tag:old.nabble.com,2006:forum-1736Nabble - Gnu - Make2009-12-12T01:09:15ZGnu Make is a tool which controls the generation of executables and other non-source files of a program from the program's source files. Gnu - Make home is <a href="http://www.gnu.org/software/make/" target="_top" rel="nofollow">here</a>.tag:old.nabble.com,2006:post-26756015[bug #28126] bug with Windows interface: echo. in command only works when redirected2009-12-12T01:09:15Z2009-12-12T01:09:15ZEli Zaretskii-4<br>Follow-up Comment #3, bug #28126 (project make):
<br><br>I omitted those commands on purpose: they are executable programs on some
<br>older Windows versions, not shell built-ins.
<br><br>But if you, or someone else, can present a real-life use-case where it is
<br>good to have them as built-ins, I will consider changing my mind.
<br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28126" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28126</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26756015&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26754997[bug #28126] bug with Windows interface: echo. in command only works when redirected2009-12-11T20:49:24Z2009-12-11T20:49:24ZEli Zaretskii-4<br>Follow-up Comment #2, bug #28126 (project make):
<br><br>[Submitted by J. David Bryan, jdbryan at acm dot org]
<br><br>Would you consider adding the "move," "popd," "pushd," and "start" built-in
<br>commands as well?
<br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28126" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28126</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26754997&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26753469Re: was: RE: auto-dep cannot possibly work?2009-12-11T16:06:42Z2009-12-11T16:06:42ZDavid BoyceOn Fri, Dec 11, 2009 at 4:22 PM, Mark Galeck (CW) <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26753469&i=0" target="_top" rel="nofollow">mgaleck@...</a>> wrote:
<br>> Hello, this is about a thread that was here a while ago, where I argued that some "bizarre" cases should be taken care of by auto-dependency generation.
<br>>
<br>> I want to say, that after chasing this wild goose for a while, I came to understand that it is not feasible in principle to cover the case of new header files appearing, without changing existing sources.
<br>>
<br>> What is more, I finally came to understand that Paul was right, of course, that it is up to the developer to edit the makefile, do "make clean" or whatever needs to be done, if they introduce new files without touching existing sources. Make should not be expected to take care of that.
<br><br>> Thank you Paul, it is a pleasure to learn from you, even though I am a little slow with it, please excuse that.
<br><br>A classy touch.
<br><br>You know, there was something I was going to say back when this thread
<br>was active but everyone seemed a little touchy and I didn't want to
<br>make it worse ... but now that's over and we're all friends I think I
<br>will.
<br><br>I wish both you and Paul would 'come to understand' that I was right
<br>way back in the early part of the thread, that it is absolutely
<br>"feasible in principle to cover the case of new header files appearing
<br>..." by using the directories themselves as prerequisites. By this I
<br>mean that it's an objectively correct solution. The objection raised,
<br>as I recall, was that people may make any kind of unrelated changes in
<br>include dirs (temp files, backup files) which will trigger
<br>"unnecessary" builds. To which I would say then don't do that - you
<br>must design your structure and process such that you're not searching
<br>directories in which people work all the time. This is not so hard -
<br>for instance you can make a "./include" directory and populate it with
<br>symlinks to the header files where they actually live. Since this dir
<br>exists only to be searched at build time, and nothing really lives
<br>there, it won't get visited much.
<br><br>A related objection was that even if the change really does reflect a
<br>new header file, it probably won't eclipse an existing one and/or
<br>won't change the resulting binaries. This again is true but the same
<br>can be said about files too. In fact, if you think it out carefully
<br>the analogy between these directories and plain files is almost
<br>perfect. When I modify an existing header file, more often than not
<br>I'm simply adding a new macro or declaration to it. Or I might be
<br>adding comments or updating the copyright to 2010 (reminder to all).
<br>In all these cases make takes an appropriately pessimistic approach -
<br>since it has no way to know the semantic content of the change, it
<br>rebuilds everything that could possibly be affected. And if someone
<br>complained about the changes causing "unneeded" rebuilds, you'd say
<br>either don't make the changes or get used to the rebuilds. When you
<br>use dependencies on directories in search paths make behaves in
<br>precisely the same way - in both cases, in some sense you're getting
<br>unnecessary rebuilds but it's the optimal set given what we know a
<br>priori.
<br><br>Bottom line, there is a solution to this problem which is both elegant
<br>and 100% correct. It just may not be the answer you want it to be.
<br>Anyway, that's my contention and I'd be interested (and surprised) to
<br>be shown wrong now that the other shouting has died down.
<br><br>BTW, I don't remember if it came up before but it sounds like ccache
<br>ought to be mentioned in this thread because it can make
<br>determinations based on semantic content and thus partners well with
<br>make. You might want to look into it if you haven't before.
<br><br>David Boyce
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26753469&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26751627was: RE: auto-dep cannot possibly work?2009-12-11T13:22:37Z2009-12-11T13:22:37ZMark Galeck (CW)Hello, this is about a thread that was here a while ago, where I argued that some "bizarre" cases should be taken care of by auto-dependency generation.
<br><br>I want to say, that after chasing this wild goose for a while, I came to understand that it is not feasible in principle to cover the case of new header files appearing, without changing existing sources.
<br><br>What is more, I finally came to understand that Paul was right, of course, that it is up to the developer to edit the makefile, do "make clean" or whatever needs to be done, if they introduce new files without touching existing sources. Make should not be expected to take care of that.
<br><br>Thank you Paul, it is a pleasure to learn from you, even though I am a little slow with it, please excuse that.
<br><br>Mark
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26751627&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26746826Re: bug with Windows interface: echo. in command only works when redirected2009-12-11T08:00:56Z2009-12-11T08:00:56ZEli Zaretskii<div class='shrinkable-quote'>> From: "Mark Galeck (CW)" <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26746826&i=0" target="_top" rel="nofollow">mgaleck@...</a>>
<br>> Date: Sun, 29 Nov 2009 20:31:51 -0800
<br>> Accept-Language: en-US
<br>> acceptlanguage: en-US
<br>>
<br>> Hello, I asked for help on this issue (with GNU make on Windows) on the help-make list, and Paul Smith says, this is a bug. I am using WinXP Pro, SP3. Here is an SSCCE:
<br>>
<br>> makefile:
<br>> -------------------------
<br>> SHELL=cmd.exe
<br>> foobar:
<br>> @echo.>foo
<br>> ------------------------
<br>>
<br>> gives me:
<br>> C:\tmp>make foobar
<br>>
<br>> C:\tmp>
<br>>
<br>>
<br>> However, this one:
<br>>
<br>> makefile:
<br>> -------------------------
<br>> SHELL=cmd.exe
<br>> foobar:
<br>> @echo.
<br>> ------------------------
<br>>
<br>> gives:
<br>>
<br>> C:\tmp>make foobar
<br>> process_begin: CreateProcess(NULL, echo., ...) failed.
<br>> make (e=2): The system cannot find the file specified.
<br>> make: *** [foobar] Error 2
</div><br>Thanks, I just fixed this in CVS. Patch below.
