| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit changes the following in relation to string use
- The implicit convert from 'QCString' to 'const char *' is removed
- Strings parameters use 'const QCString &' as much as possible in favor
over 'const char *'
- 'if (s)' where s is a QCString has been replaced by 'if(!s.isEmpty())'
- data() now always returns a valid C-string and not a 0-pointer.
- when passing a string 's' to printf and related functions 'qPrint(s)' is
used instead of 's.data()'
- for empty string arguments 'QCString()' is used instead of '0'
- The copy() operation has been removed
- Where possible 'qstrcmp(a,b)==0' has been replaces by 'a==b' and
'qstrcmp(a,b)<0' has been replaced by 'a<b'
- Parameters of string type that were default initialized with '= 0' are
no initialized with '= QCString()'
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
slash or not.
The problem is that full qualified paths in `*_OUTPUT` were not handled properly they were handled as it were relative paths.
In the documentation it is stated for e.g. LATEX_OUTPUT:
> The LATEX_OUTPUT tag is used to specify where the LATEX docs will be put. If a relative path is entered the value of OUTPUT_DIRECTORY will be put in front of it.
So in case of a non relative path the given path should be used and this path can be unrelated to `OUTPUT_DIRECTORY`, so we have to store the designated output path as well.
|
| | |
|
| |\
| |
| |
| | |
https://github.com/cheoljoo/doxygen into cheoljoo-run_java_once_3rd_trial
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- this is doxygen documents.
- draw the plantuml
# Explanation how to process for FAST plantuml
```puml
@startuml
folder config.xml {
artifact PLANTUML_RUN_FAST
folder config.xml [
artifact xml
====
and style
]
}
folder debug.cpp {
artifact "Debug::Plantuml"
folder debug.h {
artifact "Plantuml=0x4000"
}
}
config.xml -down-> debug.cpp
debug.cpp -down-> docvisitor.cpp
folder docvisitor.cpp {
folder writePlantUMLSource {
folder writePlantUMLSource.old [
write plantuml file (inline_umlgraph_#.pu)
----
writePlantUMLFile()
generatePlantUMLOutput()
]
folder writePlantUMLSource.new [
write plantuml file (inline_umlgraph_#)
writePlantUMLFile()
generatePlantUMLOutput()
----
add PUML_Type as arguments
PlantumlManager::instance()->insert(generateType,puName,format,text);
]
writePlantUMLSource -down-> writePlantUMLSource.old : Original
writePlantUMLSource -down-> writePlantUMLSource.new : PLANTUML_RUN_FAST
}
artifact docbookvisitor.cpp
artifact htmldocvisitor.cpp
artifact latexdocvisitor.cpp
artifact rtfdocvisitor.cpp
artifact vhdldocgen.cpp
}
database PlantumlManager.DB [
static PlantumlManager *m_theInstance;
QDict< QList<QCString> > m_pngPlantumlFiles;
QDict< QList<QCString> > m_svgPlantumlFiles;
QDict< QList<QCString> > m_epsPlantumlFiles;
QDict< QCString > m_pngPlantumlContent;
QDict< QCString > m_svgPlantumlContent;
QDict< QCString > m_epsPlantumlContent;
]
writePlantUMLSource.new -down-> PlantumlManager.DB : PLANTUML_RUN_FAST
folder plantuml.cpp {
folder generatePlantUMLOutput.old [
run java for each inline_umlgraph_#.pu
]
folder generatePlantUMLOutput.new [
No run java for run once at the end of process
Just add image into list
]
writePlantUMLSource.old -down-> generatePlantUMLOutput.old : original
writePlantUMLSource.new -down-> generatePlantUMLOutput.new : PLANTUML_RUN_FAST
}
folder doxygen.cpp {
folder plantuml.h [
class PlantumlManager
]
}
folder plantuml.cpp {
folder PlantumlManager_run [
PlantumlManager::instance()->run();
- run java once at the last time of doxygen process like drawing dot graph
- read a inline_umlgraph_cache.pu
- if exist , reuse it.
- if not exist , add inline_umlgraph_type.pu
- save all plantuml into inline_umlgraph_cache.pu for caching or reusing next time.
- When we make inline_umlgraph_type.pu , we divide into 4 files for multi-processing.
- But , prcessing image count is less than 8 , we will use only one file.
]
}
plantuml.h -down-> writePlantUMLSource : PLANTUML_RUN_FAST
PlantumlManager.DB -down-> PlantumlManager_run : PLANTUML_RUN_FAST
@enduml
```
## cache
- QCString m_cachePlantumlAllContent; // = fileToString = readInputFile
// findScopePattern
// QCString.find contains
- instance() -> read cache file
- writePlantULMSource() -> compare the text with cache contents
- if matched , pass.
- if not matched , add it.
