X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=pod%2Fperldebguts.pod;h=cde70b2fb022f347950ce0d63899ab5d55251266;hb=1a6071381913ad8fca476a56b6dbb62e5c99c1bd;hp=45c33c7ec437133247cefef19c1c0ad8e6c4ede7;hpb=ee8c7f5465f003860e2347a2946abacac39bd9b9;p=p5sagit%2Fp5-mst-13.2.git diff --git a/pod/perldebguts.pod b/pod/perldebguts.pod index 45c33c7..cde70b2 100644 --- a/pod/perldebguts.pod +++ b/pod/perldebguts.pod @@ -5,9 +5,10 @@ perldebguts - Guts of Perl debugging =head1 DESCRIPTION This is not the perldebug(1) manpage, which tells you how to use -the debugger. This manpage describes low-level details ranging -between difficult and impossible for anyone who isn't incredibly -intimate with Perl's guts to understand. Caveat lector. +the debugger. This manpage describes low-level details concerning +the debugger's internals, which range from difficult to impossible +to understand for anyone who isn't incredibly intimate with Perl's guts. +Caveat lector. =head1 Debugger Internals @@ -18,12 +19,13 @@ usable only if a special Perl is built per the instructions in the F podpage in the Perl source tree. For example, whenever you call Perl's built-in C function -from the package DB, the arguments that the corresponding stack -frame was called with are copied to the @DB::args array. The -general mechanisms is enabled by calling Perl with the B<-d> switch, the -following additional features are enabled (cf. L): +from the package C, the arguments that the corresponding stack +frame was called with are copied to the C<@DB::args> array. These +mechanisms are enabled by calling Perl with the B<-d> switch. +Specifically, the following additional features are enabled +(cf. L): -=over +=over 4 =item * @@ -32,20 +34,22 @@ Perl inserts the contents of C<$ENV{PERL5DB}> (or C holds the lines of $filename for all -files compiled by Perl. The same for Ced strings that contain -subroutines, or which are currently being executed. The $filename -for Ced strings looks like C<(eval 34)>. Code assertions -in regexes look like C<(re_eval 19)>. +Each array C<@{"_<$filename"}> holds the lines of $filename for a +file compiled by Perl. The same is also true for Ced strings +that contain subroutines, or which are currently being executed. +The $filename for Ced strings looks like C<(eval 34)>. +Code assertions in regexes look like C<(re_eval 19)>. + +Values in this array are magical in numeric context: they compare +equal to zero only if the line is not breakable. =item * -The hash C<%{"_<$filename"}> contains breakpoints and actions keyed +Each hash C<%{"_<$filename"}> contains breakpoints and actions keyed by line number. Individual entries (as opposed to the whole hash) are settable. Perl only cares about Boolean true here, although the values used by F have the form -C<"$break_condition\0$action">. Values in this hash are magical -in numeric context: they are zeros if the line is not breakable. +C<"$break_condition\0$action">. The same holds for evaluated strings that contain subroutines, or which are currently being executed. The $filename for Ced strings @@ -53,7 +57,7 @@ looks like C<(eval 34)> or C<(re_eval 19)>. =item * -The scalar C<${"_<$filename"}> contains C<"_<$filename">. This is +Each scalar C<${"_<$filename"}> contains C<"_<$filename">. This is also the case for evaluated strings that contain subroutines, or which are currently being executed. The $filename for Ced strings looks like C<(eval 34)> or C<(re_eval 19)>. @@ -82,8 +86,8 @@ Cs, or C<(re_eval 19)> for those within regex code assertions. =item * When the execution of your program reaches a point that can hold a -breakpoint, the C subroutine is called any of the variables -$DB::trace, $DB::single, or $DB::signal is true. These variables +breakpoint, the C subroutine is called if any of the variables +C<$DB::trace>, C<$DB::single>, or C<$DB::signal> is true. These variables are not Cizable. This feature is disabled when executing inside C, including functions called from it unless C<< $^D & (1<<30) >> is true. @@ -92,37 +96,42 @@ unless C<< $^D & (1<<30) >> is true. When execution of the program reaches a subroutine call, a call to C<&DB::sub>(I) is made instead, with C<$DB::sub> holding