/*
* Copyright 1995, 1996 Perforce Software. All rights reserved.
*
* This file is part of Perforce - the FAST SCM System.
*/
/*
* strings.cc - support for StrBuf, StrPtr
*/
# define NEED_QUOTE
# include <stdhdrs.h>
# include <charman.h>
# include <charset.h>
# include <debug.h>
# include "strbuf.h"
# include "strdict.h"
# include "strops.h"
/*
* StrOps::Words() - break a string apart at white space
*
* Quotes are handles as thus:
*
* "foo bar" -> (foo) (bar)
* foo" "bar -> (foo) (bar)
* "foo""bar" -> (foo"bar)
* foo""bar -> (foo"bar)
*
* Note that the only user of Words() that requires quote handling
* is RunCommand::RunChild() on UNIX, which is used for P4PORT=rsh:
*/
int
StrOps::Words( StrBuf &tmp, const char *buf, char *vec[], int maxVec )
{
// Ensure tmp clear and big enough to avoid realloc
tmp.Clear();
tmp.Alloc( strlen( buf ) + 1 );
tmp.Clear();
int count = 0;
while( count < maxVec )
{
// Skip blanks
while( isAspace( buf ) ) buf++;
if( !*buf ) break;
// First letter of new word
vec[ count++ ] = tmp.End();
// Eat word
int quote = 0;
for( ; *buf; ++buf )
{
if( buf[0] == '"' && buf[1] == '"' )
tmp.Extend( buf[0] ), ++buf;
else if( buf[0] == '"' )
quote = !quote;
else if( quote || !isAspace( buf ) )
tmp.Extend( buf[0] );
else break;
}
tmp.Extend( '\0' );
}
return count;
}
/*
* StrPtr::Lines() - break a string apart at line endings.
* Supports '\r', '\n', and '\r\n' termination.
*/
int
StrOps::Lines( StrBuf &o, char *vec[], int maxVec )
{
int count = 0;
int lastwas_cr = 0;
char *buf = o.Text();
while ( count < maxVec )
{
if ( *buf )
vec[ count++ ] = buf;
else
break;
while ( *buf )
{
if ( *buf == '\r' )
lastwas_cr =1;
else if ( *buf == '\n' && lastwas_cr )
{
*(buf-1) = '\0';
*buf = '\0';
buf++;
lastwas_cr = 0;
break;
}
else if ( *buf == '\n' )
{
*buf = '\0';
buf++;
break;
}
else if ( lastwas_cr )
{
// '\r' line termination only
*(buf-1) = '\0';
lastwas_cr = 0;
break;
}
buf++;
}
if ( lastwas_cr )
{
// Final \r of a \r line terminated buffer
*(buf-1) = '\0';
}
}
return count;
}
/*
* StrPtr::Dump() - print out a string for debugging purposes
*/
void
StrOps::Dump( const StrPtr &o )
{
unsigned char *s = o.UText();
unsigned char *e = o.UEnd();
for( ; s < e; s++ )
{
if( isprint( *s ) )
p4debug.printf( "%c", *s );
else
p4debug.printf( "<%02x>", *s );
}
p4debug.printf( "\n" );
}
/*
* StrPtr::Lower() - lowercase each character (in place) in a string
*/
void
StrOps::Lower( StrBuf &o )
{
char *p = o.Text();
int l = o.Length();
for( ; l--; p++ )
*p = tolowerq( *p );
}
/*
* StrPtr::Upper() - uppercase each character (in place) in a string
*/
void
StrOps::Upper( StrBuf &o )
{
char *p = o.Text();
int l = o.Length();
for( ; l--; p++ )
*p = toupperq( *p );
}
/*
* StrPtr::Caps() - uppercase first character (in place) in a string
*/
void
StrOps::Caps( StrBuf &o )
{
char *p = o.Text();
if( o.Length() && !isAhighchar( p ) && islower( *p ) )
*p = toAupper( p );
}
/*
* StrBuf::Ident() - copy in a string, indenting a tabstop
*/
void
StrOps::Indent( StrBuf &o, const StrPtr &buf )
{
const char *p = buf.Text();
const char *t;
while( *p )
{
o.Append( "\t", 1 );
if( t = strchr( p, '\n' ) )
{
o.Append( p, t - p + 1 );
p = t + 1;
}
else
{
// just in case - all lines end with \n
o.Append( p );
o.Append( "\n", 1 );
p += strlen( p );
}
}
}
/*
* StrBuf::Sub() - replace one character with another
*/
void
StrOps::Sub( StrPtr &string, char target, char replacement )
{
for ( char * temp = string.Text();
*temp != '\0';
temp++ )
{
if ( *temp == target )
*temp = replacement;
}
}
/*
* StrOps::Replace() - Replace substrings. Replaces all occurrences of
* string s in string i with string r and writes the results to o.