- add text to current plantuml
- run() -> save current plantuml into cache file
## todo list
- done
- html example
- cache
- ing
- multiple thread : divide into several files before running java. (already know the count)
- epstopdf run : change the code like the java run
|
| |/ |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
# What is it
- #6465 is the first trial. So I followed up the albert advice.
- in case of debug you use printf statements, I think that here there better options would be
- create an -d flag (e.g. -d plantuml) that gives the information, so no need for using the configuration option
- Follow up : create -d plantuml [debug.h, debug.cpp]
- usage of printf
- printf should not be used,all output is steered over calls to routines as defined in message.cpp,
- Follow up : use Debug::Print(Debug::Plantuml, ... and msg()
- usage of std::string / std::map
- in doxygen the string manipulation is. mostly, done by means of Qt classes (e.g. classes QCString, QString). For consistency these classes should be used.
- map manipulation is also done by means of Qt classes (dictionary and list classes)
- Follow up : use QDict QList QCString [plantuml.h, plantuml.cpp]
- dirent / opendir /readdir
- when browsing through directories the routines line readFileOrDirectory should probably be used or the techniques used in this routine
- Follow up : I will follow when I try to reduce redundancy of creating plantuml image.
- #6467 is the second trial. This is work items to satisfiy comments.
- In the debug statements the __PRETTY_FUNCTION__ is a GCC extension. But this project can run in windows.
- I got it. I removed __PRETTY_FUNCTION__.
- In the PLANTUML_RUN_JAVA_ONCE all the .pu files are given to the plantuml.jar with their full path (see -d extcmd when running). I think this will lead to problems when there are quite a few files ("command line" overflow). Seen this merging the files into one file like done for formulas. A small test with one file revealed that it didn't bring a lot / nothing time wise, but here the problem with the "command line" overflow shouldn't / can't happen.
- Yes. The command line has overflow problem. So I change into one file.
- I think it still would make sense to implement the 'md5' possibility for, especially, the non PLANTUML_RUN_JAVA_ONCE version, i.e. if the file already exists don't try to recreate the image.
- This is a next item.
- PlantumlManager::instance()->insert() : this function will be changed to skip image.
# Configuration
- add configuration value in Doxyfile
- PLANTUML_RUN_JAVA_ONCE = YES
# Debugging
- doxygen -d plantuml
- doxygen -d time -d extcmd
# Tested with following options
- source code of 3 plantuml
- PLANTUML_RUN_JAVA_ONCE = YES or NO
- DOT_IMAGE_FORMAT = png or svg
- DOT_CLEANUP = YES or NO
- -d plantuml
- -d time
- -d extcmd
|
| |
|
|
| |
not final
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
# What is it
- #6465 is the first reqeust. So I followed up the albert advice.
- in case of debug you use printf statements, I think that here there better options would be
- create an -d flag (e.g. -d plantuml) that gives the information, so no need for using the configuration option
- Follow up : create -d plantuml [debug.h, debug.cpp]
- usage of printf
- printf should not be used,all output is steered over calls to routines as defined in message.cpp,
- Follow up : use Debug::Print(Debug::Plantuml, ... and msg()
- usage of std::string / std::map
- in doxygen the string manipulation is. mostly, done by means of Qt classes (e.g. classes QCString, QString). For consistency these classes should be used.
- map manipulation is also done by means of Qt classes (dictionary and list classes)
- Follow up : use QDict QList QCString [plantuml.h, plantuml.cpp]
- dirent / opendir /readdir
- when browsing through directories the routines line readFileOrDirectory should probably be used or the techniques used in this routine
- Follow up : I will follow when I try to reduce redundancy of creating plantuml image.
- Run java minimally
- Test Case : 4 plantuml
- Original Run without PLANTUML_RUN_JAVA_ONCE
- 8.2 sec = 4 times * ( 1.6(java vm) + 0.1 * 1(load multiple file) + process each plantuml (0.35) *1 ) <- prediction
- New with PLANTUML_RUN_JAVA_ONCE
- 3.499 sec = 1.6(java vm) + 0.1 * 4(load multiple file) + process each plantuml (0.35) *4 <- prediction
- Creating Java Virtual Machine has a big portion.
- Result
- Improving Performance : 8.2 -> 3.499 (in case of 4 plantuml)
# Configuration
- add configuration value in Doxyfile
- PLANTUML_RUN_JAVA_ONCE = YES
# Debugging
- doxygen -d plantuml
- doxygen -d time -d plantuml
# Tested with following options
- source code of 3 plantuml
- PLANTUML_RUN_JAVA_ONCE = YES or NO
- DOT_IMAGE_FORMAT = png or svg
- DOT_CLEANUP = YES or NO
- -d plantuml
|
| | |
|
| |
|