the -name of the called subroutine. This doesn't happen if the subroutine +name of the called subroutine. (This doesn't happen if the subroutine was compiled in the C package.) =back Note that if C<&DB::sub> needs external data for it to work, no -subroutine call is possible until this is done. For the standard -debugger, the C<$DB::deep> variable (how many levels of recursion -deep into the debugger you can go before a mandatory break) gives -an example of such a dependency. +subroutine call is possible without it. As an example, the standard +debugger's C<&DB::sub> depends on the C<$DB::deep> variable +(it defines how many levels of recursion deep into the debugger you can go +before a mandatory break). If C<$DB::deep> is not defined, subroutine +calls are not possible, even though C<&DB::sub> exists. =head2 Writing Your Own Debugger -The minimal working debugger consists of one line +=head3 Environment Variables + +The C environment variable can be used to define a debugger. +For example, the minimal "working" debugger (it actually doesn't do anything) +consists of one line: sub DB::DB {} -which is quite handy as contents of C environment -variable: +It can easily be defined like this: $ PERL5DB="sub DB::DB {}" perl -d your-script -Another brief debugger, slightly more useful, could be created +Another brief debugger, slightly more useful, can be created with only the line: sub DB::DB {print ++$i; scalar } -This debugger would print the sequential number of encountered -statement, and would wait for you to hit a newline before continuing. +This debugger prints a number which increments for each statement +encountered and waits for you to hit a newline before continuing +to the next statement. -The following debugger is quite functional: +The following debugger is actually useful: { package DB; @@ -130,39 +139,79 @@ The following debugger is quite functional: sub sub {print ++$i, " $sub\n"; &$sub} } -It prints the sequential number of subroutine call and the name of the -called subroutine. Note that C<&DB::sub> should be compiled into the -package C. +It prints the sequence number of each subroutine call and the name of the +called subroutine. Note that C<&DB::sub> is being compiled into the +package C through the use of the C directive. -At the start, the debugger reads your rc file (F<./.perldb> or -F<~/.perldb> under Unix), which can set important options. This file may -define a subroutine C<&afterinit> to be executed after the debugger is -initialized. +When it starts, the debugger reads your rc file (F<./.perldb> or +F<~/.perldb> under Unix), which can set important options. +(A subroutine (C<&afterinit>) can be defined here as well; it is executed +after the debugger completes its own initialization.) After the rc file is read, the debugger reads the PERLDB_OPTS -environment variable and parses this as the remainder of a C -line as one might enter at the debugger prompt. +environment variable and uses it to set debugger options. The +contents of this variable are treated as if they were the argument +of an C debugger command (q.v. in L). + +=head3 Debugger internal variables +In addition to the file and subroutine-related variables mentioned above, +the debugger also maintains various magical internal variables. + +=over 4 + +=item * + +C<@DB::dbline> is an alias for C<@{"::_, which +holds the lines of the currently-selected file (compiled by Perl), either +explicitly chosen with the debugger's C command, or implicitly by flow +of execution. + +Values in this array are magical in numeric context: they compare +equal to zero only if the line is not breakable. -The debugger also maintains magical internal variables, such as -C<@DB::dbline>, C<%DB::dbline>, which are aliases for -C<@{"::_ C<%{"::_. Here C -is the currently selected file, either explicitly chosen with the +=item * + +C<%DB::dbline>, is an alias for C<%{"::_, which +contains breakpoints and actions keyed by line number in +the currently-selected file, either explicitly chosen with the debugger's C command, or implicitly by flow of execution. -Some functions are provided to simplify customization. See -L for description of options parsed by -C. The function C skips the specified number of frames and returns a list -containing information about the calling frames (all of them, if -C is missing). Each entry is reference to a hash with -keys C (either C<.