part). Archive handling was broken at least since the move from BSD ar/ranlib
to GNU binutils because of the different archive format. This rewrite fixes
this by making make to carry around the defines for all formats (it supports)
so it can support all of them independent of the actually used one. The
supported formats are: traditional BSD (this seems to come from V7 at least,
short names only and __.SYMDEF), BSD4.4 (long names with #1/ and __.SYMDEF)
and SysV (extra name table and //). The only format not supported are broken
traditional archives where the member names are truncated to 15 characters.
Errors in the archive are not ignored anymore, but cause make to stop with
an error message. The command line option -A causes these errors to become
non-fatal. This is almost compatible with previous usage except for the
error message printed in any case.
Use a type-safe intrusive list for the archive cache.
Reviewed by: Max Okumoto <okumoto@ucsd.edu> (without new error handling)
module. The only module accessing it (the current line number) was the
condition module, so pass the current line number as a function argument.
Centralize the pushing of new input sources into one function
ParsePushInput() and rename the function handling the popping from ParseEOF()
to ParsePopInput(). Make the entire thing a little bit clearer, by holding
the current input source in the top element of the stack instead of
using extra variables for this. Use a type-safe intrusive list for the
input stack.
characters so it is not safe to move around code from
before it to after it. This should fix problems with building the
documentation.
Patch: 7.170
Submitted by: Max Okumoto <okumoto@ucsd.edu>
A Path is now a TAILQ of PathElements each of which just points to
a reference counted directory. Rename all functions dealing with Paths
from the Dir_ prefix to a Path_ prefix.
the closing brace so it is unwise to keep a pointer to it. Make
the variable static to fix this.
Patch: 7.152
Submitted by: Max Okumoto <okumoto@ucsd.edu>
Remove unreachable code for VAR_NOSUBST - it was never set.
Replace redundant code with calls to VarGetPattern().
Patch: 7.143-7.145
Submitted by: Max Okumoto <okumoto@ucsd.edu>
Change the parsing of the C modifier flags so that specifying both
'1' and 'g' gives an error.
Patch: 7.141,7.142
Submitted by: Max Okumoto <okumoto@ucsd.edu>
is always a '$'. This is not always correct, for example for
conditionals: .if defined(foobar).
Clean up some comments.
Move common code out of if-statements.
Patch: 7.140
Submitted by: Max Okumoto <okumoto@ucsd.edu>
by the caller. Don't pass 'endc' - it can be computed from 'startc'.
Remove unneccessary temporary variables. Remove constant if-expressions
and remove code before call Fatal() - there is no point to cleanup before
aborting.
Patch: 7.134,7.135
Submitted by: Max Okumoto <okumoto@ucsd.edu>
occurence which actually holds always the same constant value.
Shorten the name pattern to patt.
Patch: 7.133
Submitted by: Max Okumoto <okumoto@ucsd.edu>
by putting them into struct VarParser or computing them (consumed).
Change the loop termination condition in VarParseLong from endc to \0.
Patch: 7.128-7.132
Submitted by: Max Okumoto <okumoto@ucsd.edu>
Reduce the number of arguments passed between these functions by
creating a special-purpose struct.
Patch: 7.120,7.121
Submitted by: Max Okumoto <okumoto@ucsd.edu>
would access memory before the beginning of the string to match (the
suffix match starts at the end of both the string and the suffix and
proceedes to the begin until either the start of the suffix is hit
or the character does not match). This could lead to a memcpy copying
into random memory. Fix this by checking the length of the string to
match too and replacing the Lst_Find calls with LST_FOREACH loops
(last part by me).
Submitted by: Matt Dillon <dillon@apollo.backplane.com> (in principle)
Rename result variable so common code becomes more visible.
Rename freePtr to freeResult to make clear what pointer must be freed.
Patch: 7.116, 7.116a
Submitted by: Max Okumoto <okumoto@ucsd.edu>
the loop. Add a comment why the 'consumed' variable is updated.
Rename lengthPtr to consumed.
Patch: 7.115
Submitted by: Max Okumoto <okumoto@ucsd.edu>
It turns out that some ports use the obscure feature of spreading
a dependency block across multiple include files. While this seems
bad style, allow it for now and call said function only at end of
all input to process the really last line of everything.
