Including files

>
>>> How would this know about a filepath specified on the command line,
>>> for example?  (see gst_smalltalk_args and process_file in lib.c)
>>>    
>>
>> It wouldn't.
>
> The obvious solution being to call FileStream>>#fileIn: there too.  We
> only need to use process_file to bootstrap the system, everything else
> had better be done with Smalltalk call-ins.  We're actually doing this
> more often in post-2.2 development versions, to allow filing in a
> ReadStream.
>
>> I'm afraid I'm having difficulty coming to grips with the
>> recent fad for doing everything outside of Smalltalk.
>>  
>
> ???

>Stephen Compall wrote:
> > Mike Anderson wrote:
> >
> >>Adding a primitive would be overkill, wouldn't it? Surely you just need
> >>a small change to FileStream >> #fileIn:?
> >
> >
> > How would this know about a filepath specified on the command line,
> > for example?  (see gst_smalltalk_args and process_file in lib.c)
>
>It wouldn't. I'm afraid I'm having difficulty coming to grips with the
>recent fad for doing everything outside of Smalltalk.
>
>Mike
>
>
>
>_______________________________________________
>help-smalltalk mailing list
>[hidden email]
>http://lists.gnu.org/mailman/listinfo/help-smalltalk

>>>Yes, touch doesn't update the timestamp. It seems that the open-close is
>>>optimized out of existence. I tried other means, but got the same
>>>results (like adding a position 0 before the close).
>>>A primitive using utimes would fix it, but I'm not sure how to do it (at
>>>least I haven't tried it). I tried to utimes on Linux and Mac and it
>>>worked beautifully, not sure about other platforms.
>>>
>>man utime says it's POSIX. I'd be in favour of adding #utime and
>>removing #touch if anyone cared to do the work ;)
>Like this?  I will commit this as soon as you confirm it's what you could
>have done.
>
>That was just laziness on my side.  I recall I was using Cygwin at the
>time I added File>>#touch.  When I added VFS support, that one method had
>not changed, so I didn't bother testing if it still worked.
>
>Thanks all for the activity that is going on these days!
>
>Paolo
>
>
>_______________________________________________
>help-smalltalk mailing list
>[hidden email]
>http://lists.gnu.org/mailman/listinfo/help-smalltalk

