August 2008 - Darwin-dev - lists.apple.com

Re: select() and svn 1.5
by Jordan K. Hubbard 31 Aug '08

31 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 30, 2008, at 5:54 PM, Luke Daley wrote: On 31/08/2008, at 3:59 AM, Jordan K. Hubbard wrote: - Jordan _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… And the grand prize goes to Jonas, for correctly intuiting that svn 1.5 is clearly built and running 64 bit, whereas the interpositioning library is probably not built with the same architectural support! I might further speculate that Luke is running a SnowLeopard seed since 64 bit execution is not the default in Leopard, and he would have to either force 64 bit execution or perhaps even have an svn 1.5 binary *without* the right 32 bit section which is now "falling back" to 64 bit, but that's pretty far-fetched. I'm going with SnowLeopard seed. :-) Either way, the solution is clear: Make sure that otool -f displays the same architecture sections (or use use file(1) - it's more human readable) for both the interposing dylib and the svn binary(s). This _is_ on Leopard. :) Yeah, sorry for the misinformation there - I was misremembering the Leopard defaults. So that would be the same? Here is how we build this thing… Well, clearly the 64 bit interpositioning case is the one that is biting you, and if you've already clarified that all the select() variants are STILL converging on your interpositioning routines even in the 64 bit ABI case, then I guess you need to look very carefully at your interpositioning code and see if you have any 32 vs 64 bit size issues anywhere, that's the only thing I can think of (getting a 64 bit value in and truncating it to 32 bits then passing it on to a syscall which expects 64 bits would do it). This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Jordan K. Hubbard 31 Aug '08

31 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 30, 2008, at 5:50 PM, Clark Cox wrote: - Jordan _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This is not true, Leopard will prefer 64-bit executables wherever possible. Whoops! You're right. Sorry, I thought this hadn't been changed in Leopard. This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Luke Daley 31 Aug '08

31 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 31/08/2008, at 3:59 AM, Jordan K. Hubbard wrote: This _is_ on Leopard. :) So that would be the same? Here is how we build this thing… #!/usr/bin/env bash #NDEBUG=1 SCRIPT_DIR="$(dirname "$0")" SRC_DIR="$SCRIPT_DIR/src" DST_DIR="$SCRIPT_DIR/build" LIB_NAME="tm_interactive_input.dylib" function build { DEPLOYMENT=10.4 SDK=/Developer/SDKs/MacOSX10.4u.sdk if [[ "$1" = *64 ]]; then DEPLOYMENT=10.5 SDK=/Developer/SDKs/MacOSX10.5.sdk fi if [[ -d "$DST_DIR" ]]; then echo "Cleaning old build dir (‘"$DST_DIR"’)…" rm -rf "$DST_DIR" fi if ! mkdir "$DST_DIR"; then exit; fi for ARCH in ppc i386 ppc64 x86_64; do build "$ARCH"; done SUPPORT_LIB="$TM_SUPPORT_PATH/lib" echo "Copying to support…" cp "$LIB_NAME" "$SUPPORT_LIB" -- LD. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… And the grand prize goes to Jonas, for correctly intuiting that svn 1.5 is clearly built and running 64 bit, whereas the interpositioning library is probably not built with the same architectural support! I might further speculate that Luke is running a SnowLeopard seed since 64 bit execution is not the default in Leopard, and he would have to either force 64 bit execution or perhaps even have an svn 1.5 binary *without* the right 32 bit section which is now "falling back" to 64 bit, but that's pretty far-fetched. I'm going with SnowLeopard seed. :-) Either way, the solution is clear: Make sure that otool -f displays the same architecture sections (or use use file(1) - it's more human readable) for both the interposing dylib and the svn binary(s). » file /usr/local/bin/svn /usr/local/bin/svn: Mach-O universal binary with 4 architectures /usr/local/bin/svn (for architecture ppc): Mach-O executable ppc /usr/local/bin/svn (for architecture i386): Mach-O executable i386 /usr/local/bin/svn (for architecture ppc64): Mach-O 64-bit executable ppc64 /usr/local/bin/svn (for architecture x86_64): Mach-O 64-bit executable x86_64 » file tm_interactive_input.dylib tm_interactive_input.dylib: Mach-O universal binary with 4 architectures tm_interactive_input.dylib (for architecture i386): Mach-O dynamically linked shared library i386 tm_interactive_input.dylib (for architecture ppc): Mach-O dynamically linked shared library ppc tm_interactive_input.dylib (for architecture ppc64): Mach-O 64-bit dynamically linked shared library ppc64 tm_interactive_input.dylib (for architecture x86_64): Mach-O 64-bit dynamically linked shared library x86_64 echo "Building ‘tm_interactive_input_$1.dylib’ (for ${DEPLOYMENT}$ {NDEBUG:+, no debug})…" gcc -dynamiclib -Wmost -Os -fno-common \ -framework CoreFoundation \ -DDATE=\"$(date +%Y-%m-%d)\" \ ${NDEBUG:+-DNDEBUG=1} \ -mmacosx-version-min="$DEPLOYMENT" -isysroot "$SDK" \ -arch "$1" \ -o "$DST_DIR/tm_interactive_input_$1.dylib" \ "$SRC_DIR"/*.c [ $? = 0 ] || exit 1 } echo "Merging…" cd "$DST_DIR" lipo -create tm_interactive_input_*.dylib -output "$LIB_NAME" && rm tm_interactive_input_*.dylib lipo -info "$LIB_NAME" Do I need to do some different things in my code depending on what I am building for maybe? This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Clark Cox 31 Aug '08

