Merge remote-tracking branch 'origin/master' into github-api-token

src/libutil/archive.cc

@@ -27,6 +27,8 @@ struct ArchiveSettings : Config
         #endif
         "use-case-hack",
         "Whether to enable a Darwin-specific hack for dealing with file name collisions."};
+    Setting<bool> preallocateContents{this, true, "preallocate-contents",
+        "Whether to preallocate files when writing objects with known size."};
 };
 
 static ArchiveSettings archiveSettings;
@@ -66,9 +68,7 @@ static void dump(const Path & path, Sink & sink, PathFilter & filter)
 {
     checkInterrupt();
 
-    struct stat st;
-    if (lstat(path.c_str(), &st))
-        throw SysError("getting attributes of path '%1%'", path);
+    auto st = lstat(path);
 
     sink << "(";
 
@@ -325,6 +325,9 @@ struct RestoreSink : ParseSink
 
     void preallocateContents(uint64_t len)
     {
+        if (!archiveSettings.preallocateContents)
+            return;
+
 #if HAVE_POSIX_FALLOCATE
         if (len) {
             errno = posix_fallocate(fd.get(), 0, len);

src/libutil/args.hh

@@ -192,7 +192,7 @@ public:
     {
         expectArgs({
             .label = label,
-            .optional = true,
+            .optional = optional,
             .handler = {dest}
         });
     }

src/libutil/callback.hh

@@ -0,0 +1,46 @@
+#pragma once
+
+#include <future>
+#include <functional>
+
+namespace nix {
+
+/* A callback is a wrapper around a lambda that accepts a valid of
+   type T or an exception. (We abuse std::future<T> to pass the value or
+   exception.) */
+template<typename T>
+class Callback
+{
+    std::function<void(std::future<T>)> fun;
+    std::atomic_flag done = ATOMIC_FLAG_INIT;
+
+public:
+
+    Callback(std::function<void(std::future<T>)> fun) : fun(fun) { }
+
+    Callback(Callback && callback) : fun(std::move(callback.fun))
+    {
+        auto prev = callback.done.test_and_set();
+        if (prev) done.test_and_set();
+    }
+
+    void operator()(T && t) noexcept
+    {
+        auto prev = done.test_and_set();
+        assert(!prev);
+        std::promise<T> promise;
+        promise.set_value(std::move(t));
+        fun(promise.get_future());
+    }
+
+    void rethrow(const std::exception_ptr & exc = std::current_exception()) noexcept
+    {
+        auto prev = done.test_and_set();
+        assert(!prev);
+        std::promise<T> promise;
+        promise.set_exception(exc);
+        fun(promise.get_future());
+    }
+};
+
+}

src/libutil/config.cc

@@ -1,6 +1,6 @@
 #include "config.hh"
 #include "args.hh"
-#include "abstractsettingtojson.hh"
+#include "abstract-setting-to-json.hh"
 
 #include <nlohmann/json.hpp>
 

src/libutil/fmt.hh

@@ -1,7 +1,6 @@
 #pragma once
 
 #include <boost/format.hpp>
-#include <boost/algorithm/string/replace.hpp>
 #include <string>
 #include "ansicolor.hh"
 

src/libutil/serialise.cc

@@ -93,7 +93,7 @@ void Source::operator () (unsigned char * data, size_t len)
 }
 
 
-std::string Source::drain()
+void Source::drainInto(Sink & sink)
 {
     std::string s;
     std::vector<unsigned char> buf(8192);
@@ -101,12 +101,19 @@ std::string Source::drain()
         size_t n;
         try {
             n = read(buf.data(), buf.size());
-            s.append((char *) buf.data(), n);
+            sink(buf.data(), n);
         } catch (EndOfFile &) {
             break;
         }
     }
-    return s;
+}
+
+
+std::string Source::drain()
+{
+    StringSink s;
+    drainInto(s);
+    return *s.s;
 }
 
 

src/libutil/serialise.hh

@@ -69,6 +69,8 @@ struct Source
 
     virtual bool good() { return true; }
 
+    void drainInto(Sink & sink);
+
     std::string drain();
 };
 
@@ -340,7 +342,7 @@ T readNum(Source & source)
         ((uint64_t) buf[6] << 48) |
         ((uint64_t) buf[7] << 56);
 
-    if (n > std::numeric_limits<T>::max())
+    if (n > (uint64_t)std::numeric_limits<T>::max())
         throw SerialisationError("serialised integer %d is too large for type '%s'", n, typeid(T).name());
 
     return (T) n;
@@ -404,4 +406,93 @@ struct StreamToSourceAdapter : Source
 };
 
