daniel:// stenberg://

1 day ago

daniel:// stenberg://
1 day ago

So tell me, what does a comparison of #wcurl and #wget need to include?

github.com/bagder/docs/blob/ma…

docs/wcurl-vs-wget.md at master · bagder/docs

Internet protocols and tools related documentation - bagder/docs

^GitHub

in reply to daniel:// stenberg://

Wolf480pl

in reply to daniel:// stenberg:// 1 day ago

how the default name of the destination file is computed

This entry was edited (1 day ago)

in reply to daniel:// stenberg://

Todor Petkov

in reply to daniel:// stenberg:// 1 day ago

Wouldn't mind "--report-speed=bits" like in wget, to see the progress in bits. But this is more like for fancy output, not a real future

in reply to Todor Petkov

daniel:// stenberg://

in reply to Todor Petkov 1 day ago

@tppetkov I'm documenting differences.

@Todor Petkov

in reply to daniel:// stenberg://

🍉🌻Different🚀

in reply to daniel:// stenberg:// 1 day ago

i was a wget user, but time comes when you have to curl

in reply to daniel:// stenberg://

Redd

in reply to daniel:// stenberg:// 1 day ago

dependencies, executable file size, usability

in reply to Redd

daniel:// stenberg://

in reply to Redd 1 day ago

@redd how do you compare usability?

@Redd

in reply to daniel:// stenberg://

Redd

in reply to daniel:// stenberg:// 1 day ago

usefulness of error messages when something goes wrong? Maybe test cases like domain cannot be resolved, target not reachable,...?

in reply to daniel:// stenberg://

DistroWatch

in reply to daniel:// stenberg:// 1 day ago

The thing I tend to be most interested in is if the tool can report remote file size without downloading the file.

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

what I miss from curl is downloading multiple files, keeping their file names. WGet does that by default. How do I do: curl -LO example.org/file1.bin example.org/file2.png?

in reply to Petr Menšík

daniel:// stenberg://

in reply to Petr Menšík 1 day ago

@pemensik you either do one -O for each URL since each URL needs info where store them, or you use --remote-name-all which is basically saying "use -O for all URLs"

@Petr Menšík

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

nice. Has --remote-name-all some short variant? I like to use shell expansion like file{1,2,3,4}.bin. Option -O cannot be used with such expansion. Could wcurl do it by default perhaps?

in reply to Petr Menšík

daniel:// stenberg://

in reply to Petr Menšík 1 day ago

@pemensik curl has its own "globbing" that -O works with:

curl -O ".jpg]https://example.com/[1-1000].jpg" etc

@Petr Menšík

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

would it work like pkg{,-test,-dev,-lib}? I use it to download rpm subpackages.

in reply to Petr Menšík

daniel:// stenberg://

in reply to Petr Menšík 1 day ago

@pemensik curl.se/docs/manpage.html#GLOB…

curl - How To Use

^curl.se

@Petr Menšík

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

oh. I have to prevent shell expansion only and curl will make it itself. That is new, thanks.

in reply to Petr Menšík

daniel:// stenberg://

in reply to Petr Menšík 1 day ago

@pemensik a new feature we added the autumn of 1999 😀

@Petr Menšík

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

new to me. But short variant of --remote-name-all would be useful anyway. Quotes could be avoided.

in reply to daniel:// stenberg://

Petr Menšík

in reply to daniel:// stenberg:// 1 day ago

I think #HappyEyeballs in curl should be mentioned. Very useful on lacking IPv6 connectivity. Https DNS record support is not it WGet, is it?

#happyeyeballs

in reply to daniel:// stenberg://

B'ad Samurai 🐐

in reply to daniel:// stenberg:// 1 day ago

A difference? PowerShell v5 says curl and wget are just aliases of Invoke-WebRequest :D

in reply to daniel:// stenberg://

vsz

in reply to daniel:// stenberg:// 1 day ago

With a tiny fix and a few extra steps, I managed to run #wcurl on Win dows: github.com/curl/wcurl/pull/73

replace `>/dev/stderr` with `>&2` for portability by vszakats · Pull Request #73 · curl/wcurl

Fixes this error in environments missing /dev/stderr: wcurl: line 88: can't create /dev/stderr: nonexistent directory Seen it on Windows with busybox.exe: https://frippery.org/busybox/

^GitHub

#wcurl

This entry was edited (1 day ago)

in reply to daniel:// stenberg://

Eric Curtin

in reply to daniel:// stenberg:// 1 day ago

now that wcurl is relatively mature, any thoughts around porting it to something like Rust or Go? Would be useful for platforms like Windows, no dependancy on a Unix shell. I understand there are tradeoffs here.

in reply to Eric Curtin

daniel:// stenberg://

in reply to Eric Curtin 23 hours ago

@ecurtin see github.com/curl/wcurl/issues/7…

Compiled language

Has it been considered to write this wrapper in a compiled language? Or even contribute it back to curl itself?

^{xnox (GitHub)}

@Eric Curtin

in reply to daniel:// stenberg://

Eric Curtin

in reply to daniel:// stenberg:// 11 hours ago

Thank you. PR posted btw: github.com/curl/wcurl/pull/74

Add wcurl in Rust by ericcurtin · Pull Request #74 · curl/wcurl

To have a compiled binary free of dependancies

^GitHub

in reply to Eric Curtin

daniel:// stenberg://

in reply to Eric Curtin 11 hours ago

@ecurtin I noticed, You should know that I'm not a main driver of wcurl, but @samueloph is one of them.

@Eric Curtin @Samuel Henrique

in reply to daniel:// stenberg://

Samuel Henrique

in reply to daniel:// stenberg:// 11 hours ago

@ecurtin thanks for the PR, I want to approach something like this very carefully so I ask for a few weeks to think about this and consider the implications. Just to set the expectations; it might take some months until we get to a point where we can approve/reject the PR.

@Eric Curtin

⇧

daniel:// stenberg:// 1 day ago • •

daniel:// stenberg://
1 day ago