31 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:to :subject:cc:in-reply-to:mime-version:content-type :content-transfer-encoding:content-disposition:references; bh=u0LbuJ2tsZcSbBCgLNG4g4QKfybHe3PAtJAnZ75+/XE=; b=o7PaF5lWZamYo+04wOuYxsXp4mnrKgdGDIjWJBfXywgpFrWfgR/yp9ktoLpaSAQyo3 jawSkvVAQ5e65I9WlIn+vj2Qqn4SZKTCEZEFo/AKdgkpUkSMSBWCj7dfrthLBJqjtHxt V3RpcIJLpuxupSvt3AuPUM5moG9U+tVMI0+as= Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version :content-type:content-transfer-encoding:content-disposition :references; b=q/FrJOuQ2tnQnGnj2zlisR/SXmJMhx+N2DTXJsj+iSD9fZ3/Z9BCcBWGxF/bGEBX0V iC/uVA0jbpBd5eNn678nC1QAYimVZgjJJ0yN7jr4rHOrtYNxFW0gqdBH9qOGJXu1Pv2s /J2xtiw5ml6WXtksQCQx83UNrBw2idvJblSPc= On Sat, Aug 30, 2008 at 10:59 AM, Jordan K. Hubbard <jkh(a)apple.com> wrote: > I might further speculate that Luke is running a SnowLeopard seed since 64 > bit execution is not the default in Leopard This is not true, Leopard will prefer 64-bit executables wherever possible. -- Clark S. Cox III clarkcox3(a)gmail.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Jordan K. Hubbard 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Jonas Maebe 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 30 Aug 2008, at 09:24, Luke Daley wrote: Interesting. Here is svn 1.4… [snip] And 1.5… [snip] Jonas _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… svn: select(nfds=4, readfds=bfffe918, writefds=0, errorfds=0, timeout=bfffe998) svn: select(nfds=4, readfds=7fff5fbfe580, writefds=7fff5fbfe500, errorfds=0, timeout=7fff5fbfe4f0) Tried several times and got the same behaviour all times. The command was `svn log http://macromates.com/svn/Bundles --limit 1`, without my library injected. This unfortunately doesn't ring any bells for me. What I don't understand is why the syscall to select in my code, using the exact params that were passed to select untouched, fails. Is your library perhaps 32 bit only? It appears the copy of svn 1.4 you have is 32 bit, while the 1.5 one is 64 bit. This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Luke Daley 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On 30/08/2008, at 5:56 PM, Jordan K. Hubbard wrote: Yes please :) Interesting. Here is svn 1.4… And 1.5… -- LD. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… (some nice hooks in Libsystem, for example, right below the point where all the API variants converge so you don't have do all that crazy entrypoint gymnastics... Hurm... Mumble... Moving on...). What happens with svn (say `svn log http://blah.com`) is that it calls select() with a single fd in the writefds set (presumably the fd that the network request is going out on) and a timeout of 3600. With svn < 1.5 with this implementation injected everything works fine. With svn 1.5, the syscall() returns immediately with -1 and the error is 'Bad file descriptor'. Well, one obvious question, I suppose, is what happens when you trace both svn 1.4 and svn 1.5 and then compare the results? It's easy enough to do stuff like this with drace, to wit: svn: select(nfds=4, readfds=bfffe918, writefds=0, errorfds=0, timeout=bfffe998) svn: select(nfds=4, readfds=bfffe988, writefds=0, errorfds=0, timeout=bfffea08) svn: select(nfds=4, readfds=bfffe918, writefds=0, errorfds=0, timeout=bfffe998) svn: select(nfds=4, readfds=bfffe928, writefds=0, errorfds=0, timeout=bfffe9a8) svn: select(nfds=4, readfds=bfffe928, writefds=0, errorfds=0, timeout=bfffe9a8) svn: select(nfds=4, readfds=bfffe988, writefds=0, errorfds=0, timeout=bfffea08) svn: select(nfds=4, readfds=bfffea88, writefds=0, errorfds=0, timeout=bfffeb08) svn: select(nfds=4, readfds=bfffea98, writefds=0, errorfds=0, timeout=bfffeb18) svn: select(nfds=4, readfds=bfffe9e8, writefds=0, errorfds=0, timeout=bfffea68) svn: select(nfds=4, readfds=bfffe9f8, writefds=0, errorfds=0, timeout=bfffea78) svn: select(nfds=4, readfds=bfffea58, writefds=0, errorfds=0, timeout=bfffead8) svn: select(nfds=4, readfds=bfffeaf8, writefds=0, errorfds=0, timeout=bfffeb78) svn: select(nfds=4, readfds=bfffea48, writefds=0, errorfds=0, timeout=bfffeac8) svn: select(nfds=4, readfds=bfffeab8, writefds=0, errorfds=0, timeout=bfffeb38) svn: select(nfds=4, readfds=bfffec78, writefds=0, errorfds=0, timeout=bfffecf8) svn: select(nfds=4, readfds=bfffef48, writefds=0, errorfds=0, timeout=bfffefc8) svn: select(nfds=4, readfds=7fff5fbfe580, writefds=7fff5fbfe500, errorfds=0, timeout=7fff5fbfe4f0) svn: select(nfds=4, readfds=7fff5fbfe660, writefds=7fff5fbfe5e0, errorfds=0, timeout=7fff5fbfe5d0) svn: select(nfds=4, readfds=7fff5fbfe260, writefds=7fff5fbfe1e0, errorfds=0, timeout=7fff5fbfe1d0) svn: select(nfds=4, readfds=7fff5fbfe2f0, writefds=7fff5fbfe270, errorfds=0, timeout=7fff5fbfe260) svn: select(nfds=4, readfds=7fff5fbfe380, writefds=7fff5fbfe300, errorfds=0, timeout=7fff5fbfe2f0) svn: select(nfds=4, readfds=7fff5fbfe4e0, writefds=7fff5fbfe460, errorfds=0, timeout=7fff5fbfe450) svn: select(nfds=4, readfds=7fff5fbfe4c0, writefds=7fff5fbfe440, errorfds=0, timeout=7fff5fbfe430) svn: select(nfds=4, readfds=7fff5fbfe3c0, writefds=7fff5fbfe340, errorfds=0, timeout=7fff5fbfe330) svn: select(nfds=4, readfds=7fff5fbfe450, writefds=7fff5fbfe3d0, errorfds=0, timeout=7fff5fbfe3c0) svn: select(nfds=4, readfds=7fff5fbfe550, writefds=7fff5fbfe4d0, errorfds=0, timeout=7fff5fbfe4c0) svn: select(nfds=5, readfds=7fff5fbfe740, writefds=7fff5fbfe6c0, errorfds=0, timeout=7fff5fbfe6b0) svn: select(nfds=6, readfds=7fff5fbfe580, writefds=7fff5fbfe500, errorfds=0, timeout=7fff5fbfe4f0) svn: select(nfds=4, readfds=7fff5fbfe5e0, writefds=7fff5fbfe560, errorfds=0, timeout=7fff5fbfe550) svn: select(nfds=5, readfds=7fff5fbfe660, writefds=7fff5fbfe5e0, errorfds=0, timeout=7fff5fbfe5d0) svn: select(nfds=6, readfds=7fff5fbfe660, writefds=7fff5fbfe5e0, errorfds=0, timeout=7fff5fbfe5d0) svn: select(nfds=6, readfds=7fff5fbfe450, writefds=7fff5fbfe3d0, errorfds=0, timeout=7fff5fbfe3c0) svn: select(nfds=6, readfds=7fff5fbfe260, writefds=7fff5fbfe1e0, errorfds=0, timeout=7fff5fbfe1d0) svn: select(nfds=6, readfds=7fff5fbfe2f0, writefds=7fff5fbfe270, errorfds=0, timeout=7fff5fbfe260) svn: select(nfds=6, readfds=7fff5fbfe380, writefds=7fff5fbfe300, errorfds=0, timeout=7fff5fbfe2f0) svn: select(nfds=6, readfds=7fff5fbfe3c0, writefds=7fff5fbfe340, errorfds=0, timeout=7fff5fbfe330) svn: select(nfds=6, readfds=7fff5fbfe4e0, writefds=7fff5fbfe460, errorfds=0, timeout=7fff5fbfe450) svn: select(nfds=6, readfds=7fff5fbfe3d0, writefds=7fff5fbfe350, errorfds=0, timeout=7fff5fbfe340) svn: select(nfds=6, readfds=7fff5fbfe4f0, writefds=7fff5fbfe470, errorfds=0, timeout=7fff5fbfe460) svn: select(nfds=6, readfds=7fff5fbfe800, writefds=7fff5fbfe780, errorfds=0, timeout=7fff5fbfe770) svn: select(nfds=6, readfds=7fff5fbfe840, writefds=7fff5fbfe7c0, errorfds=0, timeout=7fff5fbfe7b0) Tried several times and got the same behaviour all times. The command was `svn log http://macromates.com/svn/Bundles --limit 1`, without my library injected. This unfortunately doesn't ring any bells for me. What I don't understand is why the syscall to select in my code, using the exact params that were passed to select untouched, fails. This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Jordan K. Hubbard 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com On Aug 29, 2008, at 11:16 PM, Luke Daley wrote: HTH - Jordan _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I have been developing a library that interposes several system calls . This is used in TextMate to intercept attempts to read from STDIN for the purpose of displaying a dialog to allow the user to enter input. This works for everything I have tested so far (e.g. Java, Ruby, Python, svn < 1.5), but not svn 1.5. Ah, you naughty boy... Just wait until you happen to execute something which, for security or other reasons, completely cleanses its environment and kicks the crutches out from under you on the next fork(). :-) Still, it's hard to blame you for using dyld inter- positioning when we don't really provide you with any other choice for dealing with scenarios like this one (some nice hooks in Libsystem, for example, right below the point where all the API variants converge so you don't have do all that crazy entrypoint gymnastics... Hurm... Mumble... Moving on...). DISCLAIMER: This is all magic to me and I don't understand what __asm() is really doing and that could very well be my problem. All the __asm()s are doing in your example is spitting out extra assembler labels. The label emitted by each _asm() overrides the label that would have been used instead (e.g. "select$UNIX03" instead of "select"). To put it another way, it's a poor-man's versioned symbol hack with attribute tags instead of version numbers (and no help from the dynamic loader). Sometimes you have to work with what you've got! What happens with svn (say `svn log http://blah.com`) is that it calls select() with a single fd in the writefds set (presumably the fd that the network request is going out on) and a timeout of 3600. With svn < 1.5 with this implementation injected everything works fine. With svn 1.5, the syscall() returns immediately with -1 and the error is 'Bad file descriptor'. Well, one obvious question, I suppose, is what happens when you trace both svn 1.4 and svn 1.5 and then compare the results? It's easy enough to do stuff like this with drace, to wit: #!/usr/sbin/dtrace -s #pragma D option quiet syscall::select*:entry { printf("%s: %s(nfds=%d, readfds=%x, writefds=%x, errorfds=%x, timeout= %x)\n", execname, probefunc, arg0, arg1, arg2, arg3, arg4); } There, a stand-alone shell script which will show system-wide usage of select (for extra credit cleverness, I took care to display probefunc so you also get all the aliased calls). That should show you fairly quickly how the behavior is diverging, at which point you're at least a lot closer to putting your finger on the root cause. (I also left it as an exercise for the reader the task of translating the individual arguments to select() and printing them all purty-like). This email sent to site_archiver(a)lists.apple.com

