Merge pull request #68 from JuliaIO/sk/list-without-mark

Tar.list: refactor to avoid using mark/unmark/reset

Author: StefanKarpinski

src/Tar.jl

@@ -131,12 +131,8 @@ function list(
 )
     raw && strict &&
         error("`raw=true` and `strict=true` options are incompatible")
-    read_hdr = raw ? read_standard_header :
-        let globals = Dict{String,String}()
-            (tar; buf) -> read_header(tar, globals, buf=buf)
-        end
     arg_read(tarball) do tar
-        iterate_headers(callback, tar, read_hdr, strict=strict)
+        iterate_headers(callback, tar, raw=raw, strict=strict)
     end
 end
 

src/create.jl

@@ -26,7 +26,7 @@ function recreate_tarball(
     check_skeleton_header(skeleton, buf=buf)
     globals = Dict{String,String}()
     while !eof(skeleton)
-        hdr = read_header(skeleton, globals, buf=buf, tee=tar)
+        hdr = read_header(skeleton, globals=globals, buf=buf, tee=tar)
         hdr === nothing && break
         check_header(hdr)
         sys_path = joinpath(root, hdr.path)

src/extract.jl

@@ -8,25 +8,34 @@ end
 
 function iterate_headers(
     callback::Function,
-    tar::IO,
-    read_hdr::Function;
-    strict::Bool = true,
+    tar::IO;
+    raw::Bool = false,
+    strict::Bool = !raw,
     buf::Vector{UInt8} = Vector{UInt8}(undef, DEFAULT_BUFFER_SIZE),
 )
-    if !eof(tar)
-        mark(tar)
-        hdr = read_standard_header(tar, buf=buf)
-        if hdr.type == :g && hdr.path == SKELETON_MAGIC
-            skip_data(tar, hdr.size)
-            skeleton = true
-            unmark(tar)
-        else
-            skeleton = false
-            reset(tar)
+    eof(tar) && return
+    hdr = read_standard_header(tar, buf=buf)
+    hdr == nothing && return
+    if !raw
+        globals = Dict{String,String}()
+    end
+    if hdr.type == :g && hdr.path == SKELETON_MAGIC
+        skip_data(tar, hdr.size)
+        skeleton = true
+    else
+        if !raw
+            hdr = read_header(tar, hdr, globals=globals, buf=buf)
         end
+        skeleton = false
+        @goto loop
     end
     while !eof(tar)
-        hdr = read_hdr(tar, buf=buf)
+        hdr = if raw
+            read_standard_header(tar, buf=buf)
+        else
+            read_header(tar, globals=globals, buf=buf)
+        end
+    @label loop
         hdr === nothing && break
         strict && check_header(hdr)
         callback(hdr)
@@ -298,7 +307,7 @@ function read_tarball(
     paths = Dict{String,Union{Symbol,String}}()
     globals = Dict{String,String}()
     while !eof(tar)
-        hdr = read_header(tar, globals, buf=buf, tee=skeleton)
+        hdr = read_header(tar, globals=globals, buf=buf, tee=skeleton)
         hdr === nothing && break
         # check if we should extract or skip
         if !predicate(hdr)
@@ -332,13 +341,23 @@ function read_tarball(
 end
 
 function read_header(
-    io::IO,
-    globals::Dict{String,String} = Dict{String,String}();
+    io::IO;
+    globals::Dict{String,String} = Dict{String,String}(),
     buf::Vector{UInt8} = Vector{UInt8}(undef, DEFAULT_BUFFER_SIZE),
     tee::IO = devnull,
 )
     hdr = read_standard_header(io, buf=buf, tee=tee)
     hdr === nothing && return nothing
+    read_header(io, hdr, globals=globals, buf=buf, tee=tee)
+end
+
+function read_header(
+    io::IO,
+    hdr::Header; # initial header
+    globals::Dict{String,String} = Dict{String,String}(),
+    buf::Vector{UInt8} = Vector{UInt8}(undef, DEFAULT_BUFFER_SIZE),
+    tee::IO = devnull,
+)
     # process zero or more extended headers
     metadata = copy(globals)
     while true

test/runtests.jl

@@ -39,21 +39,42 @@ include("setup.jl")
 end
 
 @testset "empty tarball" begin
-    dir = mktempdir()
-    tarball = Tar.create(dir)
-    rm(dir, recursive=true)
-    @test Tar.list(tarball) == [Tar.Header(".", :directory, 0o755, 0, "")]
-    test_empty_hashes(tarball)
-    dir = Tar.extract(tarball)
-    @test isempty(readdir(dir))
-    rm(dir, recursive=true)
-    open(tarball, append=true) do io
-        write(io, zeros(UInt8, 512))
+    @testset "empty file as tarball" begin
+        tarball = devnull
+        @test Tar.list(tarball) == Tar.Header[]
+        skel = tempname()
+        dir = Tar.extract(tarball, skeleton=skel)
+        @test isempty(readdir(dir))
+        rm(dir, recursive=true)
+        @test isfile(skel)
+        @test Tar.list(skel) == Tar.Header[]
+        @test Tar.list(skel, raw=true) == Tar.Header[]
+        rm(skel)
+    end
+
+    @testset "create an empty tarball" begin
+        dir = mktempdir()
+        tarball = Tar.create(dir)
+        rm(dir, recursive=true)
+        @test Tar.list(tarball) == [Tar.Header(".", :directory, 0o755, 0, "")]
+        @test Tar.list(tarball, raw=true) == [Tar.Header(".", :directory, 0o755, 0, "")]
+        test_empty_hashes(tarball)
+        skel = tempname()
+        dir = Tar.extract(tarball, skeleton=skel)
+        @test isempty(readdir(dir))
+        rm(dir, recursive=true)
+        @test isfile(skel)
+        @test Tar.list(skel) == [Tar.Header(".", :directory, 0o755, 0, "")]
+        @test Tar.list(skel, raw=true) == [Tar.Header(".", :directory, 0o755, 0, "")]
+        rm(skel)
+        open(tarball, append=true) do io
+            write(io, zeros(UInt8, 512))
+        end
+        test_empty_hashes(tarball)
+        dir = Tar.extract(tarball)
+        @test isempty(readdir(dir))
+        rm(dir, recursive=true)
     end
-    test_empty_hashes(tarball)
-    dir = Tar.extract(tarball)
-    @test isempty(readdir(dir))
-    rm(dir, recursive=true)
 end
 
 @testset "test tarball" begin