*/
void
StrOps::Replace( StrBuf &o, const StrPtr &i, const StrPtr &s, const StrPtr &r )
{
char *start, *end;
o.Clear();
start = i.Text();
while ( end = strstr( start, s.Text() ) )
{
o.Append( start, end - start );
o.Append( r.Text() );
start += end - start + s.Length();
}
if ( *start )
o.Append( start );
}
/*
* StrBuf::Expand() - expand a string doing %var%, %x substitutions
*/
void
StrOps::Expand( StrBuf &o, const StrPtr &buf, StrDict &dict, StrDict *u )
{
StrPtr *val;
const char *p = buf.Text();
const char *q;
while( q = strchr( p, '%' ) )
{
// look for closing %
o.Append( p, q++ - p );
if( !( p = strchr( q, '%' ) ) )
{
// handle %junk
p = q;
break;
}
else if( p == q )
{
// handle %%
o.Extend( '%' );
}
else
{
// handle %var%
StrBuf var;
var.Extend( q, p - q );
var.Terminate();
if( val = dict.GetVar( var ) )
o.Append( val );
else
{
o << "%" << var << "%";
if( u )
u->SetVar( var.Text() );
}
}
++p;
}
o.Append( p );
}
/*
* StrOps::Expand2() - expand a string doing [%var%|opt] substitutions
*/
void
StrOps::Expand2( StrBuf &o, const StrPtr &buf, StrDict &dict )
{
StrPtr *val;
const char *p = buf.Text();
const char *q, *r, *s, *t;
// Handle sequences of
// text %var% ...
// text [ stuff1 %var% stuff2 ] ...
while( q = strchr( p, '%' ) )
{
// variables: (p)text (r)[ stuff (q)%var(s)% stuff2 (t)]
if( !( s = strchr( q + 1, '%' ) ) )
{
// %junk
break;
}
else if( s == q + 1 )
{
// %% - [ %% ] not handled!
o.Append( p, s - p );
p = s + 1;
continue;
}
// Pick out var name and look up value
StrVarName var( q + 1, s - q - 1 );
val = dict.GetVar( var );
// Now handle %var% or [ %var% | alt ]
if( !( r = (char*)memchr( p, '[', q - p ) ) )
{
// %var%
o.Append( p, q - p );
if( val ) o.Append( val );
p = s + 1;
}
else if( !( t = strchr( s + 1, ']' ) ) )
{
// [ junk
break;
}
else
{
// [ stuff1 %var% stuff2 | alternate ]
o.Append( p, r - p );
// [ | alternate ]
const char *v = (char *)memchr( s, '|', t - s );
if( !v ) v = t;
if( val && val->Length() )
{
// stuff1, val, stuff2
o.Append( r + 1, q - r - 1 );
o.Append( val );
o.Append( s + 1, v - s - 1 );
}
else if( v < t )
{
// alternate
o.Append( v + 1, t - v - 1 );
}
p = t + 1;
}
}
o.Append( p );
}
/*
* StrBuf::OtoX() - turn an octet stream into hex
*/
void
StrOps::OtoX( const StrPtr &octet, StrBuf &hex )
{
OtoX( (const unsigned char *)octet.Text(), octet.Length(), hex );
}
void
StrOps::OtoX( const unsigned char *octet, int octetLen, StrBuf &hex )
{
char *b = hex.Alloc( 2 * octetLen );
for( int i = 0; i < octetLen; i++ )
{
*b++ = OtoX( ( octet[i] >> 4 ) & 0x0f );
*b++ = OtoX( ( octet[i] >> 0 ) & 0x0f );
}
hex.Terminate();
}
/*
* StrOps::XtoO() - turn hex into an octet stream
*
* No error checking.