<br><br>2009-12-11 Eli Zaretskii <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26746826&i=1" target="_top" rel="nofollow">eliz@...</a>>
<br><br> * job.c (construct_command_argv_internal) <sh_cmds_dos>
<br> [WINDOWS32]: Add "echo." and a few more commands that are built
<br> into cmd.exe. Fixes Savannah bug #28126.
<br><br>--- job.c~ 2008-09-30 13:15:03.465375000 +0200
<br>+++ job.c 2009-12-11 17:22:22.698250000 +0200
<br>@@ -2287,12 +2287,15 @@
<br> 0 };
<br> #elif defined (WINDOWS32)
<br> static char sh_chars_dos[] = "\"|&<>";
<br>- static char *sh_cmds_dos[] = { "break", "call", "cd", "chcp", "chdir", "cls",
<br>- "copy", "ctty", "date", "del", "dir", "echo",
<br>- "erase", "exit", "for", "goto", "if", "if", "md",
<br>- "mkdir", "path", "pause", "prompt", "rd", "rem",
<br>- "ren", "rename", "rmdir", "set", "shift", "time",
<br>- "type", "ver", "verify", "vol", ":", 0 };
<br>+ static char *sh_cmds_dos[] = { "assoc", "break", "call", "cd", "chcp",
<br>+ "chdir", "cls", "color", "copy", "ctty",
<br>+ "date", "del", "dir", "echo", "echo.",
<br>+ "endlocal", "erase", "exit", "for", "ftype",
<br>+ "goto", "if", "if", "md", "mkdir", "path",
<br>+ "pause", "prompt", "rd", "rem", "ren",
<br>+ "rename", "rmdir", "set", "setlocal",
<br>+ "shift", "time", "title", "type", "ver",
<br>+ "verify", "vol", ":", 0 };
<br> static char sh_chars_sh[] = "#;\"*?[]&|<>(){}$`^";
<br> static char *sh_cmds_sh[] = { "cd", "eval", "exec", "exit", "login",
<br> "logout", "set", "umask", "wait", "while", "for",
<br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26746826&i=2" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26746972[bug #28126] bug with Windows interface: echo. in command only works when redirected2009-12-11T07:59:10Z2009-12-11T07:59:10ZEli Zaretskii-4<br>Update of bug #28126 (project make):
<br><br> Status: None => Fixed
<br> Open/Closed: Open => Closed
<br> Fixed Release: None => CVS
<br><br> _______________________________________________________
<br><br>Follow-up Comment #1:
<br><br>Fixed with the attached patch (which adds a couple more builtin commands).
<br><br><br>(file #19239)
<br> _______________________________________________________
<br><br>Additional Item Attachment:
<br><br>File name: w32builtins.dif Size:1 KB
<br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28126" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28126</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26746972&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26743025Re: behavior on Win depends on whether \ or / is used??2009-12-11T03:58:41Z2009-12-11T03:58:41ZEli Zaretskii<div class='shrinkable-quote'>> From: "Mark Galeck (CW)" <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743025&i=0" target="_top" rel="nofollow">mgaleck@...</a>>
<br>> Date: Thu, 10 Dec 2009 13:06:14 -0800
<br>>
<br>> I have 2 subdirectories foobar0 and foobar1, and only in foobar1 there is a=
<br>> file foobar.s . There is no file foobar.o .
<br>>
<br>> The makefile is:
<br>>
<br>> SHELL=Dcmd.exe
<br>>
<br>> vpath %.s .\foobar0 .\foobar1
<br>>
<br>> %.o : %.s
<br>> echo $<
<br>>
<br>> foobar0\foobar.s:;
<br>>
<br>>
<br>> This behaves as expected:
<br>>
<br>> C:\tmp>make foobar.o
<br>> echo .\foobar1/foobar.s
<br>>
<br>> However, if I change all the backslashes to forward-slashes, then I get:
<br>>
<br>> C:\tmp>make foobar.o
<br>> make: *** No rule to make target `foobar.s', needed by `foobar.o'. Stop.
</div><br>The behavior with the forward slashes is the "correct" one (see below
<br>for the reasons I put ``correct'' in quotes). Backslashes are not
<br>treated by Make on Windows in a 100% consistent manner, because in
<br>some cases they have special semantics.
<br><br>One such special case is in file-name wildcards, where a backslash
<br>quotes special wildcard characters.
<br><br>Another is in searching for files along vpath: Make looks up its
<br>database of known files by literally comparing hash values of the file
<br>names as strings, so foobar1/foobar.s and foobar1\foobar.s are not
<br>considered equal, and the search fails. This is your case. If you
<br>replace the backslash with forward slashes only in the last line of
<br>your Makefile, it will work (after applying the patch below). That is
<br>because, when Make searches vpath, it looks for .\foobar0/foobar.s and
<br>.\foobar1/foobar.s, but your Makefile specifies foobar0\foobar.s, so
<br>Make thinks the file foobar0\foobar.s is not mentioned in the
<br>Makefile. If you use forward slashes everywhere, then Make realizes
<br>that file foobar0/foobar.s _is_ mentioned in the Makefile.
<br><br>I found and fixed a small bug in how the leading ".\" was treated on
<br>Windows: it should be treated the same as "./", but it wasn't. The
<br>patch below fixes that. However, this does not (and cannot) make the
<br>backslash version of your Makefile behave 100% like the version with
<br>forward slashes, due to the fact that targets of the Makefile are
<br>looked up by comparing hash values of their strings, and the different
<br>flavor of slashes defeats the comparison.
<br><br>Paul, do you see any potential problems in providing on Windows a
<br>special version of the hash macros that would treat forward and
<br>back-slashes the same?
<br><br>The above explains why the behavior with backslashes was different.
<br>As to why the behavior with forward slashes is correct: this seems to
<br>be because you have the foobar0/foobar.s target in the Makefile, but
<br>not foobar1/foobar.s, while in reality the file that exists is
<br>foobar1/foobar.s. This seems to dupe Make into thinking that it found
<br>the right prerequisite as foobar0/foobar.s, and it does not proceed
<br>looking for foobar1/foobar.s. If you change the order of directories
<br>in the vpath line, the commands will work as you expect.
<br><br>There is a long comment in the function vpath.c:selective_vpath_search
<br>which seems to describe why files mentioned as targets in the Makefile
<br>are treated specially, but I don't understand what it says. Perhaps
<br>Paul or someone else could help in explaining the current behavior.
<br><br>(I added help-make to the list of addressees, because this question
<br>was first asked there.)
<br><br>Here's the patch mentioned above that I will install in the
<br>development sources:
<br><br>2009-12-11 Eli Zaretskii <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743025&i=1" target="_top" rel="nofollow">eliz@...</a>>
<br><br> * file.c (lookup_file) [HAVE_DOS_PATHS]: Treat '\\' like we do
<br> with '/'.
<br><br>--- file.c~ 2006-03-17 18:24:20.000000000 +0200
<br>+++ file.c 2009-12-11 13:12:13.760750000 +0200
<br>@@ -102,10 +102,20 @@
<br> while (name[0] == '[' && name[1] == ']' && name[2] != '\0')
<br> name += 2;
<br> #endif
<br>- while (name[0] == '.' && name[1] == '/' && name[2] != '\0')
<br>+ while (name[0] == '.'
