gts-cache
gts-caches are files that are kept in user cache directories to avoid re-executing previously executed commands. By using a gts-cache, gts(1) commands will not only skip the command specific computations but also parsing, which can provide significant performance boosts in certain scenarios. A cache file consists of a header which contains information for validating the content of the cache file, and a body which can be any form of data of any length.
A gts(1) command will first check for the availability of a cache. This is done by computing a SHA-1 hash value using the inputs given to the command. First, the primary input file is digested to produce an input sum. Other inputs to the command is combined into a list of key-value pairs along with the name of the command and the command version. This list is then serialized and digested to produce the data sum. The input sum and data sum are concatenated and digested to produce the output sum. This output sum will then be encoded as a hexadecimal string, and this value will be used as the cache filename.
If a cache file with the computed filename does not exist, a command will attempt to create a cache file unless caching is disabled or an output file is specified explicitly. This means that even if caching is enabled, the cache file will only be created if the command is writing to standard output. Furthermore, a cache file will be deleted if an output file is specified by the user. This is done to minimize duplicate data from existing within the system. While caches do provide temporal benefits, they do occupy disk space which is generally undesierable. Therefore, gts(1) commands will try to minimize the amount of cache existing within the system at a given moment. Once the cache file is created, the input sum, data sum and the hash value of the digested body designated the body sum will be written into the file comprising the header, followed by the body content.
If a cache file with the computed filename does exist, a command will attempt to open the file unless caching is disabled. Once open, the header is read to verify that the input sum and data sum produces an output sum whose hexadecimal encoding is identical to the filename. The body is then digested to also verify that the body content is intact. If either of this verification fails, the cache is immediately discarded and a new cache file is created. If both verifications pass, the body content is then written to the specified output stream. If the output is a file explicitly specified by the user, the cache file is removed.
gts(1), gts-cache(1), gts-cache-list(1), gts-cache-path(1), gts-cache-purge(1)