| … | |
… | |
| 98 | current length is less than the new value. This does not affect the |
98 | current length is less than the new value. This does not affect the |
| 99 | contents of the scalar, but is only useful to "pre-allocate" memory |
99 | contents of the scalar, but is only useful to "pre-allocate" memory |
| 100 | space if you know the scalar will grow. The return value is the |
100 | space if you know the scalar will grow. The return value is the |
| 101 | modified scalar (the scalar is modified in-place). |
101 | modified scalar (the scalar is modified in-place). |
| 102 | |
102 | |
| 103 | scalar = extend scalar, addlen |
103 | scalar = extend scalar, addlen=64 |
| 104 | Reserves enough space in the scalar so that addlen bytes can be |
104 | Reserves enough space in the scalar so that addlen bytes can be |
| 105 | appended without reallocating it. The actual contents of the scalar |
105 | appended without reallocating it. The actual contents of the scalar |
| 106 | will not be affected. The modified scalar will also be returned. |
106 | will not be affected. The modified scalar will also be returned. |
| 107 | |
107 | |
| 108 | This function is meant to make append workloads efficient - if you |
108 | This function is meant to make append workloads efficient - if you |
| … | |
… | |
| 112 | |
112 | |
| 113 | If you instead use "extend $scalar, length $shortstring", then |
113 | If you instead use "extend $scalar, length $shortstring", then |
| 114 | Convert::Scalar will use a "size to next power of two, roughly" |
114 | Convert::Scalar will use a "size to next power of two, roughly" |
| 115 | algorithm, so as the scalar grows, perl will have to resize and copy |
115 | algorithm, so as the scalar grows, perl will have to resize and copy |
| 116 | it less and less often. |
116 | it less and less often. |
|
|
117 | |
|
|
118 | nread = extend_read fh, scalar, addlen=64 |
|
|
119 | Calls "extend scalar, addlen" to ensure some space is available, |
|
|
120 | then do the equivalent of "sysread" to the end, to try to fill the |
|
|
121 | extra space. Returns how many bytes have been read, 0 on EOF or |
|
|
122 | undef> on eror, just like "sysread". |
|
|
123 | |
|
|
124 | This function is useful to implement many protocols where you read |
|
|
125 | some data, see if it is enough to decode, and if not, read some |
|
|
126 | more, where the naive or easy way of doing this would result in bad |
|
|
127 | performance. |
|
|
128 | |
|
|
129 | nread = read_all fh, scalar, length |
|
|
130 | Tries to read "length" bytes into "scalar". Unlike "read" or |
|
|
131 | "sysread", it will try to read more bytes if not all bytes could be |
|
|
132 | read in one go (this is often called "xread" in C). |
|
|
133 | |
|
|
134 | Returns the total nunmber of bytes read (normally "length", unless |
|
|
135 | an error or EOF occured), 0 on EOF and "undef" on errors. |
|
|
136 | |
|
|
137 | nwritten = write_all fh, scalar |
|
|
138 | Like "readall", but for writes - the equivalent of the "xwrite" |
|
|
139 | function often seen in C. |
| 117 | |
140 | |
| 118 | refcnt scalar[, newrefcnt] |
141 | refcnt scalar[, newrefcnt] |
| 119 | Returns the current reference count of the given scalar and |
142 | Returns the current reference count of the given scalar and |
| 120 | optionally sets it to the given reference count. |
143 | optionally sets it to the given reference count. |
| 121 | |
144 | |
