diff options
Diffstat (limited to 'apt-inst')
| -rw-r--r-- | apt-inst/contrib/arfile.cc | 2 | ||||
| -rw-r--r-- | apt-inst/contrib/extracttar.cc | 2 | ||||
| -rw-r--r-- | apt-inst/extract.cc | 4 | ||||
| -rw-r--r-- | apt-inst/filelist.cc | 4 |
4 files changed, 6 insertions, 6 deletions
diff --git a/apt-inst/contrib/arfile.cc b/apt-inst/contrib/arfile.cc index 9d84c1784..77dbc55d6 100644 --- a/apt-inst/contrib/arfile.cc +++ b/apt-inst/contrib/arfile.cc @@ -6,7 +6,7 @@ AR File - Handle an 'AR' archive AR Archives have plain text headers at the start of each file - section. The headers are aligned on a 2 byte boundry. + section. The headers are aligned on a 2 byte boundary. Information about the structure of AR files can be found in ar(5) on a BSD system, or in the binutils source. diff --git a/apt-inst/contrib/extracttar.cc b/apt-inst/contrib/extracttar.cc index fb4db42f8..2437c9749 100644 --- a/apt-inst/contrib/extracttar.cc +++ b/apt-inst/contrib/extracttar.cc @@ -6,7 +6,7 @@ Extract a Tar - Tar Extractor Some performance measurements showed that zlib performed quite poorly - in comparision to a forked gzip process. This tar extractor makes use + in comparison to a forked gzip process. This tar extractor makes use of the fact that dup'd file descriptors have the same seek pointer and that gzip will not read past the end of a compressed stream, even if there is more data. We use the dup property to track extraction diff --git a/apt-inst/extract.cc b/apt-inst/extract.cc index 2c95fba92..b3dfccfc6 100644 --- a/apt-inst/extract.cc +++ b/apt-inst/extract.cc @@ -10,7 +10,7 @@ object is unpacked to '.dpkg.new' then the original is hardlinked to '.dpkg.tmp' and finally the new object is renamed to overwrite the old one. From an external perspective the file never ceased to exist. - After the archive has been sucessfully unpacked the .dpkg.tmp files + After the archive has been successfully unpacked the .dpkg.tmp files are erased. A failure causes all the .dpkg.tmp files to be restored. Decisions about unpacking go like this: @@ -22,7 +22,7 @@ [Note, this is reduced to only check if a file was expected to be there] - If the existing link/file is not a directory then it is replaced - irregardless + regardless - If the existing link/directory is being replaced by a directory then absolutely nothing happens. - If the existing link/directory is being replaced by a link then diff --git a/apt-inst/filelist.cc b/apt-inst/filelist.cc index 879c07855..defc4f4df 100644 --- a/apt-inst/filelist.cc +++ b/apt-inst/filelist.cc @@ -5,14 +5,14 @@ File Listing - Manages a Cache of File -> Package names. - Diversions add some signficant complexity to the system. To keep + Diversions add some significant complexity to the system. To keep storage space down in the very special case of a diverted file no extra bytes are allocated in the Node structure. Instead a diversion is inserted directly into the hash table and its flag bit set. Every lookup for that filename will always return the diversion. The hash buckets are stored in sorted form, with diversions having - the higest sort order. Identical files are assigned the same file + the highest sort order. Identical files are assigned the same file pointer, thus after a search all of the nodes owning that file can be found by iterating down the bucket. |