 
+/* A source that reads a distinct format of concatenated chunks back into its
+   logical form, in order to guarantee a known state to the original stream,
+   even in the event of errors.
+
+   Use with FramedSink, which also allows the logical stream to be terminated
+   in the event of an exception.
+*/
+struct FramedSource : Source
+{
+    Source & from;
+    bool eof = false;
+    std::vector<unsigned char> pending;
+    size_t pos = 0;
+
+    FramedSource(Source & from) : from(from)
+    { }
+
+    ~FramedSource()
+    {
+        if (!eof) {
+            while (true) {
+                auto n = readInt(from);
+                if (!n) break;
+                std::vector<unsigned char> data(n);
+                from(data.data(), n);
+            }
+        }
+    }
+
+    size_t read(unsigned char * data, size_t len) override
+    {
+        if (eof) throw EndOfFile("reached end of FramedSource");
+
+        if (pos >= pending.size()) {
+            size_t len = readInt(from);
+            if (!len) {
+                eof = true;
+                return 0;
+            }
+            pending = std::vector<unsigned char>(len);
+            pos = 0;
+            from(pending.data(), len);
+        }
+
+        auto n = std::min(len, pending.size() - pos);
+        memcpy(data, pending.data() + pos, n);
+        pos += n;
+        return n;
+    }
+};
+
+/* Write as chunks in the format expected by FramedSource.
+
+   The exception_ptr reference can be used to terminate the stream when you
+   detect that an error has occurred on the remote end.
+*/
+struct FramedSink : nix::BufferedSink
+{
+    BufferedSink & to;
+    std::exception_ptr & ex;
+
+    FramedSink(BufferedSink & to, std::exception_ptr & ex) : to(to), ex(ex)
+    { }
+
+    ~FramedSink()
+    {
+        try {
+            to << 0;
+            to.flush();
+        } catch (...) {
+            ignoreException();
+        }
+    }
+
+    void write(const unsigned char * data, size_t len) override
+    {
+        /* Don't send more data if the remote has
+            encountered an error. */
+        if (ex) {
+            auto ex2 = ex;
+            ex = nullptr;
+            std::rethrow_exception(ex2);
+        }
+        to << len;
+        to(data, len);
+    };
+};
+
+
 }

src/libutil/split.hh

@@ -9,7 +9,7 @@ namespace nix {
 
 // If `separator` is found, we return the portion of the string before the
 // separator, and modify the string argument to contain only the part after the
-// separator. Otherwise, wer return `std::nullopt`, and we leave the argument
+// separator. Otherwise, we return `std::nullopt`, and we leave the argument
 // string alone.
 static inline std::optional<std::string_view> splitPrefixTo(std::string_view & string, char separator) {
     auto sepInstance = string.find(separator);

src/libutil/url-parts.hh

@@ -0,0 +1,44 @@
+#pragma once
+
+#include <string>
+#include <regex>
+
+namespace nix {
+
+// URI stuff.
+const static std::string pctEncoded = "(?:%[0-9a-fA-F][0-9a-fA-F])";
+const static std::string schemeRegex = "(?:[a-z+.-]+)";
+const static std::string ipv6AddressRegex = "(?:\\[[0-9a-fA-F:]+\\])";
+const static std::string unreservedRegex = "(?:[a-zA-Z0-9-._~])";
+const static std::string subdelimsRegex = "(?:[!$&'\"()*+,;=])";
+const static std::string hostnameRegex = "(?:(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + ")*)";
+const static std::string hostRegex = "(?:" + ipv6AddressRegex + "|" + hostnameRegex + ")";
+const static std::string userRegex = "(?:(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + "|:)*)";
+const static std::string authorityRegex = "(?:" + userRegex + "@)?" + hostRegex + "(?::[0-9]+)?";
+const static std::string pcharRegex = "(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + "|[:@])";
+const static std::string queryRegex = "(?:" + pcharRegex + "|[/? \"])*";
+const static std::string segmentRegex = "(?:" + pcharRegex + "+)";
+const static std::string absPathRegex = "(?:(?:/" + segmentRegex + ")*/?)";
+const static std::string pathRegex = "(?:" + segmentRegex + "(?:/" + segmentRegex + ")*/?)";
+
+// A Git ref (i.e. branch or tag name).
+const static std::string refRegexS = "[a-zA-Z0-9][a-zA-Z0-9_.-]*"; // FIXME: check
+extern std::regex refRegex;
+
+// Instead of defining what a good Git Ref is, we define what a bad Git Ref is
+// This is because of the definition of a ref in refs.c in https://github.com/git/git
+// See tests/fetchGitRefs.sh for the full definition
+const static std::string badGitRefRegexS = "//|^[./]|/\\.|\\.\\.|[[:cntrl:][:space:]:?^~\[]|\\\\|\\*|\\.lock$|\\.lock/|@\\{|[/.]$|^@$|^$";
+extern std::regex badGitRefRegex;
+
+// A Git revision (a SHA-1 commit hash).
+const static std::string revRegexS = "[0-9a-fA-F]{40}";
+extern std::regex revRegex;
+
+// A ref or revision, or a ref followed by a revision.
+const static std::string refAndOrRevRegex = "(?:(" + revRegexS + ")|(?:(" + refRegexS + ")(?:/(" + revRegexS + "))?))";
+
+const static std::string flakeIdRegexS = "[a-zA-Z][a-zA-Z0-9_-]*";
+extern std::regex flakeIdRegex;
+
+}

src/libutil/url.cc

@@ -1,4 +1,5 @@
 #include "url.hh"
+#include "url-parts.hh"
 #include "util.hh"
 
 namespace nix {

src/libutil/url.hh

@@ -2,8 +2,6 @@
 
 #include "error.hh"
 
