|--==> "SY" == Steve Youngs <youngs(a)xemacs.org&gt; writes: SY> - A picture paints a 1000 words... Mike? I suppose it'd help if I actually attached the picture... -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|

|--==> "jpw" == John Paul Wallington <jpw(a)shootybangbang.com&gt; writes: jpw> Steve Youngs <youngs(a)xemacs.org&gt; wrote: jpw> This sounds like it could be related to making jpw> `line-beginning-position' an alias for `point-at-bol'. jpw> I tried (require 'jde-db) with JDE package version 1.38 and XEmacs jpw> 21.5.9 (+CVS-20021228) but didn't get the error. I don't use JDE. Is Probably because the version of JDE that you have was built with XEmacs 21.4 (Rendhalver builds packages with 21.4). And besides, I only see the problem when building the package. jpw> there an easy way to reproduce it? Also, could you post a backtrace? To reproduce it just build the JDE package with XEmacs 21.5. The backtrace: While compiling jde-db-breakpoint-marker::initialize-instance in file /home/steve/test-it/build/packages/xemacs-packages/jde/lisp/jde-db.el: !! Wrong type argument ((listp #<subr point-at-bol>)) backtrace(nil t) # bind (error-info) byte-compile-report-error((wrong-type-argument listp #<subr point-at-bol>)) # bind (error-info) #<compiled-function (error-info) "...(4)" [error-info byte-compile-report-error] 2>((wrong-type-argument listp #<subr point-at-bol>)) # bind (fn name form) byte-compile-inline-expand((point-at-bol)) # bind (fn name form) byte-compile-inline-expand((line-beginning-position)) # bind (fn name form) byte-compile-inline-expand((jde-line-beginning-position)) # bind (opt new for-effect form) byte-optimize-form((jde-line-beginning-position)) mapcar(byte-optimize-form ((jde-line-beginning-position))) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((extent-at (jde-line-beginning-position)) nil) # bind (for-effect form) byte-optimize-form((extent-at (jde-line-beginning-position))) mapcar(byte-optimize-form ((extent-at (jde-line-beginning-position)))) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((not (extent-at (jde-line-beginning-position))) nil) # bind (for-effect form) byte-optimize-form((not (extent-at (jde-line-beginning-position)))) mapcar(byte-optimize-form ((not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position)))))) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) nil) # bind (for-effect form) byte-optimize-form((or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) nil) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) nil) # bind (for-effect form) byte-optimize-form((if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) nil) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)) nil) # bind (for-effect form) byte-optimize-form((if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))) mapcar(byte-optimize-form (this (quote marker) (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)))) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((eieio-oset this (quote marker) (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))) nil) # bind (for-effect form) byte-optimize-form((eieio-oset this (quote marker) (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))) nil) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))) nil) # bind (for-effect form) byte-optimize-form((oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))) nil) # bind (rest result fe new all-for-effect forms) byte-optimize-body(((call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)))) nil) # bind (fn tmp for-effect form) byte-optimize-form-code-walker((progn (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)))) nil) # bind (for-effect form) byte-optimize-form((progn (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)))) nil) # bind (byte-compile-constants byte-compile-variables byte-compile-tag-number byte-compile-depth byte-compile-maxdepth byte-compile-output output-type for-effect form) byte-compile-top-level((progn (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t)))) nil lambda) # bind (int doc body byte-compile-bound-variables arglist fun) byte-compile-lambda((lambda (this &rest fields) "Create a breakpoint overlay at LINE in FILE." (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))))) # bind (my-outbuffer class arg1 lamparams params key meth form) byte-compile-file-form-defmethod((defmethod initialize-instance ((this jde-db-breakpoint-marker) &rest fields) "Create a breakpoint overlay at LINE in FILE." (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))))) # bind (byte-compile-current-form handler form) byte-compile-file-form((defmethod initialize-instance ((this jde-db-breakpoint-marker) &rest fields) "Create a breakpoint overlay at LINE in FILE." (call-next-method) (oset this marker (if (featurep (quote xemacs)) (if (or (not (extent-at (jde-line-beginning-position))) (not (jde-db-breakpoint-marker-p (extent-at (jde-line-beginning-position))))) (make-extent (jde-line-beginning-position) (1+ (jde-line-end-position)))) (make-overlay (jde-line-beginning-position) (1+ (jde-line-end-position)) (current-buffer) nil t))))) # (unwind-protect ...) #<compiled-function nil "...(43)" [byte-compile-unresolved-functions byte-compile-inbuffer 1 " \n" nil looking-at ";" byte-compile-file-form read byte-compile-flush-pending byte-compile-warn-about-unresolved-functions] 3>() # (unwind-protect ...) call-with-condition-handler(#<compiled-function (error-info) "...(4)" [error-info byte-compile-report-error] 2> #<compiled-function nil "...(43)" [byte-compile-unresolved-functions byte-compile-inbuffer 1 " \n" nil looking-at ";" byte-compile-file-form read byte-compile-flush-pending byte-compile-warn-about-unresolved-functions] 3>) # (unwind-protect ...) # bind (byte-compile-warnings-beginning byte-compile-point-max-prev byte-compile-log-buffer byte-compile-macro-environment byte-compile-function-environment byte-compile-autoload-environment byte-compile-unresolved-functions byte-compile-bound-variables byte-compile-free-references byte-compile-free-assignments byte-compile-verbose byte-optimize byte-compile-emacs19-compatibility byte-compile-dynamic byte-compile-dynamic-docstrings byte-compile-warnings byte-compile-file-domain byte-compile-outbuffer float-output-format case-fold-search print-length print-level byte-compile-constants byte-compile-variables byte-compile-tag-number byte-compile-depth byte-compile-maxdepth byte-compile-output eval filename byte-compile-inbuffer) byte-compile-from-buffer(#<buffer " *Compiler Input*"> "/home/steve/test-it/build/packages/xemacs-packages/jde/lisp/jde-db.el") # bind (byte-compile-current-file target-file input-buffer output-buffer byte-compile-dest-file load filename) byte-compile-file("lisp/jde-db.el") byte-code("..." [file byte-compile-file t] 2) # (condition-case ... . ((error (byte-code "ÂÃ!.Â!.ÂÄ!.ÅÆ!«.Æ Ç\".ª. ＠ÈN®. ＠!.ÉÊ A\".ÂË!.Ç." [file err princ ">>Error occurred processing " ": " fboundp display-error nil error-message mapcar #<compiled-function (x) "...(8)" [x princ " " prin1] 2> "\n"] 3)))) # bind (file) batch-byte-compile-1("lisp/jde-db.el") # bind (error file-to-process) batch-byte-compile-one-file() # bind (error) batch-byte-compile() # bind (arg) command-line-do-funcall("-f") # bind (dir file-count line end-of-options file-p arg tem) command-line-1() # bind (command-line-args-left) command-line() # bind (error-data) normal-top-level() # (condition-case ... . error) # (catch top-level ...) Wrong type argument: listp, #<subr point-at-bol> Done gmake: *** [lisp/jde-db.elc] Error 1 Compilation exited abnormally with code 2 at Mon Dec 30 15:32:16 -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|