<br>+#ifdef HAVE_DOS_PATHS
<br>+ && (name[1] == '/' || name[1] == '\\')
<br>+#else
<br>+ && name[1] == '/'
<br>+#endif
<br>+ && name[2] != '\0')
<br> {
<br> name += 2;
<br>- while (*name == '/')
<br>+ while (*name == '/'
<br>+#ifdef HAVE_DOS_PATHS
<br>+ || *name == '\\'
<br>+#endif
<br>+ )
<br> /* Skip following slashes: ".//foo" is "foo", not "/foo". */
<br> ++name;
<br> }
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743025&i=2" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26743035Re: behavior on Win depends on whether \ or / is used??2009-12-11T03:58:41Z2009-12-11T03:58:41ZEli Zaretskii<div class='shrinkable-quote'>> From: "Mark Galeck (CW)" <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743035&i=0" target="_top" rel="nofollow">mgaleck@...</a>>
<br>> Date: Thu, 10 Dec 2009 13:06:14 -0800
<br>>
<br>> I have 2 subdirectories foobar0 and foobar1, and only in foobar1 there is a=
<br>> file foobar.s . There is no file foobar.o .
<br>>
<br>> The makefile is:
<br>>
<br>> SHELL=Dcmd.exe
<br>>
<br>> vpath %.s .\foobar0 .\foobar1
<br>>
<br>> %.o : %.s
<br>> echo $<
<br>>
<br>> foobar0\foobar.s:;
<br>>
<br>>
<br>> This behaves as expected:
<br>>
<br>> C:\tmp>make foobar.o
<br>> echo .\foobar1/foobar.s
<br>>
<br>> However, if I change all the backslashes to forward-slashes, then I get:
<br>>
<br>> C:\tmp>make foobar.o
<br>> make: *** No rule to make target `foobar.s', needed by `foobar.o'. Stop.
</div><br>The behavior with the forward slashes is the "correct" one (see below
<br>for the reasons I put ``correct'' in quotes). Backslashes are not
<br>treated by Make on Windows in a 100% consistent manner, because in
<br>some cases they have special semantics.
<br><br>One such special case is in file-name wildcards, where a backslash
<br>quotes special wildcard characters.
<br><br>Another is in searching for files along vpath: Make looks up its
<br>database of known files by literally comparing hash values of the file
<br>names as strings, so foobar1/foobar.s and foobar1\foobar.s are not
<br>considered equal, and the search fails. This is your case. If you
<br>replace the backslash with forward slashes only in the last line of
<br>your Makefile, it will work (after applying the patch below). That is
<br>because, when Make searches vpath, it looks for .\foobar0/foobar.s and
<br>.\foobar1/foobar.s, but your Makefile specifies foobar0\foobar.s, so
<br>Make thinks the file foobar0\foobar.s is not mentioned in the
<br>Makefile. If you use forward slashes everywhere, then Make realizes
<br>that file foobar0/foobar.s _is_ mentioned in the Makefile.
<br><br>I found and fixed a small bug in how the leading ".\" was treated on
<br>Windows: it should be treated the same as "./", but it wasn't. The
<br>patch below fixes that. However, this does not (and cannot) make the
<br>backslash version of your Makefile behave 100% like the version with
<br>forward slashes, due to the fact that targets of the Makefile are
<br>looked up by comparing hash values of their strings, and the different
<br>flavor of slashes defeats the comparison.
<br><br>Paul, do you see any potential problems in providing on Windows a
<br>special version of the hash macros that would treat forward and
<br>back-slashes the same?
<br><br>The above explains why the behavior with backslashes was different.
<br>As to why the behavior with forward slashes is correct: this seems to
<br>be because you have the foobar0/foobar.s target in the Makefile, but
<br>not foobar1/foobar.s, while in reality the file that exists is
<br>foobar1/foobar.s. This seems to dupe Make into thinking that it found
<br>the right prerequisite as foobar0/foobar.s, and it does not proceed
<br>looking for foobar1/foobar.s. If you change the order of directories
<br>in the vpath line, the commands will work as you expect.
<br><br>There is a long comment in the function vpath.c:selective_vpath_search
<br>which seems to describe why files mentioned as targets in the Makefile
<br>are treated specially, but I don't understand what it says. Perhaps
<br>Paul or someone else could help in explaining the current behavior.
<br><br>(I added help-make to the list of addressees, because this question
<br>was first asked there.)
<br><br>Here's the patch mentioned above that I will install in the
<br>development sources:
<br><br>2009-12-11 Eli Zaretskii <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743035&i=1" target="_top" rel="nofollow">eliz@...</a>>
<br><br> * file.c (lookup_file) [HAVE_DOS_PATHS]: Treat '\\' like we do
<br> with '/'.
<br><br>--- file.c~ 2006-03-17 18:24:20.000000000 +0200
<br>+++ file.c 2009-12-11 13:12:13.760750000 +0200
<br>@@ -102,10 +102,20 @@
<br> while (name[0] == '[' && name[1] == ']' && name[2] != '\0')
<br> name += 2;
<br> #endif
<br>- while (name[0] == '.' && name[1] == '/' && name[2] != '\0')
<br>+ while (name[0] == '.'
<br>+#ifdef HAVE_DOS_PATHS
<br>+ && (name[1] == '/' || name[1] == '\\')
<br>+#else
<br>+ && name[1] == '/'
<br>+#endif
<br>+ && name[2] != '\0')
<br> {
<br> name += 2;
<br>- while (*name == '/')
<br>+ while (*name == '/'
<br>+#ifdef HAVE_DOS_PATHS
<br>+ || *name == '\\'
<br>+#endif
<br>+ )
<br> /* Skip following slashes: ".//foo" is "foo", not "/foo". */
<br> ++name;
<br> }
<br><br><br>_______________________________________________
<br>Make-w32 mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26743035&i=2" target="_top" rel="nofollow">Make-w32@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/make-w32" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/make-w32</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---W32-f1740.html" embed="fixTarget[1740]" target="_top" >Gnu - Make - W32</a></p>tag:old.nabble.com,2006:post-26740319behavior on Win depends on whether \ or / is used??2009-12-10T13:06:14Z2009-12-10T13:06:14ZMark Galeck (CW)<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:m="http://schemas.microsoft.com/office/2004/12/omml" xmlns="http://www.w3.org/TR/REC-html40">
<head>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=us-ascii">
<meta name=Generator content="Microsoft Word 12 (filtered medium)">
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=EN-US link=blue vlink=purple>
<div class=Section1>
<p class=MsoPlainText>Hello, this is a Windows question, <o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>I have 2 subdirectories foobar0 and foobar1, and only in
foobar1 there is a file foobar.s . There is no file foobar.o . <o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>The makefile is:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>SHELL=cmd.exe<o:p></o:p></p>
<p class=MsoPlainText> <o:p></o:p></p>
<p class=MsoPlainText>vpath %.s .\foobar0 .\foobar1 <o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>%.o : %.s <o:p></o:p></p>
<p class=MsoPlainText> echo $<<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>foobar0\foobar.s:;<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>This behaves as expected:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>C:\tmp>make foobar.o<o:p></o:p></p>
<p class=MsoPlainText>echo .\foobar1/foobar.s<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>However, if I change all the backslashes to
forward-slashes, then I get:<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>C:\tmp>make foobar.o<o:p></o:p></p>
<p class=MsoPlainText>make: *** No rule to make target `foobar.s', needed by
`foobar.o'. Stop.<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>(this is just a small example, what I observe is that
forward slash in directories, misbehaves with vpath on Windows)<o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Is this a bug in Windows make, or in my head? <o:p></o:p></p>
<p class=MsoPlainText><o:p> </o:p></p>
<p class=MsoPlainText>Mark<o:p></o:p></p>
<p class=MsoNormal><o:p> </o:p></p>
</div>
</body>
</html>
<br />_______________________________________________
<br>Make-w32 mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26740319&i=0" target="_top" rel="nofollow">Make-w32@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/make-w32" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/make-w32</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---W32-f1740.html" embed="fixTarget[1740]" target="_top" >Gnu - Make - W32</a></p>tag:old.nabble.com,2006:post-26731506Re: Dependency graphs for a Makefile2009-12-10T09:37:09Z2009-12-10T09:37:09ZPaul Smith-20On Fri, 2009-12-11 at 11:18 +1800, Peng Yu wrote:
<br>> GraphViz can be used to plot a graph. I'm wondering if there is a easy
<br>> way to generate dependency graph (exported to a figure that can be
<br>> visualized) for a Makefile.