-#include <regex>
-
 namespace nix {
 
 struct ParsedURL
@@ -29,40 +27,4 @@ std::map<std::string, std::string> decodeQuery(const std::string & query);
 
 ParsedURL parseURL(const std::string & url);
 
-// URI stuff.
-const static std::string pctEncoded = "(?:%[0-9a-fA-F][0-9a-fA-F])";
-const static std::string schemeRegex = "(?:[a-z+.-]+)";
-const static std::string ipv6AddressRegex = "(?:\\[[0-9a-fA-F:]+\\])";
-const static std::string unreservedRegex = "(?:[a-zA-Z0-9-._~])";
-const static std::string subdelimsRegex = "(?:[!$&'\"()*+,;=])";
-const static std::string hostnameRegex = "(?:(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + ")*)";
-const static std::string hostRegex = "(?:" + ipv6AddressRegex + "|" + hostnameRegex + ")";
-const static std::string userRegex = "(?:(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + "|:)*)";
-const static std::string authorityRegex = "(?:" + userRegex + "@)?" + hostRegex + "(?::[0-9]+)?";
-const static std::string pcharRegex = "(?:" + unreservedRegex + "|" + pctEncoded + "|" + subdelimsRegex + "|[:@])";
-const static std::string queryRegex = "(?:" + pcharRegex + "|[/? \"])*";
-const static std::string segmentRegex = "(?:" + pcharRegex + "+)";
-const static std::string absPathRegex = "(?:(?:/" + segmentRegex + ")*/?)";
-const static std::string pathRegex = "(?:" + segmentRegex + "(?:/" + segmentRegex + ")*/?)";
-
-// A Git ref (i.e. branch or tag name).
-const static std::string refRegexS = "[a-zA-Z0-9][a-zA-Z0-9_.-]*"; // FIXME: check
-extern std::regex refRegex;
-
-// Instead of defining what a good Git Ref is, we define what a bad Git Ref is
-// This is because of the definition of a ref in refs.c in https://github.com/git/git
-// See tests/fetchGitRefs.sh for the full definition
-const static std::string badGitRefRegexS = "//|^[./]|/\\.|\\.\\.|[[:cntrl:][:space:]:?^~\[]|\\\\|\\*|\\.lock$|\\.lock/|@\\{|[/.]$|^@$|^$";
-extern std::regex badGitRefRegex;
-
-// A Git revision (a SHA-1 commit hash).
-const static std::string revRegexS = "[0-9a-fA-F]{40}";
-extern std::regex revRegex;
-
-// A ref or revision, or a ref followed by a revision.
-const static std::string refAndOrRevRegex = "(?:(" + revRegexS + ")|(?:(" + refRegexS + ")(?:/(" + revRegexS + "))?))";
-
-const static std::string flakeIdRegexS = "[a-zA-Z][a-zA-Z0-9_-]*";
-extern std::regex flakeIdRegex;
-
 }

src/libutil/util.hh

@@ -12,13 +12,9 @@
 #include <signal.h>
 
 #include <functional>
-#include <limits>
-#include <cstdio>
 #include <map>
 #include <sstream>
 #include <optional>
-#include <future>
-#include <iterator>
 
 #ifndef HAVE_STRUCT_DIRENT_D_TYPE
 #define DT_UNKNOWN 0
@@ -480,43 +476,8 @@ std::optional<typename T::mapped_type> get(const T & map, const typename T::key_
 }
 
 
-/* A callback is a wrapper around a lambda that accepts a valid of
-   type T or an exception. (We abuse std::future<T> to pass the value or
-   exception.) */
 template<typename T>
-class Callback
-{
-    std::function<void(std::future<T>)> fun;
-    std::atomic_flag done = ATOMIC_FLAG_INIT;
-
-public:
-
-    Callback(std::function<void(std::future<T>)> fun) : fun(fun) { }
-
-    Callback(Callback && callback) : fun(std::move(callback.fun))
-    {
-        auto prev = callback.done.test_and_set();
-        if (prev) done.test_and_set();
-    }
-
-    void operator()(T && t) noexcept
-    {
-        auto prev = done.test_and_set();
-        assert(!prev);
-        std::promise<T> promise;
-        promise.set_value(std::move(t));
-        fun(promise.get_future());
-    }
-
-    void rethrow(const std::exception_ptr & exc = std::current_exception()) noexcept
-    {
-        auto prev = done.test_and_set();
-        assert(!prev);
-        std::promise<T> promise;
-        promise.set_exception(exc);
-        fun(promise.get_future());
-    }
-};
+class Callback;
 
 
 /* Start a thread that handles various signals. Also block those signals