*/
void
StrOps::XtoO( const StrPtr &hex, StrBuf &octet )
{
int len = hex.Length() / 2;
XtoO( hex.Text(), (unsigned char *)octet.Alloc( len ), len );
octet.Terminate();
}
void
StrOps::XtoO( char *hex, unsigned char *octet, int octetLen )
{
for( ; octetLen--; hex += 2 )
{
*octet++
= ( XtoO( hex[0] ) << 4 )
| ( XtoO( hex[1] ) << 0 );
}
}
static const char *valid = "0123456789abcdefABCDEF";
static int
IsX( char p )
{
for( int i = 0; i < 22; i++ )
if( valid[i] == p )
return 1;
return 0;
}
int
StrOps::IsDigest( const StrPtr &hex )
{
if( hex.Length() != 32 )
return 0;
for( int i = 0; i < 32; i++ )
if( !IsX( hex.Text()[i] ) )
return 0;
return 1;
}
/*
* StrOps::WildToStr() - turn wildcards into %x escaped string
*
* No error checking.
*/
void
StrOps::WildToStr( const StrPtr &i, StrBuf &o )
{
// format special characters '@#*%'
WildToStr( i, o, "@#%*" );
}
void
StrOps::WildToStr( const StrPtr &i, StrBuf &o, const char *t )
{
o.Clear();
char *p = i.Text();
const char *f;
char *s;
while( *p )
{
s = p;
while( *p )
{
f = t;
while( *f && *f != *p )
++f;
if( *f )
break;
++p;
}
o.Append( s, p - s );
if( *p )
{
char buf[3];
buf[0] = '%';
buf[1] = OtoX( ( *p >> 4 ) & 0x0f );
buf[2] = OtoX( ( *p++ >> 0 ) & 0x0f );
o.Append( buf, 3 );
}
}
}
void
StrOps::MaskNonPrintable( const StrPtr &i, StrBuf &o )
{
o.Clear();
o.Alloc( i.Length() + 1 );
o.Clear();
unsigned char *s = i.UText();
unsigned char *e = i.UEnd();
for( ; s < e; s++ )
{
if( isAprint( s ) )
o.Extend( *s );
else
o.Extend( '_' );
}
o.Terminate();
}
/*
* StrOps::StrToWild() - turn %x escaped string into wildcards.
*
* No error checking.
*/
void
StrOps::StrToWild( const StrPtr &i, StrBuf &o )
{
// expand %x character back to '@#*%'
o.Clear();
char *p = i.Text();
char *s;
while( *p )
{
s = p;
while( *p )
{
if( p[0] == '%' && p[1] == '%' )
p += 2;
else if( p[0] != '%' )
p++;
else
break;
}
o.Append( s, p - s );
if( *p && ( p + 2 < i.End() ) )
{
char b[2];
b[0] = ( XtoO( p[1] ) << 4 ) | ( XtoO( p[2] ) << 0 );
// Only translate the @#%* wildcards
if( b[0] == '@' || b[0] == '#' || b[0] == '%' || b[0] == '*' )
o.Append( b, 1 );
else
o.Append( p, 3 );
if( *(p+2) == '\0')
break;
p += 3;
}
else if( *p )
{
o.Append( p++, 1 );
}
else
{
break;
}
}
}
/*
* StrOps::WildCompat() - turn %%d into %d for 'p4 where' compatability.
*
* No error checking.