>, C<$>, or C<@>), C (subroutine +As previously noted, individual entries (as opposed to the whole hash) +are settable. Perl only cares about Boolean true here, although +the values used by F have the form +C<"$break_condition\0$action">. + +=back + +=head3 Debugger customization functions + +Some functions are provided to simplify customization. + +=over 4 + +=item * + +See L for description of options parsed by +C parses debugger options; see +L for a description of options recognized. + +=item * + +C skips the specified number of frames +and returns a list containing information about the calling frames (all +of them, if C is missing). Each entry is reference to a hash +with keys C (either C<.>, C<$>, or C<@>), C (subroutine name, or info about C), C (C or a reference to an array), C, and C. -The function C prints +=item * + +C prints formatted info about caller frames. The last two functions may be convenient as arguments to C<< < >>, C<< << >> commands. +=back + Note that any variables and functions that are not documented in this manpages (or in L) are considered for internal use only, and as such are subject to change without notice. @@ -191,9 +240,9 @@ information. For example, contrast this expression trace: main::bar((eval 170):2): 42 -with this one, once the Cption C has been set: +with this one, once the Cption C has been set: - DB<4> O f=2 + DB<4> o f=2 frame = '2' DB<5> t print foo() * bar() 3: foo() * bar() @@ -362,45 +411,60 @@ compile time and run time. It is not lexically scoped. The debugging output at compile time looks like this: - compiling RE `[bc]d(ef*g)+h[ij]k$' - size 43 first at 1 - 1: ANYOF(11) - 11: EXACT (13) - 13: CURLYX {1,32767}(27) - 15: OPEN1(17) - 17: EXACT (19) - 19: STAR(22) - 20: EXACT (0) - 22: EXACT (24) - 24: CLOSE1(26) - 26: WHILEM(0) - 27: NOTHING(28) - 28: EXACT (30) - 30: ANYOF(40) - 40: EXACT (42) - 42: EOL(43) - 43: END(0) - anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating) - stclass `ANYOF' minlen 7 + Compiling REx `[bc]d(ef*g)+h[ij]k$' + size 45 Got 364 bytes for offset annotations. + first at 1 + rarest char g at 0 + rarest char d at 0 + 1: ANYOF[bc](12) + 12: EXACT (14) + 14: CURLYX[0] {1,32767}(28) + 16: OPEN1(18) + 18: EXACT (20) + 20: STAR(23) + 21: EXACT (0) + 23: EXACT (25) + 25: CLOSE1(27) + 27: WHILEM[1/1](0) + 28: NOTHING(29) + 29: EXACT (31) + 31: ANYOF[ij](42) + 42: EXACT (44) + 44: EOL(45) + 45: END(0) + anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating) + stclass `ANYOF[bc]' minlen 7 + Offsets: [45] + 1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1] + 0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0] + 11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0] + 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0] + Omitting $` $& $' support. The first line shows the pre-compiled form of the regex. The second shows the size of the compiled form (in arbitrary units, usually -4-byte words) and the label I of the first node that does a -match. +4-byte words) and the total number of bytes allocated for the +offset/length table, usually 4+C*8. The next line shows the +label I of the first node that does a match. + +The -The last line (split into two lines above) contains optimizer + anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating) + stclass `ANYOF[bc]' minlen 7 + +line (split into two lines above) contains optimizer information. In the example shown, the optimizer found that the match should contain a substring C at offset 1, plus substring C at some offset between 3 and infinity. Moreover, when checking for these substrings (to abandon impossible matches quickly), Perl will check for the substring C before checking for the substring C. The optimizer may also use the knowledge that the match starts (at the -C I) with a character class, and the match cannot be -shorter than 7 chars. +C I) with a character class, and no string +shorter than 7 characters can possibly match. -The fields of interest which