<br><br>There is nothing built into make. You can use "make -p" to see a copy
<br>of the database which will give you any explicit dependencies and could
<br>be postprocessed with perl/awk/etc.
<br><br>However, there is no way to know what the implicit dependencies are
<br>since make doesn't compute them ever unless it actually wants to build
<br>the target.
<br><br>It *might* be possible to do this after a complete build if you give the
<br>-B option, as well as -d, but the parsing will be a lot more complicated
<br>since you'll have to watch the debug output to see all the prerequisites
<br>that make considers when building a target.
<br><br><br>There is also the remake project which you can find on sourceforge,
<br>which might have some help for you on this:
<br><br><a href="http://bashdb.sourceforge.net/remake/" target="_top" rel="nofollow">http://bashdb.sourceforge.net/remake/</a><br><br>The "download" link here seems to lead me off into the weeds, but if you
<br>go to the sourceforge.net project page link and use the "View all files"
<br>button you can find it.
<br><br>I can't say for sure but I think the debugging output of remake is more
<br>conducive to parsing, at any rate.
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26731506&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26731292Dependency graphs for a Makefile2009-12-10T09:18:59Z2009-12-10T09:18:59ZPeng YuGraphViz can be used to plot a graph. I'm wondering if there is a easy
<br>way to generate dependency graph (exported to a figure that can be
<br>visualized) for a Makefile.
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26731292&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26721316RE: behavior on Win depends on whether \ or / is used??2009-12-09T18:42:04Z2009-12-09T18:42:04ZMark Galeck (CW)Here's another example, now this one is just terrible, I spent several hours tracking down this problem:
<br><br><br>Again, 2 subdirectories foobar0 and foobar1, and only in foobar1 there is a file foobar.s . There is no file foobar.o .
<br><br>The makefile is:
<br><br>SHELL=cmd.exe
<br>
<br>vpath %.s .\foobar0 .\foobar1
<br><br>%.o : %.s
<br> echo $<
<br><br>.\foobar0\foobar.s:;
<br><br><br>This behaves as expected:
<br><br>C:\tmp>make foobar.o
<br>echo .\foobar1/foobar.s
<br><br><br>But, that is after my change! The last line in the makefile is generated automatically - it comes out of what make outputs as $< to some rule, and make, outputs forward slashes! If I keep it the way make wants it
<br><br>.\foobar/foobar.s:;
<br><br><br>Now this bombs! Because it says
<br><br>C:\tmp>make foobar.o
<br>echo .\foobar0/foobar.s
<br><br>and of course, the file foobar0/foobar.s does not exist - and then later make dutifully exclaims "cannot open file foobar0/foobar.s for reading" - well yes you can't, why did you think you found it???
<br><br>What is going on here, can someone enlighten me?
<br><br>How do I get vpaths and forward slashes, output by make itself, to behave on Windows? It seems to me that the only solution, in those more complex cases like above, is every time make outputs a forward slash, to edit it myself to backslash, before I give it back to make!!
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26721316&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26720921behavior on Win depends on whether \ or / is used??2009-12-09T17:45:46Z2009-12-09T17:45:46ZMark Galeck (CW)Hello, this is a Windows question,
<br><br>I have 2 subdirectories foobar0 and foobar1, and only in foobar1 there is a file foobar.s . There is no file foobar.o .
<br><br>The makefile is:
<br><br><br>SHELL=cmd.exe
<br>
<br>vpath %.s .\foobar0 .\foobar1
<br><br>%.o : %.s
<br> echo $<
<br><br>foobar0\foobar.s:;
<br><br><br>This behaves as expected:
<br><br>C:\tmp>make foobar.o
<br>echo .\foobar1/foobar.s
<br><br><br>However, if I change all the backslashes to forward-slashes, then I get:
<br><br><br>C:\tmp>make foobar.o
<br>make: *** No rule to make target `foobar.s', needed by `foobar.o'. Stop.
<br><br>(this is just a small example, what I observe is that forward slash in directories, misbehaves with vpath on Windows)
<br><br>Is this a bug in Windows make, or in my head?
<br><br>Mark
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26720921&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26710498Target with fake dependencies2009-12-09T05:43:40Z2009-12-09T05:43:40ZBeber-2Hi,
<br><br>I'm trying to wrote a generic Makefile that will have to do what I ask
<br>in all SUBDIRS (declared inside Makefile). For exemple :
<br><br>$ make pdf odt
<br>make -C ni_a_gomz_brezhoneg pdf odt
<br>...
<br>make -C troit_e_brezhoneg pdf odt
<br><br>$ make pdf dummy-target
<br>make -C ni_a_gomz_brezhoneg pdf dummy-target
<br>make[1]: Entering directory `/home/beber/bzh/skol_ober/ni_a_gomz_brezhoneg'
<br>make: *** No rule to make target `dummy-target'. Stop.
<br><br>I actually did this like that :
<br><br>---
<br>SUBDIRS = \
<br> ni_a_gomz_brezhoneg \
<br> troit_e_brezhoneg
<br><br>all: $(SUBDIRS)
<br><br>%:
<br> @for DIR in $(SUBDIRS) ; do \
<br> echo $(MAKE) -C $$DIR $@ ; \
<br> $(MAKE) -C $$DIR $@ ; \
<br> done
<br><br>.PHONY: $(SUBDIRS)
<br>---
<br><br>This is ok when I use :
<br> - make all
<br> - make clean
<br> - make dummy-target
<br><br>Not important : I can't get it make go in the subdir with all previous target as
<br>argument. I get the following scheme :
<br><br>$ make pdf odt
<br>make -C ni_a_gomz_brezhoneg pdf
<br>..
<br>make -C troit_e_brezhoneg pdf
<br>..
