position of having to skip a number the next time they want to do their own major release.
masklinn | a day ago
It does not though? It can create a bit of confusion as e.g. pypdf's history can attest, but you can have foo 2.0 and foo2 2.0 as separate packages with separate versioning.
And at the same time it really helps with discovery, looking for a successor to a package it's much easier to find a foo2 or foo-next than something named completely differently.
algernon | a day ago
Not that I'm writing any Python nowadays, but... this would be a reason for me to be vary of Pydantic's fork. I prefer my dependencies on reliable platforms that aren't slop infested. GitHub is neither of those.
This is a very stupid, shortsighted reason that will not age well.
xvello | a day ago
Pydantic is a VC-backed company embracing code extrusion. Of course they need to stay on github for value-signaling purposes.
viraptor | 21 hours ago
It makes me sad every time I hear a popular person making that kind of choice. People staying on X because that's where their audience is (guess why they're there?). People staying on GitHub because they want popularity/stars. People staying with CloudFlare because it's the popular default. And they get the internet they deserve...
nikaro | 22 hours ago
This a quote from the Starlette author, not Pydantic.
algernon | 22 hours ago
Said person is working for Pydantic, though, aren't they, and was speaking in that capacity?
nikaro | 19 hours ago
You are right, i was not aware.
JaDogg | a day ago
Nice to see people adulting instead of going mine mine mine.
simonw | 21 hours ago
The decision to port nearly 1,000 old issues - both open and closed - from the httpx repository is a bit weird. The full comment history of each has been recreated but with one user as the author of every comment (albeit with a note crediting the original author), eg https://github.com/pydantic/httpx2/issues/362
HTTPX maintainer
@lovelydinosaurhad previously shut down the old issues section on https://github.com/encode/httpx/issues/947 so I understand the desire to make all of that material available again but this is a bit of a weird way to do it. I would have gone with a static set of archived pages instead.simonw | 21 hours ago
httpx v1.0 development seems to have paused, but it was threatening to ship quite substantial breaking changes - which would have been a big problem for the Python ecosystem for reasons I explained at length here.
In that comment I actually suggested they ship their new changes as a new package called
httpx2to allow users to have the two incompatible versions installed at the same time.