ParseRestModifier() and ParseRestEnd(): move advancement of ptr to remove
a confusing calculation.
VarParseLong(): cleanup calculation of consumed.
Patch: 7.114
Submitted by: Max Okumoto <okumoto@ucsd.edu>
actual variable lookup. Consistently rename lengthPtr to consumed.
Update a number of comments to match the code.
Patch: 7.111-113
Submitted by: Max Okumoto <okumoto@ucsd.edu>
patch differs from the previous one in that it calls the function
only when a real file hits EOF. The bodies of .for loops are also
handled as files, but for these we don't want to end a dependency block
on the 'EOF' as in:
foo:
do-this
.for ...
do-something
.endfor
do-more
variables and expands archive specifications, one that expands
wild cards and a driver that loops over the children list and
expands each child if necessary replacing it with it's expansions.
that is now used for both the 'M'/'N' branch and the 'S' branch of
the switch statement into a common scope.
Patch: 7.102-105
Submitted by: Max Okumoto <okumoto@ucsd.edu>
is terminated with a slash. Although we are not System V, ar has
been configured to put that slash in. This format allows filenames
with trailing spaces.
paranthesis or brace) into the loop and don't leak the buffer in this
case. Remove the check for Var_Parse returning NULL - it can't.
Patch: 7.92
Submitted by: Max Okumoto <okumoto@ucsd.edu>
Return always malloc()-ed strings from VarParseShort() to get
rid of warnings when returning string constants from a non-const char *
function.
Patch: 7.90
Submitted by: Max Okumoto <okumoto@ucsd.edu>
to VarExpand down into the branches of the if as well as cleanup code.
Eliminate code that is now obviously dead.
Patch: 7.83
Submitted by: Max Okumoto <okumoto@ucsd.edu>
an else clause. Move the assignment to the lengthPtr down to
just before the return statements.
Patch: 7.81
Submitted by: Max Okumoto <okumoto@ucsd.edu>
functions: one for the single letter variables, one for the others
and one that does the recursive expansion.
Patches: 7.68-7.79
Submitted by: Max Okumoto <okumoto@ucsd.edu>
after a return. Move assignments to {freePtr, dynamic, start} closer to the
return statements to clarify which variables are actually used for
communication between the losely coupled blocks of the code. Clear up
an if-expression to make common structures of the conditions clearer.
Use strchr instead of switch statements to check for a character beeing
a member of a set.
Patches: 7-62.2, 7-62.3, 7-64, 7-65.1, 7-65.2
Submitted by: Max Okumoto <okumoto@ucsd.edu>
to the Lst_ForEach function this macro reduces the number of function
calls per invocation by N + 1 (where N is the number of list elements)
and increases code locality thereby increasing readability and
(maybe) performance.
when looking into an already hashed archive, the code tried to use
the name shortened to the maximum length allowed for the archive.
Unfortunately it passed a buffer of junk to the hashing routine when
the name actually wasn't too long. Theoretically this could lead to
a false positive.
it is actually needed. This makes clear in which subblocks the variables
are not needed and which can easier be split out.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
to prepare for function splitting and slightly reorganise the code
in anticipation of Var_Subst returning a Buffer.
Submitted by: Max Okumoto <okumoto@ucsd.edu> (with slight changes)
to be split out into a function soon). Also there is no need to write
back the colon that we have NUL-ed - the string is going to be freed
anyway.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
use a more consistent style with regard to *str and str[0];
simplify code by introducing a temporary variable;
shift a break around and add braces where appropriate.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
Move some assignments nearer to where they actually used. Convert a loop
from a for() to a while() to make it clearer and add braces to the long
body of it. Split assignment from variable declaration.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
parantheses. This helps editors to find its way through the horrible
mess of Var_Parse. Rewrite a for() loop into a while() to make it clearer.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
- convert Buf_AddByte from a macro to a function
- move #define's into the header file
- remove unused field in struct Buffer
- remove size fields - they can be easily computed
- inline Buf_OvAddByte
Submitted by: Max Okumoto <okumoto@ucsd.edu>
was non-NULL. This let's us eliminated an otherwise unused variable.
shellneeded can never return -1 so there is no need to check for it and
hence no need for a variable to hold the returned value.
