readbytes and readall, replaced by read and readstring per #14660

[samoconnor]

Hi, @kmsquire, @nalimilan, re JuliaIO/GZip.jl#45 (comment): I've made an attempt. I haven't looked at Compat.jl before, so I'm not sure if I'm doing it right.

[yuyichao]

Looks like it's at least reversed. Also I believe only readstring is a new name. Other ones shouldn't be overwirtten.

[samoconnor]

thanks for the tips @yuyichao!

[yuyichao]

And please use a more descriptive commit and PR title.

[jrevels]

@samoconnor Could you rebase against the current master and add some tests? Otherwise, LGTM. This will fix a lot of annoying deprecation warnings for web-related stuff.

[samoconnor]

@jrevels, done

[tkelman]

ERROR: mktempdir has no method matching mktempdir(::Function) on 0.3 (that could be added as a new feature here in a separate PR though), and ERROR: LoadError: MethodError: Cannot convert an object of type Array{Any,1} to an object of type UTF8String on nightly.

[samoconnor]

Hi @tkelman,
I don't have a nightly build to test this against, but it works with the 0.5 build I do have (Version 0.5.0-dev+2236 (2016-01-21 23:04 UTC)).
Is this possibly a type-inference regression in nightly?

This is the code that is failing:

        for text in [
            old_text,
            UTF8String(['A' + i % 52 for i in 1:(div(Base.SZ_UNBUFFERED_IO,2))]),
            UTF8String(['A' + i % 52 for i in 1:(    Base.SZ_UNBUFFERED_IO -1)]),
            UTF8String(['A' + i % 52 for i in 1:(    Base.SZ_UNBUFFERED_IO   )]),
            UTF8String(['A' + i % 52 for i in 1:(    Base.SZ_UNBUFFERED_IO +1)])
        ]

It looks to me that the compiler ought to be able to tell that those are Char arrays:
Char + Int % Int = Char

[tkelman]

A type inference issue seems likely. Adding the Char[...] element type might work around the issue for now, but this could be a bug worth opening on base as well.

[samoconnor]

See JuliaLang/julia@6396218#diff-435745fdfa02a0dc4f3b747ba977af11L180

[stevengj]

This relies on JuliaLang/julia#9017, which was introduced in Julia 0.4.

Probably we should have a Compat shim for that mktempdir feature on 0.3 as well.

[stevengj]

Oh, I see @tkelman commented on this above.

[malmaud]

How's this PR doing? Can I help? Would love to see Requests.jl green again.

[tkelman]

A separate addition of mktempdir should probably be the only thing this needs. We could do that in a separate PR, once that's good to merge we can just restart the CI builds here and they should pass.

[malmaud]

OK, I'll do that today.

On Sun, Feb 7, 2016 at 12:32 PM Tony Kelman notifications@github.com
wrote:

A separate addition of mktempdir should probably be the only thing this
needs. We could do that in a separate PR, once that's good to merge we can
just restart the CI builds here and they should pass.

—
Reply to this email directly or view it on GitHub
#162 (comment).

[samoconnor]

Hi @malmaud, that would me much appreciated.
I'm afraid I'm not at all familiar with Compat.jl, so this PR is just my best attempt.
I do all my work against v0.4.
I occasionally build a v0.5 environment if I'm trying to submit a bug fix, but I don't have any setup for testing against v0.3.
If you find anything else is broken in this PR please feel free to hijack my branch into your own fork and fix it up before merging with master.

[tkelman]

You can download and extract a tarball binary to test against 0.3. It couldn't be any easier.

[malmaud]

@samoconnor OK, I'll take you up on that

[tkelman]

thanks @malmaud, do open a new pr so appveyor runs