may appear in the last line are +The fields of interest which may appear in this line are -=over +=over 4 =item C I C I @@ -426,7 +490,7 @@ Don't scan for the found substrings. =item C -Means that the optimizer info is all that the regular +Means that the optimizer information is all that the regular expression contains, and thus one does not need to enter the regex engine at all. @@ -457,12 +521,12 @@ being C, C, or C. See the table below. If a substring is known to match at end-of-line only, it may be followed by C<$>, as in C. -The optimizer-specific info is used to avoid entering (a slow) regex -engine on strings that will not definitely match. If C flag +The optimizer-specific information is used to avoid entering (a slow) regex +engine on strings that will not definitely match. If the C flag is set, a call to the regex engine may be avoided even when the optimizer found an appropriate place for the match. -The rest of the output contains the list of I of the compiled +Above the optimizer section is the list of I of the compiled form of the regex. Each line has format C< >I: I I (I) @@ -581,6 +645,36 @@ Here are the possible types, with short descriptions: # To simplify debugging output, we mark it as if it were a node OPTIMIZED off Placeholder for dump. +=for unprinted-credits +Next section M-J. Dominus (mjd-perl-patch+@plover.com) 20010421 + +Following the optimizer information is a dump of the offset/length +table, here split across several lines: + + Offsets: [45] + 1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1] + 0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0] + 11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0] + 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0] + +The first line here indicates that the offset/length table contains 45 +entries. Each entry is a pair of integers, denoted by C. +Entries are numbered starting with 1, so entry #1 here is C<1[4]> and +entry #12 is C<5[1]>. C<1[4]> indicates that the node labeled C<1:> +(the C<1: ANYOF[bc]>) begins at character position 1 in the +pre-compiled form of the regex, and has a length of 4 characters. +C<5[1]> in position 12 +indicates that the node labeled C<12:> +(the C<< 12: EXACT >>) begins at character position 5 in the +pre-compiled form of the regex, and has a length of 1 character. +C<12[1]> in position 14 +indicates that the node labeled C<14:> +(the C<< 14: CURLYX[0] {1,32767} >>) begins at character position 12 in the +pre-compiled form of the regex, and has a length of 1 character---that +is, it corresponds to the C<+> symbol in the precompiled regex. + +C<0[0]> items indicate that there is no corresponding node. + =head2 Run-time output First of all, when doing a match, one may get no run-time output even @@ -639,7 +733,7 @@ than 32 bytes (all these examples assume 32-bit architectures, the result are quite a bit worse on 64-bit architectures). If a variable is accessed in two of three different ways (which require an integer, a float, or a string), the memory footprint may increase yet another -20 bytes. A sloppy malloc(3) implementation can make inflate these +20 bytes. A sloppy malloc(3) implementation can inflate these numbers dramatically. On the opposite end of the scale, a declaration like @@ -686,12 +780,12 @@ the following example: Total sbrk(): 215040/47:145. Odd ends: pad+heads+chain+tail: 0+2192+0+6144. It is possible to ask for such a statistic at arbitrary points in -your execution using the mstats() function out of the standard +your execution using the mstat() function out of the standard Devel::Peek module. Here is some explanation of that format: -=over +=over 4 =item C @@ -727,7 +821,7 @@ were 4 12 24 48 80 With non-C perl, the buckets starting from C<128> have -a 4-byte overhead, and thus a 8192-long bucket may take up to +a 4-byte overhead, and thus an 8192-long bucket may take up to 8188-byte allocations. =item C @@ -838,7 +932,7 @@ per glob - for glob name, and glob stringification magic. Here are explanations for other Is above: -=over +=over 4 =item C<717> @@ -892,7 +986,7 @@ these categories. If warn() string starts with -=over +=over 4 =item C @@ -920,4 +1014,4 @@ L, L L, and -L. +L.