<br>make -C ni_a_gomz_brezhoneg odt
<br>..
<br>make -C troit_e_brezhoneg odt
<br>..
<br><br>This is also fine.
<br><br>But, I can't just build one of the subdir, if I ask just
<br>ni_a_gomz_brezhoneg target for exemple :
<br><br>$ make ni_a_gomz_brezhoneg
<br>make: Nothing to be done for `ni_a_gomz_brezhoneg'.
<br><br>Also I can override this :
<br>$ make -C ni_a_gomz_brezhoneg
<br>make: Entering directory `/home/beber/bzh/skol_ober/ni_a_gomz_brezhoneg'
<br><br>This is not a worry, but i'd like to understand why there is nothing to
<br>do as it's .PHONY.
<br><br>Another thing is that I'd like to avoid the "shell for loop". foreach is
<br>not really appropriate as it make things less visibles by executing at
<br>once.
<br><br>I did this like that in past for file like :
<br><br>---
<br>SRC = \
<br> br.po
<br><br>OBJMO = $(SRC:po=mo)
<br><br># Build
<br>%.mo %.gmo: %.po
<br> $(MSGFMTPRG) $(MSGFMTOPT) -o $@ $<
<br><br># Install
<br>install: $(OBJMO)
<br> $(MAKE) $(addprefix install-, $^)
<br><br>install-%.mo: $(OBJMO)
<br> $(INSTALLPRG) $(INSTALLOPT) $< $(localedir)/$*/LC_MESSAGES/$(NAME).mo
<br>---
<br><br>I don't known if it's a good practice, but it permit me to avoid "shell loop"
<br>by re-executing make and redefined target on the fly :
<br><br>$ make install prefix=/tmp
<br>make install-br.mo install-fr.mo
<br><br>I'm trying to achieve the same thing when I have to use $(MAKE) -C to
<br>go inside subdir as demonstrated previously.
<br><br>Thanks for your advice
<br>Beber
<br><br>And sorry for the bad english.
<br><br>--
<br>Beber
<br><br /> <br />_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26710498&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><div class="small"><br/><img src="http://old.nabble.com/images/icon_attachment.gif" > <strong>signature.asc</strong> (204 bytes) <a href="http://old.nabble.com/attachment/26710498/0/signature.asc" target="_top">Download Attachment</a></div><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26703765Re: Cancelling a Implicit Rule ?2009-12-08T17:28:02Z2009-12-08T17:28:02ZPhilip Guenther-22009/12/8 L.Guo <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26703765&i=0" target="_top" rel="nofollow">leaveye.guo@...</a>>:
<br>...
<div class='shrinkable-quote'><br>> I have a main makefile 'Makefile' and two sub (common) makefile 'default.mk' and 'rules.mk'.
<br>>
<br>> The source files all in 'src' dir, the output in 'out'. As these:
<br>>
<br>> SRCS = src/a.c src/main.c src/version.h
<br>> DEPS = out/a.d out/main.d
<br>> OBJS = out/a.o out/main.o
<br>> TGTS = out/programme
<br>>
<br>> I want to use pattern rules in 'rules.mk' to handle the dependencies generating.
<br>>
<br>> $(OUTDIR)/%.d: $(SRCDIR)/%.c $(OUTDIR)
<br>> @{ set -e; $(RM) $@;\
<br>> echo 'Generate $@ from $<'\
<br>> $(CC) -M $(CFLAGS) $< > $@.$$$$;\
<br>> sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@.$$$$ > $@;\
<br>> $(RM) $@.$$$$;\
<br>> }
<br>>
<br>> $(OBJS): $(OUTDIR)/%.o: $(SRCDIR)/%.c $(OUTDIR)/%.d
<br>> $(CC) -c $(CFLAGS) -o $@ $<
</div><br>I suggest you read
<br> <a href="http://make.paulandlesley.org/autodep.html#advanced" target="_top" rel="nofollow">http://make.paulandlesley.org/autodep.html#advanced</a><br><br>to see how having actual targets for the files included by make is
<br>unnecessary and wasteful. It's sufficient (and more efficient) to
<br>just have them created as a side-effect of the rule that generates the
<br>.o files.
<br><br><br>> When using the whole thing to build my project, 'make' gives me an error report:
<br>>
<br>> make: *** No rule to make target `main.c', needed by `out/main.d'. Stop.
<br><br>I don't think this can be determined from what you've provided. I
<br>suggest you check the output of
<br> make -Rpq
<br><br>and examine all the lines that include ".d" to see whether anything
<br>stands out as "Not What I Meant".
<br><br><br>Philip Guenther
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26703765&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26701615[bug #28230] $(eval) does not understand multiline2009-12-08T14:02:01Z2009-12-08T14:02:01ZEli Zaretskii-4<br>Follow-up Comment #1, bug #28230 (project make):
<br><br>To quote the info pages:
<br><br> The `shell' function performs the same function that backquotes
<br>(``') perform in most shells: it does "command expansion". This means that
<br>it takes as an argument a shell command and evaluates to the output of the
<br>command. The only processing `make' does on the result is to convert each
<br>newline (or carriage-return / newline pair) to a single space. If there is a
<br>trailing (carriage-return and) newline it will simply be removed.
<br><br>So, $(eval) isn't actually seeing two lines in this case, but just a single
<br>one. The described behavior is as specified by the documentation, so it seems
<br>like "not a bug", though maybe a feature request for a version of $(shell)
<br>that gives a multiline result is called for.
<br><br>Philip Guenther
<br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28230" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28230</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26701615&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26699137[bug #28230] $(eval) does not understand multiline2009-12-08T10:53:48Z2009-12-08T10:53:48ZEli Zaretskii-4<br>URL:
<br> <<a href="http://savannah.gnu.org/bugs/?28230" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28230</a>>
<br><br> Summary: $(eval) does not understand multiline
<br> Project: make
<br> Submitted by: yozh
<br> Submitted on: Tue 08 Dec 2009 06:53:47 PM GMT
<br> Severity: 3 - Normal
<br> Item Group: None
<br> Status: None
<br> Privacy: Public
<br> Assigned to: None
<br> Open/Closed: Open
<br> Discussion Lock: Any
<br> Component Version: None
<br> Operating System: None
<br> Fixed Release: None
<br> Triage Status: None
<br><br> _______________________________________________________
<br><br>Details:
<br><br>====
<br># cat script.sh
<br>#!/bin/sh -e
<br><br>echo "a = 1"
<br>echo "b = 2"
<br><br># cat Makefile
<br>$(eval $(shell ./script.sh))
<br>$(info $(a))
<br>$(info $(b))
<br><br>xx:
<br># gmake
<br>1 b = 2
<br><br>gmake: Nothing to be done for `xx'.
<br>====
<br><br>"1 b = 2\n\n" is printed. "1\n2\b" expected.
<br><br><br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28230" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28230</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26699137&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26698795Bug in Makefile document2009-12-08T07:21:02Z2009-12-08T07:21:02ZSanjiv K. Bhatia-2I just saw something in the Makefile document and thought you should know.
<br><br>On Page 14 of the manual (PDF version), in Section 3.5, the last
<br>paragraph of the section says:
<br><br>"See Section 8.2 [Text Functions], page 78, for more information on
<br>the word and words functions used above."