On Mon, 2002-12-30 at 08:06, Steve Youngs wrote: Yes, I can reproduce this, and they're visible in Mats' smoketest logs too (Steve, take a look at <http://labb.contactor.se/~matsl/smoketest/logs/> if you're not familiar with the feature yet). Hmm. Which one should be fixed, JDE or 21.5? -- \/ille Skyttä scop at xemacs.org

On Tue, 2002-12-31 at 00:06, Steve Youngs wrote: Maybe; line-beginning-position appeared in 21.5, that's what's being looked for in the start of jde-db.el. Here's a somewhat minimized test case which fails byte compilation with 21.5 but not with 21.[14]. Would wrapping the (if ...) inside eval-when-compile be TRTD? Uh, and 21.[14] tells me about this: While compiling foo-test in file /home/scop/test.el: ** attempt to inline line-beginning-position before it was defined While compiling the end of the data: ** the function line-beginning-position is not known to be defined. -- \/ille Skyttä scop at xemacs.org

|--==> "JJ" == Jerry James <james(a)xemacs.org&gt; writes: JJ> Steve Youngs <youngs(a)xemacs.org&gt; wrote: [...] JJ> And I'm still not able to reproduce it. Sigh. Is there any chance we JJ> can go back to talking about me getting ssh access to your box so I can JJ> watch this in action? Sure, I'll email you privately about it. -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|

