我一直在研究一个脚本，该脚本从profiler.out文件中检索测试代码覆盖率数据。现在生成此文件已不再是问题。但是我似乎无法理解此文件的内容与触发的测试之间的关系（对我来说，它们看起来就像是随机数和字符串）...
有人知道这些数字代表什么吗？
我已经生成了一个示例探查器文件（源语言为Progress），如下所示：
谢谢！
Jacco
答案 0 :(得分：3)
OpenEdge Profiler被“轻松记录”。如果您搜索Progress kbase，它将显示几个条目，最终导致： https://community.progress.com/community_groups/openedge_general/w/openedgegeneral/1980.profiler-control-tool
这是旧的不受支持的独立“探查器控制工具”（该工具不受支持，探查器启动选项和会话句柄受支持）。
该页面链接到的zip文件中是探查器输出格式的唯一已知文档。
当前的OpenEdge 11 IDE（PDSOE）直接支持探查器分析工具。
以下代码不是很漂亮，但是它是一个独立的工具，可以读取探查器输出文件（我的首选是将它们命名为“ something.prf”而不是“ profiler.out”）并生成有关最耗时的代码行：
/* zprof_topx.p
*
* process a .prf file and produce a report
*
* pro -p lib/zprof_topx.p -param tmp/f03e40c9-8043-0aba-e611-3cc5eac467bd.2016.12.18.11.12.14
*
* note: there is no ".prf" at the end of the file name!
*
*/
define variable zprofData as character no-undo.
define variable topLines as integer no-undo initial 20.
define temp-table tt_profile
field id as integer format ">>>>9"
field pdate as date format "99/99/99"
field description as character format "x(30)"
index profile-idx is unique primary
id
index profile-date
pdate
.
define temp-table tt_source
field id as integer format ">>>>9"
field pid as integer format ">>>>>9"
field pname as character format "x(40)"
field debug_name as character format "x(40)"
index source-idx is unique primary
id pid
index source-name
pname
.
define temp-table tt_tree
field id as integer format ">>>>9"
field caller as integer format ">>>>>9"
field src_line as integer format ">>>>>9"
field callee as integer format ">>>>>9"
field call_count as integer format ">>>>>9"
index tree-idx is primary
id caller src_line callee
.
define temp-table tt_ptime
field id as integer format ">>>>9"
field pid as integer format ">>>>>9"
field src_line as integer format ">>>>>9"
field exec_count as integer format ">>>>>>>>>9"
field exe_time as decimal format ">>>>>9.999999"
field tot_time as decimal format ">>>>>9.999999"
field avg_time as decimal format ">>>>>9.999999"
index ptime-idx is unique primary
id pid src_line
index avg-idx
avg_time descending
index line-idx
src_line
index ptime-pid-t1
id pid exe_time
index ptime-pid-t3
id pid avg_time
index ptime-t1
id exe_time
index ptime-t3
id avg_time
.
define temp-table tt_code_line
field pid as integer format ">>>>>9" /* program id# */
field src_line as integer format ">>>>9" /* source line# */
field pname as character format "x(30)" /* procedure or class name */
field ipname as character format "x(40)" /* internal procedure or method name */
field t1 as decimal format ">>>>>9.999999" /* execution time */
field t2 as integer format ">>>>>>>>>9" /* calls */
field t3 as integer format ">>9" /* sessions */
field t4 as decimal format ">>>>>9.999999" /* average time */
index bad-idx1 is unique primary
pid pname src_line
index bad-idx2
t2 t3
index bad-idx3
t1
index avg-idx
t4
.
/*********************************************************/
define stream inStrm.
procedure zprofiler_load:
define input parameter zprofData as character no-undo.
define variable i as integer no-undo.
define variable v as integer no-undo.
define variable dt as date no-undo.
define variable dsc as character no-undo.
define variable profile_id as integer no-undo.
empty temp-table tt_profile.
empty temp-table tt_source.
empty temp-table tt_tree.
empty temp-table tt_ptime.
file-info:file-name = zprofData + ".prf".
if file-info:full-pathname = ? then
do:
message "Cannot find profiler .prf data file:" zprofData.
pause.
return.
end.
/* message "loading from:" file-info:full-pathname. /* session:date-format. */ pause. */
input stream inStrm from value( file-info:full-pathname ).
i = 1.
repeat: /* in theory there could be more than 1? that would probably break a lot of stuff... */
import stream inStrm v /* dt */ ^ dsc no-error. /* the profiler apparently ignores session:date-format... */
if v <> 1 then
do:
input stream inStrm close.
message "Invalid version:" v.
pause.
return.
end.
/* message v dt dsc. pause. */ /* the profiler apparently ignores session:date-format... */
profile_id = i.
create tt_profile.
assign
tt_profile.id = profile_id
tt_profile.pdate = today /* dt */
tt_profile.description = dsc
.
i = i + 1.
end.
/* message "profile id:" profile_id. pause. */
i = 1.
repeat:
create tt_source.
tt_source.id = profile_id.
import stream inStrm tt_source.pid tt_source.pname tt_source.debug_name no-error.
i = i + 1.
end.
/* message i "tt_source loaded". pause.
* message "creating tt_source session record". pause.
*/
/* create tt_source. */ /* don't CREATE -- an extra will be left over from the REPEAT logic */
assign
tt_source.id = profile_id
tt_source.pid = 0
tt_source.pname = "Session"
tt_source.debug_name = "Session"
.
/* message "tt_source session record created". pause. */
i = 1.
repeat:
create tt_tree.
tt_tree.id = profile_id.
import stream inStrm tt_tree.caller tt_tree.src_line tt_tree.callee tt_tree.call_count no-error.
i = i + 1.
end.
delete tt_tree.