*/
void
StrOps::WildCompat( const StrPtr &i, StrBuf &o )
{
// turn %%d into %d
o.Clear();
char *p = i.Text();
char *s;
while( *p )
{
s = p;
while( *p )
{
if( p[0] == '%' && p[1] == '%' && p[2] >= '0' && p[2] <= '9' )
break;
++p;
}
o.Append( s, p - s );
if( *p )
{
o.Append( ++p, 2 );
p += 2;
}
}
}
/*
* StrBuf::PackInt() - marshalling function
* StrBuf::PackChar() - marshalling function
* StrBuf::PackOctet() - marshalling function
* StrBuf::PackString() - marshalling function
* StrPtr::UnpackInt() - marshalling function
* StrPtr::UnpackChar() - marshalling function
* StrBuf::UnpackOctet() - marshalling function
* StrPtr::UnpackString() - marshalling function
*/
void
StrOps::PackInt( StrBuf &o, int v )
{
char *b = o.Alloc( 4 );
const unsigned int vv = (unsigned int)v;
b[0] = ( vv / 0x1 ) % 0x100;
b[1] = ( vv / 0x100 ) % 0x100;
b[2] = ( vv / 0x10000 ) % 0x100;
b[3] = ( vv / 0x1000000 ) % 0x100;
}
void
StrOps::PackInt64( StrBuf &o, P4INT64 v )
{
char *b = o.Alloc( 8 );
unsigned P4INT64 vv = (unsigned P4INT64)v;
b[0] = ( vv / 0x1 ) % 0x100;
b[1] = ( vv / 0x100 ) % 0x100;
b[2] = ( vv / 0x10000 ) % 0x100;
b[3] = ( vv / 0x1000000 ) % 0x100;
// On machines Without int64's, two >>16
// will avoid complaints and zero v.
vv >>= 16;
vv >>= 16;
b[4] = ( vv / 0x1 ) % 0x100;
b[5] = ( vv / 0x100 ) % 0x100;
b[6] = ( vv / 0x10000 ) % 0x100;
b[7] = ( vv / 0x1000000 ) % 0x100;
}
void
StrOps::PackIntA( StrBuf &o, int v )
{
o << v;
o.Extend(0);
}
void
StrOps::PackChar( StrBuf &o, const char *c, int length )
{
char *end;
// Append up to null, or whole thing if no null
if( end = (char *)memchr( c, 0, length ) )
length = end + 1 - c;
o.Append( c, length );
}
void
StrOps::PackOctet( StrBuf &o, const StrPtr &s )
{
o.Append( &s );
}
void
StrOps::PackString( StrBuf &o, const StrPtr &s )
{
PackInt( o, s.Length() );
o.Append( &s );
}
void
StrOps::PackStringA( StrBuf &o, const StrPtr &s )
{
PackIntA( o, s.Length() );
o.Append( &s );
}
int
StrOps::UnpackInt( StrRef &o )
{
if( o.Length() < 4 )
return 0;
const char *b = o.Text();
o += 4;
return
(unsigned char)b[0] * (unsigned int)0x1 +
(unsigned char)b[1] * (unsigned int)0x100 +
(unsigned char)b[2] * (unsigned int)0x10000 +
(unsigned char)b[3] * (unsigned int)0x1000000;
}
P4INT64
StrOps::UnpackInt64( StrRef &o )
{
if( o.Length() < 8 )
return 0;
const char *b = o.Text();
o += 8;
P4INT64 v =
(unsigned char)b[4] * (unsigned int)0x1 +
(unsigned char)b[5] * (unsigned int)0x100 +
(unsigned char)b[6] * (unsigned int)0x10000 +
(unsigned char)b[7] * (unsigned int)0x1000000;
// On machines without int64, this will just overflow
// and zero v, without complaint from the compiler.