<br><br>It should read:
<br><br>"See Section 8.2 [Text Functions], page 78, for more information on
<br>the lastword function used above."
<br><br>The example is with "lastword" and not "word" or "words" which are
<br>also functions.
<br><br>Sanjiv
<br>--
<br>Sanjiv K. Bhatia
<br><a href="http://member.acm.org/~sanjiv" target="_top" rel="nofollow">http://member.acm.org/~sanjiv</a><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26698795&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26691521Cancelling a Implicit Rule ?2009-12-08T02:13:47Z2009-12-08T02:13:47ZL.GuoHi all:
<br><br>In my C project, I need to generate version.h and add it into compiling.
<br>And also, I need a 'copy' rule, which copies the output binaries into a dir specified by a environment var COPYDIR.
<br>This is a common need for my projects.
<br>So I design a makefile structure, try to implement it and encounter a problem as followed.
<br><br><br>I have a main makefile 'Makefile' and two sub (common) makefile 'default.mk' and 'rules.mk'.
<br><br>The source files all in 'src' dir, the output in 'out'. As these:
<br><br> SRCS = src/a.c src/main.c src/version.h
<br> DEPS = out/a.d out/main.d
<br> OBJS = out/a.o out/main.o
<br> TGTS = out/programme
<br><br><br>I want to use pattern rules in 'rules.mk' to handle the dependencies generating.
<br><br>$(OUTDIR)/%.d: $(SRCDIR)/%.c $(OUTDIR)
<br> @{ set -e; $(RM) $@;\
<br> echo 'Generate $@ from $<'\
<br> $(CC) -M $(CFLAGS) $< > $@.$$$$;\
<br> sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@.$$$$ > $@;\
<br> $(RM) $@.$$$$;\
<br> }
<br><br>$(OBJS): $(OUTDIR)/%.o: $(SRCDIR)/%.c $(OUTDIR)/%.d
<br> $(CC) -c $(CFLAGS) -o $@ $<
<br><br><br>'default.mk' is omited because it just complete some necessary-but-unset vars in 'Makefile'.
<br><br><br>When using the whole thing to build my project, 'make' gives me an error report:
<br><br> make: *** No rule to make target `main.c', needed by `out/main.d'. Stop.
<br><br><br>I know that, for 'rules.mk' including the '.d' files, 'make' need 'out/main.d', and also need 'src/main.c'.
<br>I guess it is using a implict rule to generate '.d' from '.c'. And tried this additional empty rule but failed:
<br><br> $(OUTDIR)/%.d: %.c
<br><br>I have no idea why 'main.c' is needed after 'make' using 'src/main.c'.
<br>Here is the output of 'make -d'. (I use -R to make the output shorter.)
<br><br>---------------------------------------------------------------
<br>Reading makefiles...
<br>Reading makefile `Makefile'...
<br>Reading makefile `default.mk' (search path) (no ~ expansion)...
<br>Reading makefile `rules.mk' (search path) (no ~ expansion)...
<br>Reading makefile `out/a.d' (search path) (no ~ expansion)...
<br>Reading makefile `out/main.d' (search path) (no ~ expansion)...
<br>Updating makefiles....
<br> Considering target file `out/main.d'.
<br> Looking for an implicit rule for `out/main.d'.
<br> Trying pattern rule with stem `main'.
<br> Trying implicit prerequisite `src/main.c'.
<br> Trying rule prerequisite `out'.
<br> Found an implicit rule for `out/main.d'.
<br> Considering target file `src/main.c'.
<br> Looking for an implicit rule for `src/main.c'.
<br> No implicit rule found for `src/main.c'.
<br> Finished prerequisites of target file `src/main.c'.
<br> No need to remake target `src/main.c'.
<br> Considering target file `out'.
<br> Finished prerequisites of target file `out'.
<br> No need to remake target `out'.
<br> Considering target file `main.c'.
<br> File `main.c' does not exist.
<br> Looking for an implicit rule for `main.c'.
<br> No implicit rule found for `main.c'.
<br> Finished prerequisites of target file `main.c'.
<br> Must remake target `main.c'.
<br>make: *** No rule to make target `main.c', needed by `out/main.d'. Stop.
<br>----------------------------------------------------------------------
<br><br><br>Mainly structure of makefile :
<br><br>file Makefile :
<br>1: set FILE, DIR and COMPILER vars.
<br>2: include default.mk
<br>3: rule 'all' for TGTS
<br>4: other project specified rules. e.g. rule for out/programme
<br>5: include rules.mk
<br><br>file default.mk :
<br>1: complete vars. e.g. DEPS, SRCDIR, OUTDIR, etc.
<br><br>file rules.mk :
<br>1: common rules. e.g. clean, copy, $(OUTDIR)/%.d, $(OBJS), $(SRCDIR)/version.h
<br>2: include dependencies
<br><br> Regards
<br>--------------
<br>L.Guo
<br>2009-12-08
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26691521&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26680139Re: Running a command always but only once (even if included Makefiles are remade)2009-12-07T07:20:09Z2009-12-07T07:20:09ZPaul Smith-20On Mon, 2009-12-07 at 12:01 +0100, Christoph Groth wrote:
<div class='shrinkable-quote'><br>> in the Makefile of my project I would like to always run (before
<br>> everything else) a command which updates the submodules of the project.
<br>>
<br>> I do this by having the line
<br>>
<br>> $(shell to-be-run)
<br>>
<br>> somewhere in the Makefile.
<br>>
<br>> However, when the automatically generated dependencies are re-generated,
<br>> which happens every time some source file has been changed, the Makefile
<br>> is re-initialized. As a consequence, `to-be-run' is executed a second
<br>> time.
<br>>
<br>> I wonder how to tell GNU make to execute a command always, but only
<br>> once, even if some of the included Makefiles are re-generated.
</div><br>You can check the make variable MAKE_RESTARTS; if it's set then make has
<br>been restarted (due to makefiles being regenerated). So, something like
<br>this should do what you want:
<br><br>ifndef MAKE_RESTARTS
<br>$(shell to-b-run)
<br>endif
<br><br>(untested)
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26680139&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26675802Running a command always but only once (even if included Makefiles are remade)2009-12-07T03:01:44Z2009-12-07T03:01:44ZChristoph Groth-2Hello,
<br><br>in the Makefile of my project I would like to always run (before
<br>everything else) a command which updates the submodules of the project.
<br><br>I do this by having the line
<br><br>$(shell to-be-run)
<br><br>somewhere in the Makefile.
<br><br>However, when the automatically generated dependencies are re-generated,
<br>which happens every time some source file has been changed, the Makefile
<br>is re-initialized. As a consequence, `to-be-run' is executed a second
<br>time.
<br><br>I wonder how to tell GNU make to execute a command always, but only
<br>once, even if some of the included Makefiles are re-generated.
<br><br>thanks,
<br>Christoph
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26675802&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26672176new french translation2009-12-06T16:12:05Z2009-12-06T16:12:05ZChristophe CombellesHello,
<br><br>The current french translation of GNU Make has a lot of mistakes and
<br>mis-interpretations. I'm the new maintainer of this translation and I have
<br>reviewed and corrected it.