/* message i "tt_tree loaded". pause. */
i = 1.
repeat:
create tt_ptime.
tt_ptime.id = profile_id.
import stream inStrm tt_ptime.pid tt_ptime.src_line tt_ptime.exec_count tt_ptime.exe_time tt_ptime.tot_time no-error.
tt_ptime.avg_time = tt_ptime.exe_time / tt_ptime.exec_count.
i = i + 1.
end.
delete tt_ptime.
/* message i "tt_ptime loaded". pause. */
input stream inStrm close.
return.
end.
procedure zprofiler_proc:
define variable c as integer no-undo.
define variable i as integer no-undo.
define variable t1 as decimal no-undo format ">>>>>9.999999".
define variable t2 as integer no-undo format ">>>>>>>>>9".
define variable t3 as integer no-undo format ">>9".
define variable srcName as character no-undo.
define variable iprocName as character no-undo.
empty temp-table tt_code_line.
for each tt_ptime no-lock by tt_ptime.avg_time descending:
/* if exec_count < 1 /* or src_line = 0 */ then next. */
find tt_source where
tt_source.id = tt_ptime.id and
tt_source.pid = tt_ptime.pid no-error.
if not available( tt_source ) then
srcName = "session".
else
srcName = tt_source.pname.
if srcName begins "lib/zprof" then next. /* don't include the profiler */
find tt_code_line where
tt_code_line.pid = tt_ptime.pid and
tt_code_line.src_line = tt_ptime.src_line and
tt_code_line.pname = srcName /* tt_source.pname */ no-error.
if not available tt_code_line then
do:
create tt_code_line.
assign
i = i + 1
tt_code_line.pid = tt_ptime.pid
tt_code_line.src_line = tt_ptime.src_line
tt_code_line.pname = srcName
.
end.
end.
/* message i "entries processed". pause. */
for each tt_code_line:
assign
tt_code_line.t1 = 0
tt_code_line.t2 = 0
.
for
each tt_source where
tt_source.pname = tt_code_line.pname,
each tt_ptime where
tt_ptime.id = tt_source.id and
tt_ptime.pid = tt_source.pid and
tt_ptime.src_line = tt_code_line.src_line:
assign
tt_code_line.t1 = tt_code_line.t1 + tt_ptime.exe_time
tt_code_line.t2 = tt_code_line.t2 + tt_ptime.exec_count
tt_code_line.t3 = tt_code_line.t3 + 1
.
if tt_ptime.pid = 0 and tt_ptime.src_line = 0 then tt_code_line.t1 = tt_ptime.tot_time.
end.
end.
for each tt_code_line:
tt_code_line.t4 = ( tt_code_line.t1 / tt_code_line.t2 ). /* calculate the average time... */
if num-entries( tt_code_line.pname, " " ) > 1 then
assign
tt_code_line.ipname = entry( 1, tt_code_line.pname, " " )
tt_code_line.pname = entry( 2, tt_code_line.pname, " " )
.
end.
return.
end.
procedure zprofiler_topx:
define input parameter zprofData as character no-undo.
define input parameter toTTY as logical no-undo.
define input parameter topLines as integer no-undo.
define variable c as integer no-undo.
define variable i as integer no-undo.
define variable t1 as decimal no-undo format ">>>>>9.999999".
define variable t2 as integer no-undo format ">>>>>>>>>9".
define variable t3 as integer no-undo format ">>9".
define variable t9 as integer no-undo.
find first tt_profile no-lock no-error. /* assuming that they're all the same date... */
for each tt_code_line no-lock where tt_code_line.pname <> "session":
t9 = t9 + tt_code_line.t1.
end.
if toTTY = no then output to value( zprofData + ".rpt" ).
display
tt_profile.description label "Description" format "x(70)" skip
"Session Total Execution Time " string( t9, "hh:mm:ss" ) skip
"Line 0 = initialization, line -1 = cleanup" skip
with frame prof-hdr
title " Profiler: Top Results "
width 120
centered
overlay
side-labels
row 4
.
i = 0.
for each tt_code_line no-lock by tt_code_line.t1 descending:
if tt_code_line.pname = "session" then next.
i = i + 1.
display
tt_code_line.pname label "Program/Class"
tt_code_line.src_line label "Line"
tt_code_line.t1 label "Time"
tt_code_line.t4 label "Avg Time"
tt_code_line.t2 label "Calls"
tt_code_line.ipname label "Internal Procedure/Method"
with frame prof-rpt
title " Top Lines: Total Execution Time "
width 120
centered
overlay
down
row 9
.
if i > topLines then leave.
end.
if toTTY = no then
output close.
else
do:
pause.
hide frame prof-rpt.
hide frame prof-hdr.
end.
return.
end.
/* main body
*
*/
zprofData = entry( 1, session:parameter, "|" ).
if num-entries( session:parameter, "|" ) = 2 then topLines = integer( entry( 2, session:parameter, "|" )).
run zprofiler_load( zprofData ). /* load profiler data into temp-tables to analyze */
run zprofiler_proc. /* process the data */
run zprofiler_topx( zprofData, no, topLines ). /* report on the top X execution time lines -- to file */
if session:batch = no then
run zprofiler_topx( zprofData, yes, topLines ). /* report on the top X execution time lines -- to TTY */
quit.
答案 1 :(得分：2)
您可以在以下位置找到探查器输出的ANTLR4语法：Link。
OpenEdge plugin for SonarQube的商业版已经具有代码覆盖功能。这是有关Git存储库（这是一个演示项目）中新提交的代码的代码覆盖范围的示例：Link
免责声明：我为Riverside Software工作