v <<= 16; v <<= 16;
return v +
(unsigned char)b[0] * (unsigned int)0x1 +
(unsigned char)b[1] * (unsigned int)0x100 +
(unsigned char)b[2] * (unsigned int)0x10000 +
(unsigned char)b[3] * (unsigned int)0x1000000;
}
int
StrOps::UnpackIntA( StrRef &o )
{
char *buffer = o.Text();
int length = o.Length();
int v = 0;
int s = length && *buffer == '-';
if( s )
++buffer, --length;
while( length && *buffer )
{
v = v * 10 + *buffer - '0';
++buffer, --length;
}
if( length )
++buffer, --length;
o.Set( buffer, length );
return s ? -v : v;
}
void
StrOps::UnpackChar( StrRef &o, char *c, int l )
{
if( l > o.Length() )
l = o.Length();
char *end;
// Unpack whole thing
if( end = (char *)memccpy( c, o.Text(), 0, l ) )
l = end - c;
o += l;
}
void
StrOps::UnpackOctet( StrRef &o, const StrPtr &s )
{
int l = s.Length();
if( l > o.Length() )
l = o.Length();
memcpy( s.Text(), o.Text(), l );
o += l;
}
void
StrOps::UnpackString( StrRef &o, StrBuf &s )
{
int l = UnpackInt( o );
if( l > o.Length() )
l = o.Length();
s.Set( o.Text(), l );
o += l;
}
void
StrOps::UnpackStringA( StrRef &o, StrBuf &s )
{
int l = UnpackIntA( o );
if( l > o.Length() )
l = o.Length();
s.Set( o.Text(), l );
o += l;
}
void
StrOps::UnpackString( StrRef &o, StrRef &s )
{
int l = UnpackInt( o );
if( l > o.Length() )
l = o.Length();
s.Set( o.Text(), l );
o += l;
}
void
StrOps::UnpackStringA( StrRef &o, StrRef &s )
{
int l = UnpackIntA( o );
if( l > o.Length() )
l = o.Length();
s.Set( o.Text(), l );
o += l;
}
/*
* i18n
*/
int
StrOps::CharCnt( const StrPtr &s )
{
int cs = GlobalCharSet::Get();
if( !cs )
return s.Length();
CharStep *step = CharStep::Create( s.Text(), cs );
int ret = step->CountChars( s.End() );
delete step;
return ret;
}
void
StrOps::CharCopy( const StrPtr &s, StrBuf &t, int length )
{
int charSet;
if( s.Length() < length )
{
length = s.Length();
}
else if( s.Length() > length && ( charSet = GlobalCharSet::Get() ) )
{
// i18n -- copy 'length' characters (possibly many more bytes)
int i = 0;
CharStep *ss = CharStep::Create( s.Text(), charSet );
while( ss->Next() < s.End() && ++i < length )
;
length = ss->Ptr() - s.Text();
delete ss;
}
t.Set( s.Text(), length );
}
/*
* StrOps::ScrunchArgs() - try to display argv in a limited output buffer.
*
* This scrunches in two ways:
*
* 1. If any argument is too long (more than 1/4 the output if there
* are 4 or more args), the argument is clipped to be left...right.
*
* 2. If we run out of room even to display even clipped arguments,
* we just output (number-of-skipped-args) and then the last argument.
*/
void
StrOps::ScrunchArgs( StrBuf &out, int argc, StrPtr *argv, int targetLength,
int delim, const char *unsafeChars )
{
if( !argc )
return;
StrBuf dStr;
dStr.Extend( (char)delim );
dStr.Terminate();
// Each arg gets at most 1/4th the targetLength,
// unless there are fewer than 4 args.
int pieces = argc < 4 ? argc : 4;
int eachSpace = targetLength / pieces;
// Leave room for last argument -- we like to see it.
int endPost = CharCnt(out) + targetLength;
int lastargLen = CharCnt(argv[argc-1]);
endPost -= lastargLen < eachSpace ? lastargLen : eachSpace;
// For each arg
for( ; argc--; ++argv )
{
StrBuf argBuf, maskBuf;
StrPtr *theArg = argv;
if( unsafeChars )
{
WildToStr( *argv, maskBuf, unsafeChars );
EncodeNonPrintable( maskBuf, argBuf );
theArg = &argBuf;
}
int mySpace = CharCnt(*theArg);
int myLength = mySpace;
// If there are more args coming, we can take only our
// alloted eachSpace, and we may have to skip out on
// printing anything if we're already against the endPost.
if( argc )
{
// Back off to 1/4 targetLength max
if( mySpace > eachSpace )
mySpace = eachSpace;
// Will no more args fit?
// Just mention # of args skipped.
// We'll still dump the last arg.
if( mySpace + CharCnt(out) > endPost )
{
out << "(" << argc - 1 << ")" << dStr;
argv += argc - 1;
argc = 1;
continue;
}
}
// If this arg is bigger than our allocated space,
// dump out left...right of arg.