<br>The new translation is available here :
<br><br><a href="http://translationproject.org/PO-files/fr/make-3.81.fr.po" target="_top" rel="nofollow">http://translationproject.org/PO-files/fr/make-3.81.fr.po</a><br><br>regards,
<br>Christophe Combelles
<br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26672176&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26668231do formats of vpath and auto variable file names always agree?2009-12-06T11:41:17Z2009-12-06T11:41:17ZMark Galeck (CW)Hello,
<br><br>It says in the manual:
<br><br>$<
<br> The name of the first prerequisite
<br><br><br>Is it guaranteed that, if vpath is used, then the directory part of the value of $< is in fact, one of the vpath paths?
<br><br>(it could be, that the value of $< were an absolute path for example, while vpath paths are relative - is this kind of thing guaranteed not to happen?).
<br><br>What about the separator between the directory path and the file name, is it guaranteed to always be "/", even on Windows?
<br><br>Similar questions about other auto variables that depend on vpath/VPATH.
<br><br>If these things are guaranteed, can they be put in the manual?
<br><br>Mark
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26668231&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26661715[bug #28189] SIGPIPE when gmake invoked recursively and vpath specified2009-12-05T18:40:51Z2009-12-05T18:40:51ZEli Zaretskii-4<br>URL:
<br> <<a href="http://savannah.gnu.org/bugs/?28189" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28189</a>>
<br><br> Summary: SIGPIPE when gmake invoked recursively and vpath
<br>specified
<br> Project: make
<br> Submitted by: yozh
<br> Submitted on: Sun 06 Dec 2009 02:40:50 AM GMT
<br> Severity: 3 - Normal
<br> Item Group: None
<br> Status: None
<br> Privacy: Public
<br> Assigned to: None
<br> Open/Closed: Open
<br> Discussion Lock: Any
<br> Component Version: None
<br> Operating System: None
<br> Fixed Release: None
<br> Triage Status: None
<br><br> _______________________________________________________
<br><br>Details:
<br><br>=====
<br># cat m1.mk
<br>xx:
<br> gmake -j2 -f m2.mk xxx.dst yyy.dst
<br><br># cat m2.mk
<br>vpath %.dst /
<br><br>xxx.dst:
<br> true
<br><br>yyy.dst:
<br> true
<br><br># gmake -j2 -f ./m1.mk
<br>gmake -j2 -f m2.mk xxx.dst yyy.dst
<br>gmake[1]: Entering directory `/place/home/nga/tmp/qwqw'
<br>gmake[1]: warning: -jN forced in submake: disabling jobserver mode.
<br>true
<br>true
<br>gmake: *** [xx] Broken pipe: 13
<br>zsh: exit 2 gmake -j2 -f ./m1.mk
<br><br># gmake --version
<br>GNU Make 3.81
<br>Copyright (C) 2006 Free Software Foundation, Inc.
<br>This is free software; see the source for copying conditions.
<br>There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
<br>PARTICULAR PURPOSE.
<br><br>This program built for amd64-portbld-freebsd7.0
<br>=====
<br><br><br><br><br> _______________________________________________________
<br><br>Reply to this item at:
<br><br> <<a href="http://savannah.gnu.org/bugs/?28189" target="_top" rel="nofollow">http://savannah.gnu.org/bugs/?28189</a>>
<br><br>_______________________________________________
<br> Message sent via/by Savannah
<br> <a href="http://savannah.gnu.org/" target="_top" rel="nofollow">http://savannah.gnu.org/</a><br><br><br><br>_______________________________________________
<br>Bug-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26661715&i=0" target="_top" rel="nofollow">Bug-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/bug-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/bug-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Bugs-f1738.html" embed="fixTarget[1738]" target="_top" >Gnu - Make - Bugs</a></p>tag:old.nabble.com,2006:post-26647966RE: is this a bug?2009-12-04T11:30:41Z2009-12-04T11:30:41ZMark Galeck (CW)>Probably we use alloca, and this blows up the stack. 50K file names
<br>100 characters each is about 0.5MB.
<br><br>No... 5MB, I checked the size of the whole thing it's about 5MB.
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647966&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26647934Re: is this a bug?2009-12-04T11:28:25Z2009-12-04T11:28:25ZEli Zaretskii<div class='shrinkable-quote'>> From: "Mark Galeck (CW)" <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647934&i=0" target="_top" rel="nofollow">mgaleck@...</a>>
<br>> Date: Fri, 4 Dec 2009 11:22:37 -0800
<br>> Accept-Language: en-US
<br>> acceptlanguage: en-US
<br>>
<br>> >If you're then passing that wildcard list to a program, you could very well be exceeding the maximum argument length limit of your shell.
<br>>
<br>>
<br>> No, I isolated this so there is nothing else, only
<br>> foobar: \
<br>> $(wildcard \
<br>> <file path> \
<br>> <file path> \
<br>> (... 50000 file paths, average 100 characters each)
<br>> <file path>\
<br>> )
<br>>
<br>> And then
<br>>
<br>> >make foobar
<br>>
<br>> produces
<br>>
<br>> >make.exe -r foobar
<br>> make.exe: Interrupt/Exception caught (code = 0xc00000fd, addr = 0x4227d3)
</div><br>Probably we use alloca, and this blows up the stack. 50K file names
<br>100 characters each is about 0.5MB.
<br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647934&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26647820RE: is this a bug?2009-12-04T11:22:37Z2009-12-04T11:22:37ZMark Galeck (CW)>If you're then passing that wildcard list to a program, you could very well be exceeding the maximum argument length limit of your shell.
<br><br><br>No, I isolated this so there is nothing else, only
<br>foobar: \
<br>$(wildcard \
<br><file path> \
<br><file path> \
<br>(... 50000 file paths, average 100 characters each)
<br><file path>\
<br>)
<br><br>And then
<br><br>>make foobar
<br><br>produces
<br><br>>make.exe -r foobar
<br>make.exe: Interrupt/Exception caught (code = 0xc00000fd, addr = 0x4227d3)
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647820&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26647606Re: is this a bug?2009-12-04T11:07:14Z2009-12-04T11:07:14ZStephan Beal-2<div class="gmail_quote">On Fri, Dec 4, 2009 at 7:46 PM, Mark Galeck (CW) <span dir="ltr"><<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647606&i=0" target="_top" rel="nofollow">mgaleck@...</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Well, I found what was causing this. I attempted to call the<br>
$(wildcard ...<br>
function, with about 50,000 arguments (full path names, on average 100 characters, so the total size of all strings that were passed to wildcard, is about 5MB). Is this a bad idea? :)<br></blockquote><div><br></div><div>
If you're then passing that wildcard list to a program, you could very well be exceeding the maximum argument length limit of your shell. Early versions of DOS's <a href="http://command.com" target="_top" rel="nofollow">command.com</a> allowed only something like 128 characters per command line, and i _think_ the average limit on Unix is 32k (the maximum signed int value, IIRC, since that's what main() takes as its argc type).</div>
<div><br></div></div>-- <br>----- stephan beal<br><a href="http://wanderinghorse.net/home/stephan/" target="_top" rel="nofollow">http://wanderinghorse.net/home/stephan/</a><br>
<br />_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647606&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26647283RE: is this a bug?2009-12-04T10:46:17Z2009-12-04T10:46:17ZMark Galeck (CW)Well, I found what was causing this. I attempted to call the
<br>$(wildcard ...
