RE: FW: Win32 File Name Compares and Normalization

[Matthew Von-Maszewski]

All,

Ah hell, I have identified a third scenario of case sensitivity problems.  I
need to review both the current HAVE_CASE_INSENSITIVE_FS and my stuff in
light of this before submitting a formal code change.  Amazing what unit
test cases can do for you!

Later,
Matthew

-----Original Message-----
From: Rob Tulloh [mailto:address@hidden
Sent: Wednesday, September 12, 2001 4:16 PM
To: Matthew Von-Maszewski
Cc: address@hidden
Subject: Re: FW: Win32 File Name Compares and Normalization


Matthew,

Please resubmit the changes you propose with the approriate
wrappers of HAVE_CASE_INSENSITIVE_FS where needed.
With this, we can review what you propose in light of your
objectives below.

I can relate to what you are trying to
accomplish, but there are several ways which this can be
addressed when using GNU make on WIN32 platforms. One
example, you talk about single process access to pdb files.
May I suggest that this can be avoided by not using pdb
files. Instead, you can compile with -Z7 -Yd and put the
debug info directly into the obj files. The advantage here is that
you no longer depend on pdb files being up-to-date. I can't
tell you the number of times we have had to redo big
debug builds because one member of a pdb file was out-of-date
wrt to the rest of the build. Another approach is to put debug
into separate pdb files. This has the advantage of retaining
pdb format, but you now have to manage a lot more files
than before.

I am not saying the above are ideal solutions to the problem,
but they are possible solutions. We also use GNU make with
-j to build in parallel and our build environment is engineered
in such a way that it works just fine. Also, Visual Studio would
never be able to manage the complexity of what we have to
build. It can only be done from the command line and we can
do it there in a portable and scalable fashion.

Regards,

Rob Tulloh

Matthew Von-Maszewski wrote:

> 99% pure ...
>
> Ok, but I still have exception conditions based upon what is probably an
> unorthodox usage ... but necessary.  So I would still appreciate some
> consideration for this patch as part of the HAVE_CASE_INSENSITIVITY_FS
> conditional.
>
> Goal 1:  Have Microsoft C build a lib or an .exe using one invocation of
the
> C compiler.
>
> Reason 1:  .pdb/.idb/.pch files only allow single processor access.  Turn
on
> the gnumake -j2 flag and everything falls apart using traditional .c to
.obj
> to .exe constructions.
>
> Reason 2:  Microsoft C is much faster when it only has to load the
> precompiled header once for several files.
>
> Goal 2:  Have Microsoft build as fast or faster from command line than
from
> Visual Studio to ensure programmer adoption.
>
> I achieve both goals by using the filename to construct a variable name,
> e.g. f:/bob/fred.c becomes $(f:/bob/fred.c.DEPS).  This happens because I
> want the build environment (gnumake and a template) to do all the
thinking,
> not the programmer.  I must apply all of the prerequisites of a potential
> .obj directly against the .exe to ensure the compiler is invoked only when
> necessary.
>
> This is extremely possible when the file names are predictable.  My source
> for the file names and dependencies is not case predictable:  Microsoft /P
> output.  So I must create a normalized form.  This means normalizing
> $(CURDIR) as well as $(wildcard) outputs.
>
> Then the written HAVE_CASE_INSENSITIVE_FS code covers some bases that I do
> not relating to potential programmer typing errors.  I was not concerned
> here since my makefile setup is to run on Linux too using the same files
> pulled from CVS.  Here, the case problems would cause the compile to not
run
> on Linux and send the programmer back for fixups.
>
> Ok, by now you may think I am nuts ... and there is truth to that thought.
> Just to keep the thought on firm footing, I have provided a sample .d
file,
> the makefile template (mevmake.mk) and the underlying include make file
that
> has all the good stuff (basemake.mk).  Some more notes in progress are
> located at www.matthewv.org/make.html
>
> Thanks,
> Matthew
>
> -----Original Message-----
> From: Rob Tulloh [mailto:address@hidden
> Sent: Monday, September 10, 2001 1:02 PM
> To: Matthew Von-Maszewski
> Cc: address@hidden
> Subject: Re: FW: Win32 File Name Compares and Normalization
>
> >From the README.W32 that is my distribution:
>
> > Pathnames and Case insensitivity:
> >
> >     Unlike Unix, Windows 95/NT systems are case insensitive but case
> >     preserving.  For example if you tell the file system to create a
> >     file named "Target", it will preserve the case.  Subsequent access
to
> >     the file with other case permutations will succeed (i.e. opening a
> >     file named "target" or "TARGET" will open the file "Target").
> >
> >     By default, GNU make retains its case sensitivity when comparing
> target
> >     names and existing files or directories.  It can be configured,
> however,
> >     into a case preserving and case insensitive mode by adding a define
> >     for HAVE_CASE_INSENSITIVE_FS to config.h.W32.
> >
> >     For example, the following makefile will create a file named Target
> >     in the directory subdir which will subsequently be used to satisfy
> >     the dependency of SUBDIR/DepTarget on SubDir/TARGET.  Without
> >     HAVE_CASE_INSENSITIVE_FS configured, the dependency link will not be
> made:
> >
> >     subdir/Target:
> >         touch $@
> >
> >     SUBDIR/DepTarget: SubDir/TARGET
> >         cp $^ $@
> >
> >     Reliance on this behavior also eliminates the ability of gmake to
use
> >     case in comparison of matching rules.  For example, it is not
possible
> >     to set up a C++ rule using %.C : that is different than a C rule
using
> >     %.c :.  GNU make will consider these to be the same rule and will
> issue
> >     a warning.
> >
>
> I hope this is still in the README.W32 file we ship with GNU make!
>
> Rob
>
> Matthew Von-Maszewski wrote:
>
> > ... argh!  last line of the config.h, missed it.  I will recompile and
> > respond this afternoon ...
> >
> > -----Original Message-----
> > From: Rob Tulloh [mailto:address@hidden
> > Sent: Monday, September 10, 2001 12:37 PM
> > To: Matthew Von-Maszewski
> > Cc: address@hidden
> > Subject: Re: FW: Win32 File Name Compares and Normalization
> >
> > I thought GNU make already had a compile-time feature to
> > enable case insensitivity. Did you try this and find it didn't work
> > for you?
> >
> > There are users on both sides of the fence so any
> > work to enable this should be put under a feature test macro
> > that can be turned on or off.
> >
> > My $.02,
> >
> > Rob
> >
> > Matthew Von-Maszewski wrote:
> >