Submitted by: Max Okumoto <okumoto@ucsd.edu> (partly)
itself. This will ease constification (think of what 'const Ptr foo'
means if Ptr is a pointer to a struct).
Submitted by: Max Okumoto <okumoto@ucsd.edu>
double documentation comments. Remove the 3rd clause (from 4) of the
BSD license because these files have only the UCB copyright.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
that get included just where they are needed. All headers include the
headers that they need to compile (just with an empty .c file). Sort
includes alphabetically where apropriate and fix some duplicate commenting
for struct Job, struct GNode and struct Shell by removing one version and
inlining the comments into the structure declaration (the comments have been
somewhat outdated).
This patch does not contain functional changes (checked with md5).
Submitted by: Max Okumoto <okumoto@ucsd.edu>
and tabs. This is still not correct for command line variable values
ending in a backslash because this would require a larger effort.
Document this limitation in the BUGS section of the man page. The
quoting is mostly compatible with that of gmake and smake.
Tested by: Max Okumoto and Joerg Sonnenberger from DragonFly BSD
Reviewed by: ru (man page, partly)
now that their size is only two pointers. This eliminates a lot of calls
to Lst_Init and from there to malloc together with many calls to
Lst_Destroy (in places where the list is obviously empty). This also
reduces the chance to leave a list uninitilized so we can remove more
NULL pointer checks and probably eliminates a couple of memory leaks.
only in a couple of places and all of them except for one were easily
converted to use Lst_First/Lst_Succ. The one place is compatibility
mode in job.c where the it was used to advance to the next command on
each invocation of JobStart. For this case add a pointer to the node to
hold the currently executed command.
requires to make a copy of the filename in ReadMakefile and to duplicate
two small functions in suff.c. This hopefully will go away when everything
is constified.
Submitted by: Max Okumoto <okumoto@ucsd.edu> (partly)
purpose and explicitely free the input string in the one place that was
calling str_concat with that flag.
Submitted by: Max Okumoto <okumoto@ucsd.edu>
of the .SHELL target. Formerly it used to select the shell with the
longest common trailing substring, so that bash would select sh, but pocsh
would select csh. Now an exact match is required so that specifying bash
without also giving a path and the other keywords will give an error.
PR:
Submitted by:
Reviewed by: ru
Approved by:
Obtained from:
MFC after:
and the sizeof operator, missing empty lines, void casts, extra empty lines.
Checked by: diff on make *.o lst.lib/*.o
Submitted by: Max Okumoto <okumoto@soe.ucsd.edu> (partly)
versions of the structure definition for this: one with const char
pointers, because otherwise gcc won't let us initialize the fields with
constant strings, and one without the const, because we need to work
with the structure.
the semantic of Lst_Datum which formerly returned NULL when the argument
node was NULL. There was only one place in the source that relied on this
so change that place.
rid of a lot of uneccesary casts and temporary variables that have just
obfuscated the code. This also let's us implement a couple of the one-
liner list functions as macros (the first one is Lst_IsEmpty) and
simplify life once we start to throw consts on the code.
uses the brk_string function to parse the line. That function uses static
storage for both the expanded string and the returned argv[] vector.
The JobParseShell function simply stored away pointers into this static
storage. On the next use of something like ${FOO:O} this storage would
get overwritten with fatal results.
This also allows us to make the shells[] array const bringing us one step
further in making make WARNS=4 ready.
of submakes spawned during processing.
We create a fifo and stuff one character into it for each job we are
allowed to run. The name of the fifo is passed to child processes
in the MAKE_JOBS_FIFO environment variable.
A make which finds this variable on startup will open the fifo and
only spawn jobs when it managed to read a token from the fifo.
When the job completes a token is writen back to the fifo.
Slave make processes get one token for free: the one their parent
make got in order to run them. This makes the make processes
themselves invisible in the process counts.
The net effect is that "make -j 12 -s buildworld" will start at
most 12 jobs at the same time, instead of as previously up to
65 jobs would get started.
from the beginning). Make used to handle all its interrupt-time stuff
directly from the signal handler, including calls to printf, accessing
global data and so on. This is of course wrong and could provoke a core
dump when interrupting make. Just set a flag in the signal handler and
do everything else from the main thread.
PR: bin/29103
