From ed0c4d2864b421accf7d39d75f02477685cb581a Mon Sep 17 00:00:00 2001
From: Stefan Karpinski <stefan@karpinski.org>
Date: Wed, 2 Jun 2021 16:27:59 -0400
Subject: [PATCH] read_header_int: allow leading spaces (fix #111)

Some old versions of `tar` (e.g. Solaris) emitted integer fields with
leadings *spaces* which is technically contrary to the standard, but
seems harmless to support. This adds a fragment of this tarball:

    https://sparse.tamu.edu/MM/Oberwolfach/LF10.tar.gz

This is a collection of example sparse matrix files. The fragment is a
single header entry followed by one block of data and is a valid tarball
by itself with the idiosyncratic field formatting of Solaris `tar`.
---
 src/extract.jl              |   7 ++++++-
 test/data/LF10-fragment.tar | Bin 0 -> 1024 bytes
 test/runtests.jl            |   8 ++++++++
 3 files changed, 14 insertions(+), 1 deletion(-)
 create mode 100644 test/data/LF10-fragment.tar

diff --git a/src/extract.jl b/src/extract.jl
index fdc79d5..ecb06f2 100644
--- a/src/extract.jl
+++ b/src/extract.jl
@@ -601,14 +601,19 @@ end
 
 function read_header_int(buf::AbstractVector{UInt8}, offset::Int, length::Int)
     n = UInt64(0)
-    for i in index_range(offset, length)
+    before = true
+    r = index_range(offset, length)
+    for i in r
         byte = buf[i]
+        before && byte == UInt8(' ') && continue
         byte in (0x00, UInt8(' ')) && break
         UInt8('0') <= byte <= UInt8('7') ||
             error("invalid octal digit: $(repr(Char(byte)))")
         n <<= 3
         n |= byte - 0x30
+        before = false
     end
+    before && error("invalid integer value: $(repr(String(buf[r])))")
     return n
 end
 
diff --git a/test/data/LF10-fragment.tar b/test/data/LF10-fragment.tar
new file mode 100644
index 0000000000000000000000000000000000000000..814795e44adbe9c1582b86d6269f2acee68a8c93
GIT binary patch
literal 1024
zcmc&yJx{|h6wI7oaZ5IcZ9-F&u|=I)F#t1qIIq-dzGOS3KOZMBMJ$zC^@g+FhtKcy
zyxk}%wt+rYY3)zZHK3%-r9^}g^E0FtizSo{oXf{7&j6*ehb)4;#xQfu-g&YiS<<(0
z^Wqut+}s&gZ`qn8JMz|?c4Uw2QBMJM+uG7J<QW}#&$huAH*Bejll#jePVm0LYe&{G
zerLhDtt!^uw9OsfOpWJpH=*(1eOCxE3`44o<FwybDVM#l?3k=R2sg=y$%SSoUOw3l
lZMCOR8$Wr_XsG4_^8yO2{>)$Ap?ZRXiUY^DT=l<i{{XD<bZY<r

literal 0
HcmV?d00001

diff --git a/test/runtests.jl b/test/runtests.jl
index 7888b74..206e140 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -867,3 +867,11 @@ if Sys.iswindows() && Sys.which("icacls") !== nothing && VERSION >= v"1.6"
         end
     end
 end
+
+@testset "leading spaces" begin
+    # fragment of https://sparse.tamu.edu/MM/Oberwolfach/LF10.tar.gz
+    tarball = joinpath(@__DIR__, "data", "LF10-fragment.tar")
+    @test Tar.list(tarball) == [
+        Tar.Header("LF10/LF10_B.mtx", :file, 0o100600, 367, "")
+    ]
+end