// Otherwise, dump whole arg.
if( myLength > mySpace )
{
int side = ( mySpace - 3 ) / 2;
int cs = GlobalCharSet::Get();
if( !cs )
{
out << StrRef( theArg->Text(), side );
out << "...";
out << StrRef( theArg->End() - side, side );
}
else
{
CharStep *step = CharStep::Create( theArg->Text(), cs );
out << StrRef( theArg->Text(),
step->Next( side ) - theArg->Text() );
out << "...";
step->Next( myLength - 2 * side );
out << StrRef( step->Ptr(), theArg->End() - step->Ptr() );
delete step;
}
}
else
{
out << *theArg;
}
// blank between args
if( argc )
out << dStr;
}
}
/*
* StrOps::GetDepotName() - extracts the depot name from a depot path
*/
void
StrOps::GetDepotName( const char *d, StrBuf &n )
{
const char *s = strstr( d, "//" );
const char *p = d + 2;
if( !s || s != d )
return;
s = strstr( p, "/" );
if( !s )
return;
n.Append( p, s - p );
}
/*
* StrOps::CommonPath() - Build a common file path given multiple depotpaths.
*
* Common path result is passed in as the first argument, the second
* argument maintains a state that will indicate that paths are from multiple
* directories (will need ...) appended. The third argument is the next
* depotfile for consideration.
*
* e.g.
* StrBuf root;
* int mdir = 0;
*
* StrOps::CommonPath( root, mdir, depotPath1 );
* StrOps::CommonPath( root, mdir, depotPath2 );
* StrOps::CommonPath( root, mdir, depotPath3 );
* StrOps::CommonPath( root, mdir, depotPath4 );
* StrOps::CommonPath( root, mdir, depotPath5 );
*
* if( mdir )
* root.Append( "..." );
* else
* root.Append( "*" );
*/
void
StrOps::CommonPath( StrBuf &o, int &mdir, const StrPtr &n )
{
if( o.Length() )
{
char *op = o.Text();
char *np = n.Text();
while( op < o.End() && StrPtr::SEqual( *op, *np ) )
++op, ++np;
// check to see if multiple directories
if( !mdir && ( strchr( op, '/' ) || strchr( np, '/' ) ) )
mdir = 1;
if( mdir && op[-1] == '.' )
o.SetEnd( op - 1 );
else
o.SetEnd( op );
}
else
{
o = n;
char *s = o.Text();
char *e = o.End();
while( e > s && *e != '/')
--e;
o.SetEnd( ++e );
}
}
/*
* StrOps::StripNewline() - strip \r\n from end of buffer
*/
void
StrOps::StripNewline( StrBuf &o )
{
if( o.Length() && o.End()[ -1 ] == '\n' )
o.SetEnd( o.End() -1 );
if( o.Length() && o.End()[ -1 ] == '\r' )
o.SetEnd( o.End() -1 );
o.Terminate();
}
void
StrOps::EncodeNonPrintable( const StrPtr &i, StrBuf &o )
{
o.Clear();
char *p = i.Text();
const char *f;
char *s;
while( *p )
{
s = p;
while( *p )
{
if( !isAprint( p ) )
break;
++p;
}
o.Append( s, p - s );
if( *p )
{
char buf[3];
buf[0] = '%';
buf[1] = OtoX( ( *p >> 4 ) & 0x0f );
buf[2] = OtoX( ( *p++ >> 0 ) & 0x0f );
o.Append( buf, 3 );
}
}
}
void
StrOps::DecodeNonPrintable( const StrPtr &i, StrBuf &o )
{
o.Clear();
char *p = i.Text();
char *s;
while( *p )
{
s = p;
while( *p )
{
if( p[0] == '%' && p[1] == '%' )
p += 2;
else if( p[0] != '%' )
p++;
else
break;
}
o.Append( s, p - s );
if( *p )
{
char b[2];
b[0] = ( XtoO( p[1] ) << 4 ) | ( XtoO( p[2] ) << 0 );
o.Append( b, 1 );
p += 3;
}
}
}
void
StrOps::LFtoCRLF( const StrBuf *in, StrBuf *out )
{
out->Clear();
const char *s = in->Text(), *p = s;
while( ( p - s ) < in->Length() )
{
if( *p == '\n' )
out->Extend( '\r' );
out->Extend( *p );
p++;
}
out->Terminate();
}