<br>function, with about 50,000 arguments (full path names, on average 100 characters, so the total size of all strings that were passed to wildcard, is about 5MB). Is this a bad idea? :)
<br><br><br><br>>make.exe -r --version
<br>GNU Make 3.81
<br>Copyright (C) 2006 Free Software Foundation, Inc.
<br>This is free software; see the source for copying conditions.
<br>There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
<br>PARTICULAR PURPOSE.
<br><br>This program built for i386-pc-mingw32
<br><br>-----Original Message-----
<br>From: Greg Chicares [mailto:<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647283&i=0" target="_top" rel="nofollow">gchicares@...</a>]
<br><br><br>'make' caught a stack overflow.
<br>It would be helpful to know what the makefile contains, and what
<br> make --version
<br>says.
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26647283&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26644663Re: How specify dependences on URL2009-12-04T08:00:06Z2009-12-04T08:00:06ZPeng YuOn Fri, Dec 4, 2009 at 8:37 AM, Allan Odgaard
<br><<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26644663&i=0" target="_top" rel="nofollow">682E7718-3533-4FED-AF3B-CD0D7E604283@...</a>> wrote:
<div class='shrinkable-quote'><br>> On 4 Dec 2009, at 14:19, Peng Yu wrote:
<br>>
<br>>> Suppose I want to download the file <a href="http://some_website/some_file" target="_top" rel="nofollow">http://some_website/some_file</a><br>>> whenever it is newer than the local file. This needs the specification
<br>>> of the dependency on URLs. I'm wondering if it is possible to do so in
<br>>> gnu make.
<br>>
<br>> URLs don’t have dates. A GET/HEAD request may return a ‘Last-Modified’
<br>> header and you can do conditional requests using that or the entity tag, but
<br>> these things are beyond make.
</div><br>Some programs (e.g. rsync) support various network protocols. It would
<br>be nice that gnu make also support these protocols. Will this
<br>functionality be considered to be added in gnu make in the future?
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26644663&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26643516Re: How specify dependences on URL2009-12-04T06:37:45Z2009-12-04T06:37:45ZAllan Odgaard-8On 4 Dec 2009, at 14:19, Peng Yu wrote:
<br><br>> Suppose I want to download the file <a href="http://some_website/some_file" target="_top" rel="nofollow">http://some_website/some_file</a><br>> whenever it is newer than the local file. This needs the specification
<br>> of the dependency on URLs. I'm wondering if it is possible to do so in
<br>> gnu make.
<br><br>URLs don’t have dates. A GET/HEAD request may return a ‘Last-Modified’
<br>header and you can do conditional requests using that or the entity
<br>tag, but these things are beyond make.
<br><br><br><br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26643516&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26642926Re: make undefined $(MY_VARIABLE) usage query2009-12-04T05:48:28Z2009-12-04T05:48:28ZBugzilla from jg@jguk.org2009/12/4 Paul Smith <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642926&i=0" target="_top" rel="nofollow">psmith@...</a>>:
<br>> On Fri, 2009-12-04 at 11:00 +0000, Jon Grant wrote:
<br>>> I looked through the manual (3.81) and could not find a way to get GNU
<br>>> Make to display an error, or warning when a variable is used without
<br>>> being defined. like $(MY_VARIABLE)
<br>>
<br>> Not sure where you were looking, but the best place to start is the
<br>> "Summary of Options" section in the manual.
<br>>
<br>> Look there for the --warn-undefined-variables option.
<br><br>(Just to declare, I am using the Cygwin build)
<br><br>I noticed that if I add the additional parameter to our MAKEFLAGS, and
<br>misspell it, the -debug mode seems to get enabled. Is this intended? I
<br>had expected invalid parameters to be warned, and ignored.
<br><br>MAKEFLAGS := --no-print-directory
<br>MAKEFLAGS := -warn-undefined-variables
<br>export MAKEFLAGS
<br><br>output is lots of:
<br><br> Considering target file `makefile'.
<br> Looking for an implicit rule for `makefile'.
<br> No implicit rule found for `makefile'.
<br> Finished prerequisites of target file `makefile'.
<br> No need to remake target `makefile'.
<br>Updating goal targets....
<br>Considering target file `all'.
<br> File `all' does not exist.
<br> Looking for an implicit rule for `all'.
<br> No implicit rule found for `all'.
<br> Considering target file `/cygdrive/c/samples/dir.lib'.
<br> Considering target file `obj/utf8.o'.
<br><br>Best regards, Jon
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642926&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26642906Re: make undefined $(MY_VARIABLE) usage query2009-12-04T05:25:53Z2009-12-04T05:25:53ZBugzilla from jg@jguk.org2009/12/4 Paul Smith <<a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642906&i=0" target="_top" rel="nofollow">psmith@...</a>>:
<br>> On Fri, 2009-12-04 at 11:00 +0000, Jon Grant wrote:
<br>>> I looked through the manual (3.81) and could not find a way to get GNU
<br>>> Make to display an error, or warning when a variable is used without
<br>>> being defined. like $(MY_VARIABLE)
<br>>
<br>> Not sure where you were looking, but the best place to start is the
<br>> "Summary of Options" section in the manual.
<br>>
<br>> Look there for the --warn-undefined-variables option.
<br><br>Great, thank you for the pointer.
<br>Jon
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642906&i=1" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26642385How specify dependences on URL2009-12-04T05:19:03Z2009-12-04T05:19:03ZPeng YuSuppose I want to download the file <a href="http://some_website/some_file" target="_top" rel="nofollow">http://some_website/some_file</a><br>whenever it is newer than the local file. This needs the specification
<br>of the dependency on URLs. I'm wondering if it is possible to do so in
<br>gnu make.
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642385&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>tag:old.nabble.com,2006:post-26642355Re: is this a bug?2009-12-04T05:15:39Z2009-12-04T05:15:39ZGreg Chicares-2On 2009-12-04 04:07Z, Mark Galeck (CW) wrote:
<br>>>make.exe -r
<br>> make.exe: Interrupt/Exception caught (code = 0xc00000fd, addr = 0x4227d3)
<br><br>'make' caught a stack overflow. You might try adding the '-d' flag:
<br> make -dr
<br>It would be helpful to know what the makefile contains, and what
<br> make --version
<br>says.
<br><br><br>_______________________________________________
<br>Help-make mailing list
<br><a href="http://old.nabble.com/user/SendEmail.jtp?type=post&post=26642355&i=0" target="_top" rel="nofollow">Help-make@...</a>
<br><a href="http://lists.gnu.org/mailman/listinfo/help-make" target="_top" rel="nofollow">http://lists.gnu.org/mailman/listinfo/help-make</a><br><p>From forum: <a href="http://old.nabble.com/Gnu---Make---Help-f1739.html" embed="fixTarget[1739]" target="_top" >Gnu - Make - Help</a></p>