>
>>>man utime says it's POSIX. I'd be in favour of adding #utime and
>>>removing #touch if anyone cared to do the work ;)
>>Like this?  I will commit this as soon as you confirm it's what you could
>>have done.
>Like this I meant.
>
>Paolo
>
>
>--- orig/kernel/File.st
>+++ mod/kernel/File.st
>@@ -251,6 +251,18 @@ size
>      ^vfsHandler size
>  !
>
>+lastAccessTime: aDateTime
>+    "Update the last access time of the file corresponding to the receiver,
>+     to be aDateTime."
>+    vfsHandler lastAccessTime: aDateTime lastModifyTime: self lastModifyTime
>+!
>+
>+lastAccessTime: accessDateTime lastModifyTime: modifyDateTime
>+    "Update the timestamps of the file corresponding to the receiver, to be
>+     accessDateTime and modifyDateTime."
>+    vfsHandler lastAccessTime: accessDateTime lastModifyTime: modifyDateTime
>+!
>+
>  lastAccessTime
>      "Answer the last access time of the file identified by the receiver"
>      ^vfsHandler lastAccessTime
>@@ -272,6 +284,12 @@ creationTime
>      ^vfsHandler creationTime
>  !
>
>+lastModifyTime: aDateTime
>+    "Update the last modification timestamp of the file corresponding to the
>+     receiver, to be aDateTime."
>+    vfsHandler lastAccessTime: self lastAccessTime lastModifyTime: aDateTime
>+!
>+
>  lastModifyTime
>      "Answer the last modify time of the file identified by the receiver
>      (the `last modify time' has to do with the actual file contents)."
>@@ -381,8 +399,10 @@ contents
>  !
>
>  touch
>-    "Update the timestamp of the file with the given path name."
>-    (self openDescriptor: FileStream append) close
>+    "Update the timestamp of the file corresponding to the receiver."
>+    | now |
>+    now := DateTime now.
>+    self lastAccessTime: now lastModifyTime: now
>  !
>
>  open: mode
>
>
>--- orig/kernel/VFS.st
>+++ mod/kernel/VFS.st
>@@ -168,6 +168,9 @@ rewindDir: dirObject
>
>  !RealFileHandler class methodsFor: 'C call-outs'!
>
>+setTimeFor: file atime: atimeSeconds mtime: mtimeSeconds
>+    <cCall: 'utime' returning: #int args: #(#string #long #long)>!
>+
>  working
>      <cCall: 'getCurDirName' returning: #stringOut args: #()>! !
>
>@@ -357,6 +360,12 @@ isAccessible
>
>  !VFSHandler methodsFor: 'file operations'!
>
>+lastAccessTime: accessDateTime lastModifyTime: modifyDateTime
>+    "Set the receiver's timestamps to be accessDateTime and modifyDateTime.
>+     If your file system does not support distinct access and modification
>+     times, you should discard accessDateTime."
>+    self subclassResponsibility!
>+
>  open: class mode: mode ifFail: aBlock
>      "Open the receiver in the given mode (as answered by FileStream's
>      class constant methods)"
>@@ -518,6 +527,14 @@ isExecutable
>
>  !RealFileHandler methodsFor: 'file operations'!
>
>+lastAccessTime: accessDateTime lastModifyTime: modifyDateTime
>+    "Set the receiver's timestamps to be accessDateTime and modifyDateTime."
>+    self class
>+       setTimeFor: self realFileName
>+       atime: (self secondsFromDateTime: accessDateTime)
>+       mtime: (self secondsFromDateTime: modifyDateTime).
>+    File checkError!
>+
>  open: class mode: mode ifFail: aBlock
>      "Open the receiver in the given mode (as answered by FileStream's
>      class constant methods)"
>@@ -541,9 +558,16 @@ renameTo: newFileName
>
>  !RealFileHandler methodsFor: 'private'!
>
>-getDateAndTime: time
>+secondsFromDateTime: aDateTime
>      "Private - Convert a time expressed in seconds from 1/1/2000 to
>       an array of two Smalltalk Date and Time objects"
>+    ^(aDateTime asSeconds - Epoch asSeconds)
>+       - (aDateTime offset asSeconds - Epoch offset asSeconds)
>+!
>+
>+getDateAndTime: time
>+    "Private - Convert a time expressed in seconds from 1/1/2000 to
>+     a Smalltalk DateTime object."
>
>      ^(Epoch + (Duration seconds: time))
>         offset: (Duration seconds: Time timezoneBias)
>
>
>--- orig/libgst/cint.c
>+++ mod/libgst/cint.c
>@@ -200,6 +200,9 @@ static OOP classify_type_symbol (OOP sym
>  static int get_errno (void);
>
>  /* Encapsulate binary incompatibilities between various C libraries.  */
>+static int my_utime (const char *name,
>+                    long new_atime,
>+                    long new_mtime);
>  static int my_stat (const char *name,
>                     gst_stat * out);
>  static int my_lstat (const char *name,
>@@ -306,6 +309,22 @@ get_errno (void)
>  }
>
>  int
>+my_utime (const char *name,
>+         long new_atime,
>+         long new_mtime)
>+{
>+  struct timeval times[2];
>+  int result;
>+  times[0].tv_sec = new_atime + 86400 * 10957;
>+  times[1].tv_sec = new_mtime + 86400 * 10957;
>+  times[0].tv_usec = times[1].tv_usec = 0;
>+  result = utimes (name, times);
>+  if (!result)
>+    errno = 0;
>+  return (result);
>+}
>+
>+int
>  my_stat (const char *name,
>         gst_stat * out)
>  {
>@@ -497,6 +516,7 @@ _gst_init_cfuncs (void)
>    _gst_define_cfunc ("strerror", strerror);
>    _gst_define_cfunc ("stat", my_stat);
>    _gst_define_cfunc ("lstat", my_lstat);
>+  _gst_define_cfunc ("utime", my_utime);
>
>    _gst_define_cfunc ("opendir", my_opendir);
>    _gst_define_cfunc ("closedir", closedir);
>
>
>
>_______________________________________________
>help-smalltalk mailing list
>[hidden email]
>http://lists.gnu.org/mailman/listinfo/help-smalltalk

>I also fixed the other bugs you reported, thanks.
>>The touch utility creates the file unless you specify nocreate.
>>
>>It's not that hard to create primitives, is it?
>???
>
>As for the next tarball, it will be 2.2a (2.3 prerelease, that is) and
>I'll try to build it this week, or as soon as I get to test the remaining
>Unicode changes.
>
>Paolo