Steve Youngs <youngs(a)xemacs.org&gt; wrote: I can finally reproduce this on my box, after watching the build process on Steve's box. (Thanks, Steve!) Two elements must be present to produce this behavior: 1) Configure with --error-checking=all 2) Compile with no optimization This is taking place on Linux x86 boxes with gcc 2.95.3 in Steve's case, and RedHat's so-called gcc 2.96 (RedHat 7.3) in my case. This all arises from lrecord.h. The name error_check_pgconn is declared in DECLARE_LRECORD, and used in XRECORD. In the case where the misbehavior manifests itself, the preprocessed source (using the same compiler flags used by the Makefile, except with -E replacing -c) contains this before any use of error_check_pgconn (with white space changes for clarity, coupled with removal of redundant parentheses and a little constant folding): extern const struct lrecord_implementation lrecord_pgconn; inline extern Lisp_PGconn * error_check_pgconn (Lisp_Object obj, const char *file, int line); inline extern Lisp_PGconn * error_check_pgconn (Lisp_Object obj, const char *file, int line) { ((enum Lisp_Type)((unsigned long)obj & ~((1UL << 30 - 1UL) << 2)) == Lisp_Type_Record && ((struct lrecord_header *)obj)->type == (unsigned int)lrecord_type_pgconn) ? (void)0 : assert_failed(file, line, "RECORD_TYPEP (obj, lrecord_type_pgconn)")); return (Lisp_PGconn *)((void *)obj); } Doing "objdump -t postgresql.ell" on the offending postgresql.ell says (in part): postgresql.ell: file format elf32-i386 SYMBOL TABLE: 00000094 l d .hash 00000000 0000070c l d .dynsym 00000000 ... 00000000 *UND* 00000000 error_check_pgconn ... On the other hand, using *any* optimization makes the error_check_pgconn entry disappear altogether; the inlining is complete. Furthermore, a disassembly shows that error_check_pgconn is not called anywhere, so it was compiled away even without optimization ... but an undefined entry for it was inserted into the object file! Can anyone with a gcc 3 handy see if this behavior persists? Thanks, -- Jerry James http://www.ittc.ku.edu/~james/

Steve Youngs <youngs(a)xemacs.org&gt; wrote: Oh, right. Well, doesn't it look like a compiler bug? The GCC info manual says (under "C Extensions -> Inline"): ------------------------------------------------------------------------ If you specify both `inline' and `extern' in the function definition, then the definition is used only for inlining. In no case is the function compiled on its own, not even if you refer to its address explicitly. Such an address becomes an external reference, as if you had only declared the function, and had not defined it. ------------------------------------------------------------------------ That's what we did, and GCC inserted an undefined symbol with the function name into the object file. That seems wrong. It's especially suspicious that the problem goes away when the optimizer is turned on. I'll pop over to the GCC lists tomorrow and see if this has been discussed. -- Jerry James http://www.ittc.ku.edu/~james/

|--==> "JJ" == Jerry James <james(a)xemacs.org&gt; writes: JJ> Using any optimization setting at all causes it to link correctly. So JJ> now the question becomes whether we should try to work around this bug JJ> or not, and if so, how. Is the optimisation only needed with '--error-checking=all' (on gcc)? What consequences will arise from unconditionally turning on say '-O1' if '--error-checking=all'? Would backtraces be affected detrimentally? If nothing's likely to blow up in our face I'd suggest if anyone uses '--error-checking=all' then add '-O1' to the new cflags_warnings. That way it won't be overridden by someone's $CFLAGS if they have it set. What do you think? In any event, it would probably be a good idea to add an entry to PROBLEMS. -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|

|--==> "JJ" == Jerry James <james(a)xemacs.org&gt; writes: JJ> Steve Youngs <youngs(a)xemacs.org&gt; wrote: JJ> I hate to force a bug workaround on everybody, whether they are using a JJ> system with the bug or not. I kind of think we should leave things the JJ> way they are and ... Good point. It's not like this bug is going to bite a huge percentage of our user base. JJ> ... do this. "If you get an undefined symbol error when loading a JJ> module, then recompile with -O1 to work around a gcc bug." Also, note JJ> that compiling the module code with -O1 is sufficient; you don't need to JJ> compile all of XEmacs that way. JJ> If this is acceptable, I'll submit a patch against PROBLEMS. Yep, I'm happy with that. Thanks very much for tracking this down, Jerry! -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|