> > > Per Paul Smith, I am forwarding this email to this group for comments
> ...
> > >
> > > -----Original Message-----
> > > From: Matthew Von-Maszewski [mailto:address@hidden
> > > Sent: Monday, September 10, 2001 12:05 PM
> > > To: address@hidden
> > > Subject: Win32 File Name Compares and Normalization
> > >
> > > Greetings,
> > >
> > > I have been doing extensive work with GNU make and Microsoft's
compiler.
> > > This includes creating a dependency generator similar the gcc's .d
> files.
> > > It has become obvious that a couple of the make's functions need to
> become
> > > case insensitive when on a Win 32 platform.  Microsoft's file names
are
> > > internally considered to be case insensitive for its file operations
and
> > > matching.
> > >
> > > I have listed the changes below that I found sufficient to create
> > normalized
> > > filenames (all lowercase) and to remove case sensitive compares (this
> > needs
> > > a cleaner implementation).
> > >
> > > [My dependency generator depends upon a preprocessor dump from
> Microsoft's
> > > compiler.  Depending upon the command line options, sometimes the
> > > preprocessor normalizes all file names to lower case (e.g. /ZI
option).
> > > Other times case is left as found in the file system.  Lovely.]
> > >
> > > All changes are against the version 3.79.1 download.
> > >
> > > Would you please consider this change request?
> > >
> > > Thank you,
> > > Matthew
> > >
> > > Index: glob/fnmatch.c
> > > ===================================================================
> > > RCS file: /home/cvsroot/gnumake/glob/fnmatch.c,v
> > > retrieving revision 1.1.1.1
> > > retrieving revision 1.2
> > > diff -c -r1.1.1.1 -r1.2
> > > *** glob/fnmatch.c      2001/09/10 15:33:48     1.1.1.1
> > > --- glob/fnmatch.c      2001/09/10 15:37:35     1.2
> > > ***************
> > > *** 168,174 ****
> > >   # ifdef _LIBC
> > >   #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
> > >   # else
> > > ! #  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower
(c)
> :
> > > (c))
> > >   # endif
> > >
> > >     while ((c = *p++) != '\0')
> > > --- 168,176 ----
> > >   # ifdef _LIBC
> > >   #  define FOLD(c) ((flags & FNM_CASEFOLD) ? tolower (c) : (c))
> > >   # else
> > > ! /*#  define FOLD(c) ((flags & FNM_CASEFOLD) && ISUPPER (c) ? tolower
> (c)
> > :
> > > (c)
> > > )*/
> > > ! /* mev */
> > > ! #  define FOLD(c) ( ISUPPER (c) ? tolower (c) : (c))
> > >   # endif
> > >
> > > Index: w32/pathstuff.c
> > > ===================================================================
> > > RCS file: /home/cvsroot/gnumake/w32/pathstuff.c,v
> > > retrieving revision 1.1.1.1
> > > retrieving revision 1.2
> > > diff -c -r1.1.1.1 -r1.2
> > > *** w32/pathstuff.c     2001/09/10 15:33:48     1.1.1.1
> > > --- w32/pathstuff.c     2001/09/10 15:37:36     1.2
> > > ***************
> > > *** 80,85 ****
> > > --- 80,87 ----
> > >           if (*p == '\\')
> > >               *p = '/';
> > >
> > > + /* mev */
> > > +     strlwr(w32_path);
> > >       return w32_path;
> > >   }
> > >
> > > cvs server: Diffing w32/compat
> > > Index: w32/compat/dirent.c
> > > ===================================================================
> > > RCS file: /home/cvsroot/gnumake/w32/compat/dirent.c,v
> > > retrieving revision 1.1.1.1
> > > retrieving revision 1.2
> > > diff -c -r1.1.1.1 -r1.2
> > > *** w32/compat/dirent.c 2001/09/10 15:33:48     1.1.1.1
> > > --- w32/compat/dirent.c 2001/09/10 15:37:36     1.2
> > > ***************
> > > *** 116,122 ****
> > >         /* fill in struct dirent values */
> > >         pDir->dir_sdReturn.d_ino = -1;
> > >         strcpy(pDir->dir_sdReturn.d_name, wfdFindData.cFileName);
> > > !
> > >         return &pDir->dir_sdReturn;
> > >   }
> > >
> > > --- 116,123 ----
> > >         /* fill in struct dirent values */
> > >         pDir->dir_sdReturn.d_ino = -1;
> > >         strcpy(pDir->dir_sdReturn.d_name, wfdFindData.cFileName);
> > > ! /* mev */
> > > !         strlwr(pDir->dir_sdReturn.d_name);
> > >         return &pDir->dir_sdReturn;
> > >   }
> > >
> > > ======================
> > > Matthew Von-Maszewski
> > >  address@hidden
> > >  508-870-0118
> > >
> > > _______________________________________________
> > > Make-w32 mailing list
> > > address@hidden
> > > http://mail.gnu.org/mailman/listinfo/make-w32
>
>   ------------------------------------------------------------------------
>                   Name: basemake.mk
>    basemake.mk    Type: unspecified type (application/octet-stream)
>               Encoding: quoted-printable
>
>                  Name: mevmake.mk
>    mevmake.mk    Type: unspecified type (application/octet-stream)
>              Encoding: quoted-printable
>
>               Name: xmldb.d
>    xmldb.d    Type: unspecified type (application/octet-stream)
>           Encoding: quoted-printable