1 0

Re: select() and svn 1.5
by Jean-Daniel Dupas 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Le 30 août 08 à 08:16, Luke Daley a écrit : Hi, I essentially have this… A note about this Magic may help you: http://developer.apple.com/releasenotes/Darwin/SymbolVariantsRelNotes/index… _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I have been developing a library that interposes several system calls . This is used in TextMate to intercept attempts to read from STDIN for the purpose of displaying a dialog to allow the user to enter input. This works for everything I have tested so far (e.g. Java, Ruby, Python, svn < 1.5), but not svn 1.5. int select(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select"); int select_darwinextsn(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select$DARWIN_EXTSN"); int select_darwinextsn_nocancel(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select$DARWIN_EXTSN$NOCANCEL"); int select_nocancel_unix2003(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select$NOCANCEL$UNIX2003"); int select_unix2003(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select $UNIX2003"); int select(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return syscall(SYS_select, nfds, readfds, writefds, errorfds, timeout); } int select_darwinextsn(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_darwinextsn_nocancel(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_nocancel_unix2003(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_unix2003(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } DISCLAIMER: This is all magic to me and I don't understand what __asm() is really doing and that could very well be my problem. What happens with svn (say `svn log http://blah.com`) is that it calls select() with a single fd in the writefds set (presumably the fd that the network request is going out on) and a timeout of 3600. With svn < 1.5 with this implementation injected everything works fine. With svn 1.5, the syscall() returns immediately with -1 and the error is 'Bad file descriptor'. Is anyone able to shed any light on what might be happening here? Is my approach with the different symbols too naive? Been banging my head on this one for months, so any help would be appreciated. This email sent to site_archiver(a)lists.apple.com

1 0

select() and svn 1.5
by Luke Daley 30 Aug '08

30 Aug '08

site_archiver(a)lists.apple.com Delivered-To: darwin-dev(a)lists.apple.com Hi, I essentially have this… -- LD. _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev(a)lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.app… I have been developing a library that interposes several system calls . This is used in TextMate to intercept attempts to read from STDIN for the purpose of displaying a dialog to allow the user to enter input. This works for everything I have tested so far (e.g. Java, Ruby, Python, svn < 1.5), but not svn 1.5. int select(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select"); int select_darwinextsn(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select $DARWIN_EXTSN"); int select_darwinextsn_nocancel(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select$DARWIN_EXTSN$NOCANCEL"); int select_nocancel_unix2003(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select$NOCANCEL$UNIX2003"); int select_unix2003(int, fd_set * __restrict, fd_set * __restrict, fd_set * __restrict, struct timeval * __restrict) __asm("_select $UNIX2003"); int select(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return syscall(SYS_select, nfds, readfds, writefds, errorfds, timeout); } int select_darwinextsn(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_darwinextsn_nocancel(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_nocancel_unix2003(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } int select_unix2003(int nfds, fd_set * __restrict readfds, fd_set * __restrict writefds, fd_set * __restrict errorfds, struct timeval * __restrict timeout) { return select(nfds, readfds, writefds, errorfds, timeout); } DISCLAIMER: This is all magic to me and I don't understand what __asm() is really doing and that could very well be my problem. What happens with svn (say `svn log http://blah.com`) is that it calls select() with a single fd in the writefds set (presumably the fd that the network request is going out on) and a timeout of 3600. With svn < 1.5 with this implementation injected everything works fine. With svn 1.5, the syscall() returns immediately with -1 and the error is 'Bad file descriptor'. Is anyone able to shed any light on what might be happening here? Is my approach with the different symbols too naive? Been banging my head on this one for months, so any help would be appreciated. This email sent to site_archiver(a)lists.apple.com

1 0