Fix resolving of relative script paths

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63

[jansorg]

At first, the script's working directory is used to resolve a relative path. Then, the script's parent directory is used. This commit removes duplicated code shared between file.sh and filecache.sh.

• I've switched the lookup order, _Dbg_dir=('\$cdir' '\$cwd' ) to _Dbg_dir=('\$cwd' '\$cdir' ), because IMHO the current working directory should be looked at first. At runtime, that's what the debuggee does (I suppose)
• I've changed typeset _Dbg_cdir="${_Dbg_source_file%/*}" to typeset _Dbg_cdir="${_Dbg_init_cwd%/*}" because that seems more suitable to reference the script's location. In dbg-main.sh variable _Dbg_source_file is used in a loop (without a break) and does not seem to refer to the debuggee script passed to bashdb. Because of the missing break, I think it's always set to $DBG_RESTART_FILE when the loop is done.

Supersedes #61

Update: This is breaking other cases, discussed in #63